{
  "$schema": "folio-library-entry/v1",
  "id": "10665-381036",
  "blocks": [
    {
      "id": "library/10665-381036/blocks/prose-page-001",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1",
      "pageStart": 1,
      "pageEnd": 1,
      "target": "../sections/page-001.md",
      "narrative": null,
      "content": "WHO consolidated guidelines on \ntuberculosis. Module 3: Diagnosis  \n \n \n \n \n \nWeb Annex B. Evidence synthesis and \nanalysis",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title page of Web Annex B (evidence synthesis and analysis) to WHO's consolidated tuberculosis guidelines, Module 3: Diagnosis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-002",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 2",
      "pageStart": 2,
      "pageEnd": 2,
      "target": "../sections/page-002.md",
      "narrative": null,
      "content": "© World Health Organization 2025 \nSome rights reserved. This work is available under the Creative Commons Attribution-NonCommercial-\nShareAlike 3.0 IGO licence (CC BY-NC-SA 3.0 IGO; https://creativecommons.org/licenses/by-nc-\nsa/3.0/igo).  \nUnder the terms of this licence, you may copy, redistribute and adapt the work for non-commercial \npurposes, provided the work is appropriately cited, as indicated below. In any use of this work, there \nshould be no suggestion that WHO endorses any specific organization, products or services. The use of \nthe WHO logo is not permitted. If you adapt the work,",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The copyright and cataloguing page, with Creative Commons licence terms, noting that the annex supports the TB diagnosis guideline and is published for transparency. It also lists the usual WHO disclaimers on third-party materials and liability.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-003",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 3",
      "pageStart": 3,
      "pageEnd": 3,
      "target": "../sections/page-003.md",
      "narrative": null,
      "content": "ii \n \nWeb Annex B.1. Low complexity automated nucleic acid amplification tests (LC-aNAATs) for \ndetection of pulmonary TB and resistance to isoniazid/rifampicin in adults and adolescents: a \nsystematic review of diagnostic accuracy ........................................................................................... 1 \nWeb Annex B.2. Low complexity automated NAATs for pulmonary and extrapulmonary \ntuberculosis and rifampicin resistance in children: a systematic review of diagnostic accuracy ...... 45 \nWeb Annex B.3. Accuracy of LC-aNAATs for extrapulmonary tuberculosis and rifampicin \nre",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The first page of the table of contents, listing web annexes B.1 onwards that cover reviews of low complexity automated NAATs, other diagnostic tests and qualitative syntheses on targeted next-generation sequencing.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-004",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 4",
      "pageStart": 4,
      "pageEnd": 4,
      "target": "../sections/page-004.md",
      "narrative": null,
      "content": "iii \n \nWeb Annex B20. Accuracy of Mycobacterium tuberculosis antigen-based skin tests: a systematic \nreview and meta-analysis ................................................................................................................ 920 \nWeb Annex B 21. Safety of Mycobacterium tuberculosis antigen-based skin tests: a systematic \nreview and meta-analysis ................................................................................................................ 985 \nWeb Annex B.22. Cost–effectiveness of Mycobacterium tuberculosis antigen-based skin tests: a \nsystematic review ........",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The final page of the table of contents, listing annexes B.20 to B.23 on the accuracy, safety, cost-effectiveness and economic modelling of TB antigen-based skin tests.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-005",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 5",
      "pageStart": 5,
      "pageEnd": 5,
      "target": "../sections/page-005.md",
      "narrative": null,
      "content": "iv \n \nAcknowledgements \nThe World Health Organization (WHO) wishes to acknowledge the invaluable contributions \nmade by all the experts who carried out the systematic reviews included in this web annex. Their \nnames and affiliations are included in the sections concerned.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A brief acknowledgement thanking the experts who carried out the systematic reviews in the web annex.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-006",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 6",
      "pageStart": 6,
      "pageEnd": 6,
      "target": "../sections/page-006.md",
      "narrative": null,
      "content": "Web Annex B.1. Low complexity automated nucleic acid \namplification tests (LC-aNAATs) for detection of pulmonary TB \nand resistance to isoniazid/rifampicin in adults and adolescents: \na systematic review of diagnostic accuracy \n \nDavid J. Horne1, Leeberk Raja Inbaraj2, Jerry S. Zifodya3, Adrienne E. Shapiro4, Adhin Bhaskar2, E. \nChandler Church5, Jefferson Daniel6, Alexander W. Kay7, Jonah Kreniske8, Vittoria Lutje9, Mukesh \nKumar Sathya Narayanan2, Vignes Anand Srinivasalu2, Karen R. Steingart10, Yemisi Takwoingi11 \n \n1Division of Pulmonary, Critical Care & Sleep Medicine, University of Washi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title, author list and affiliations for Web Annex B.1, a systematic review of how accurately low complexity automated nucleic acid amplification tests (LC-aNAATs) detect pulmonary TB and isoniazid or rifampicin resistance in adults and adolescents.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-007",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 7",
      "pageStart": 7,
      "pageEnd": 7,
      "target": "../sections/page-007.md",
      "narrative": null,
      "content": "Background \nThere were an estimated 10.6 million people who newly developed tuberculosis (TB) in 2022 and of \nthe 7.5 million reported cases, 83% involved the lungs (WHO Global Tuberculosis Report 2023). In \n2022, there were about 410,000 new cases of rifampicin-resistant or multidrug resistant TB.  Several \nrapid molecular tests are recommended by the WHO as the initial diagnostic test for TB, some of which \ncan detect drug resistance simultaneously (WHO 2022). These tests can be used at the lower levels of \nthe health care system. \n \nWHO groups individual tests with similar characteristics a",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives background for the LC-aNAAT review, citing 2022 global estimates of TB incidence and rifampicin-resistant TB and WHO's recommendation of rapid molecular tests as initial diagnostics. It explains WHO's grouping of tests into classes and the clinical importance of detecting rifampicin and isoniazid resistance.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-008",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 8",
      "pageStart": 8,
      "pageEnd": 8,
      "target": "../sections/page-008.md",
      "narrative": null,
      "content": "Index tests  \nThe individual LC-aNAATs are: \n● GeneXpert MTB/RIF Ultra;  \n● Truenat (Truenat MTB Plus and Truenat MTB‐RIF Dx);  \n● STANDARD M10, SD Biosensor;  \n● IRON-qPCR, BIONEER \nFor ease of reference, we will refer to GeneXpert MTB/RIF Ultra as Xpert Ultra and the two Truenat \nassays as Truenat except where necessary to refer to a specific assay. \nXpert Ultra, Cepheid, USA \nXpert Ultra is a NAAT that functions as an automated closed system that performs real-time \npolymerase chain reaction (PCR). Xpert Ultra can be used for diagnosing TB and RR-TB. Xpert Ultra can \nbe performed by operato",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists the index tests in the LC-aNAAT class and describes the Xpert MTB/RIF Ultra and Truenat assays, including how they detect TB, report semi-quantitative results and target the rpoB gene for rifampicin resistance.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-009",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 9",
      "pageStart": 9,
      "pageEnd": 9,
      "target": "../sections/page-009.md",
      "narrative": null,
      "content": "The STANDARD M10 is an in vitro molecular diagnostics system that automatically extracts and \namplifies nucleic acids (DNA, RNA) from the collected samples, and analyses the data from \namplification in real-time, by combining the processes of isothermal amplification and real-time PCR \nin a single platform using all-in-one cartridge systems. There are two variations currently available for \nTB detection: STANDARD™ M10 MTB/NTM and STANDARD™ M10 MDR-TB. These assays are intended \nfor use with the STANDARD M10 system and provide qualitative results within 80 minutes from human \nsputum or sputum s",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the STANDARD M10 molecular system and its TB assays, then outlines the review's methods, including literature searches carried out in October 2023 across several databases.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-010",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 10",
      "pageStart": 10,
      "pageEnd": 10,
      "target": "../sections/page-010.md",
      "narrative": null,
      "content": "Cochrane Library, issue 10, October 2023; MEDLINE (Ovid); Embase (Ovid); Science Citation Index and \nBiosis previews (ISI Web of Knowledge); WHO Global Index Medicus; and SCOPUS (Elsevier). VL \nsearched ClinicalTrials.gov (clinicaltrials.gov/) and the WHO Clinical Trials Registry platform \n(www.who.int/trialsearch) to identify ongoing trials. VL also searched for relevant dissertations in \nProQuest Dissertations & Theses A&I. A WHO Public Call for ongoing and unpublished studies was \nreleased on November 30, 2023, to identify studies for inclusion in this systematic review. We reached \nout to",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists further sources searched for the LC-aNAAT review, including trial registries, dissertations and a WHO public call for data, and sets out eligibility criteria. It introduces the two reference standards used, culture and a composite reference standard.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-011",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 11",
      "pageStart": 11,
      "pageEnd": 11,
      "target": "../sections/page-011.md",
      "narrative": null,
      "content": "Culture (microbiological reference standard) \nCulture is considered the best reference standard for TB using either automated liquid culture or solid \nculture methods. Since both culture methods are used interchangeably in clinical practice (Moreira \n2015), a positive result either alone or in combination was accepted as a diagnosis of TB while all \nnegative cultures indicated no TB. \nComposite reference standard (CRS) \nWe defined CRS as a positive culture or a clinical decision to initiate treatment for PTB (i.e., clinically \ndiagnosed PTB). The CRS may be based on the results of microbiologi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Defines the culture and composite reference standards, then describes data extraction, quality assessment with QUADAS-2 and planned meta-analyses of sensitivity and specificity by subgroup such as smear and HIV status.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-012",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 12",
      "pageStart": 12,
      "pageEnd": 12,
      "target": "../sections/page-012.md",
      "narrative": null,
      "content": "LC-aNAAT class, we applied the pooling criteria that was developed by the WHO Global TB \nProgramme and Guideline Development Group (GDG) members for a previous GDG meeting. The \nclass pooling criteria is given in the Box above. We performed meta-analysis using the bivariate \nmodel to estimate summary sensitivity and specificity. When there was sparse data, we simplified \nthe model to a univariate random effects or fixed effect logistic regression model depending on the \nobserved variation in estimates of sensitivity or specificity between studies. We investigated \npotential sources of heteroge",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains the class pooling criteria and bivariate meta-analysis models used for the LC-aNAAT review, then reports search results for Xpert Ultra, yielding 33 included studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p013-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 13,
      "pageEnd": 13,
      "target": "../images/img-p013-1.webp",
      "narrative": "draft",
      "content": "PRISMA flow diagram of study selection for Xpert Ultra. From 1267 database records (746 duplicates removed), 521 were screened, 110 reports were assessed for eligibility and 79 were excluded with reasons listed; 8 further records came from a WHO public call. 33 studies were included in the review.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-013",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 13",
      "pageStart": 13,
      "pageEnd": 13,
      "target": "../sections/page-013.md",
      "narrative": null,
      "content": "Figure 1. Study flow diagram for Xpert Ultra",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A caption for the study flow diagram for Xpert Ultra.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p014-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 14,
      "pageEnd": 14,
      "target": "../images/img-p014-1.webp",
      "narrative": "draft",
      "content": "PRISMA flow diagram of study selection for Truenat assays. From 1175 database records (524 duplicates removed), 651 were screened and 34 reports were assessed for eligibility; 4 further reports came from citation searching and a WHO call for data. Five studies were included in the review (four for pulmonary TB and two for rifampicin resistance).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-014",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 14",
      "pageStart": 14,
      "pageEnd": 14,
      "target": "../sections/page-014.md",
      "narrative": null,
      "content": "Search Results: Truenat \nWe identified 1175 research articles for screening from searches of electronic sources. After removing \nduplicate studies (n=524), we screened the titles and abstracts of 651 unique articles and excluded \n617. We identified four reports through citation searching and WHO call for data. We performed full-\ntext screening of 38 articles and included five reports of which four contributed to the meta-analysis \nof Truenat MTB Plus and two included for the meta-analysis of rifampicin resistance (Figure 2). \n \n \n \n \n \n \n  \n \n \n \nFigure 2. Study flow diagram for Truenat assays",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports search results for the Truenat assays, with five reports included in the review, and captions the corresponding study flow diagram.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p015-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 15,
      "pageEnd": 15,
      "target": "../images/img-p015-1.webp",
      "narrative": "draft",
      "content": "PRISMA flow diagram of study selection for STANDARD M10 and other new LC-aNAATs. From 37 database records (3 duplicates removed), 34 were screened, 32 excluded and 2 reports assessed for eligibility; both were excluded (an ongoing trial and a conference abstract/case-control study). No studies were included.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-015",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 15",
      "pageStart": 15,
      "pageEnd": 15,
      "target": "../sections/page-015.md",
      "narrative": null,
      "content": "Search Results: STANDARD M10 and other index tests \nWe screened 34 titles and abstracts and excluded 32. We assessed the full texts of two articles and \nfound none eligible for this review (Figure 3). \n \nFigure 3.  Study flow diagram for STANDARD M10 and other new LC-aNAATs \n \nCharacteristics of included studies \nThere were 36 included studies: 31 that only evaluated Xpert Ultra (including one study based on two \nreports—Nicol 2018 and Zar 2019—and two studies extracted from a single report—Mishra 2020a & \n2020b), three studies that only evaluated Truenat MTB Plus (Gomathi 2020, Jose 2024, Nga",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that no eligible studies were found for STANDARD M10 or other new LC-aNAATs and summarizes the 36 included studies, mostly of Xpert Ultra, with eleven assessing rifampicin resistance detection.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-016",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 16",
      "pageStart": 16,
      "pageEnd": 16,
      "target": "../sections/page-016.md",
      "narrative": null,
      "content": "Respiratory specimen types \nMost included studies used sputum for their evaluation of LC-aNAATs. Three studies exclusively \nevaluated Xpert Ultra testing of BAL samples (Chien 2020, Park 2019, Zhang 2021) while one study \nincluded BAL in addition to sputum samples (Sava 2023).  One study (Penn-Nicholson 2021) pooled \nmultiple sputum specimens per participant in the reference laboratory and used unpooled sputum in \nthe peripheral labs. Sabi 2018 used the best result from multiple specimens (1 to 4) per participant. Li \n2023 used three sputum samples per person and examined each sample microscop",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the respiratory specimen types used in the included studies, mainly sputum, with some using bronchoalveolar lavage or pooled samples.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-017",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 17",
      "pageStart": 17,
      "pageEnd": 17,
      "target": "../sections/page-017.md",
      "narrative": null,
      "content": "12\nTable 1. Xpert Ultra: Characteristics of included studies  \nStudy \nReference \nstandard \nStudy design \nProportion of \nPLHIV \nClinical \nsetting  \nAge group \nCountry \nHigh TB \nburden? \nSpecimen type \nXpert Ultra \nnon-determinate \nfor detection of \nMTBa % \n(number/total) \nAguilera-Alonso \n2022 \nSolid & liquid – \n1 to 3 \nspecimens \nProspective \nCohort \nNone \nInpatient \nAdolescents Spain \nNo \nFresh sputum \nNR \nBerhanu 2018 \nSolid (LJ) & \nliquid \nProspective \ncohort \n62% (147/237) \nOutpatient \nAdults \nRSA \nYes \nFresh sputum \nNR \nBoyles 2020 \nLiquid culture - \n2 specimens \nCohort \n100% (207/207) \nO",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The first page of Table 1, giving characteristics of the Xpert Ultra studies, such as reference standard, design, share of people with HIV, setting, age group, country and specimen type.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-001",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 1. Xpert Ultra: Characteristics of included studies",
      "pageStart": 17,
      "pageEnd": 19,
      "target": "../tables/table-1.csv",
      "narrative": "draft",
      "content": "Characteristics of the studies included in the review of Xpert Ultra. Each row is a study; columns give the reference standard, study design, proportion of people living with HIV, clinical setting, age group, country, whether the country has a high TB burden, specimen type, and the proportion of non-determinate Xpert Ultra results for detection of MTB.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-018",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 18",
      "pageStart": 18,
      "pageEnd": 18,
      "target": "../sections/page-018.md",
      "narrative": null,
      "content": "13\nStudy \nReference \nstandard \nStudy design \nProportion of \nPLHIV \nClinical \nsetting  \nAge group \nCountry \nHigh TB \nburden? \nSpecimen type \nXpert Ultra \nnon-determinate \nfor detection of \nMTBa % \n(number/total) \nMishra 2020b \nLiquid \nCross-sectional 44% (152/346) \nUnknown  \nAdults \nRSA \nYes \nFresh sputum \nNR \nMoodley 2022 \nLiquid \nCross-sectional 38% (298/789) \nOutpatient  Adults \nRSA \nYes \nFresh sputum \n<1% (1/789) \nMukoka 2023 \nSolid & liquid \nRCT \n100% (319/319) \nOutpatient  Adults \nMalawi \nYes \nFresh sputum \nNR \nOpota 2019 \nLiquid \nRetrospective \ncohort  \nNR \nInpatient \nNR \nSwitzerland \nNo",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues Table 1 of Xpert Ultra study characteristics, covering further studies from South Africa, Malawi, Switzerland, China and elsewhere.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-019",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 19",
      "pageStart": 19,
      "pageEnd": 19,
      "target": "../sections/page-019.md",
      "narrative": null,
      "content": "14\nStudy \nReference \nstandard \nStudy design \nProportion of \nPLHIV \nClinical \nsetting  \nAge group \nCountry \nHigh TB \nburden? \nSpecimen type \nXpert Ultra \nnon-determinate \nfor detection of \nMTBa % \n(number/total) \ncohort \n& adults \nsputum \nWang 2022 \nLiquid \nCross-sectional 0% (0/1079) \nInpatient  \nAdolescents \n& adults \nChina \nYes \nFresh sputum \n<0.1% (4/1079) \nZar 2019 \nLiquid - 2 \nspecimens \nProspective \ncohort \n0% (0/5) \nInpatient  \nAdolescents RSA \nYes \nFrozen sputum \nNR \nZhang 2021 \nSolid & liquid \nCohort \n100% (99/99) \nInpatient  \nAdults \nChina \nYes \nBAL/bronchial \nwash (fresh \nNR \nLJ: Lö",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Ends Table 1 of Xpert Ultra study characteristics, with abbreviations for culture media and other terms.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-020",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 20",
      "pageStart": 20,
      "pageEnd": 20,
      "target": "../sections/page-020.md",
      "narrative": null,
      "content": "15\nTable 2. Truenat assays: Characteristics of included studies  \nStudy \nReference \nstandard \nStudy design \nProportion \nof PLHIV  \nClinical setting  \nAge group \nCountry  \nHigh TB \nburden \nSpecimen \ntype \nTruenat non-\ndeterminate \n% (number/total) \nJose 2024 \nLJ & MGIT \nCross- sectional \n(Retrospective) \nNot reported  Outpatients & \ninpatients from \ntertiary care \nhospital  \nAdolescents \n& adults \nIndia \nYes  \nFresh  \nNot reported  \nGomathi 2020 \nMGIT \nCross-sectional  \nNot reported  Outpatients \nfrom all settings \nAdults \nIndia \nYes \nFresh and \nstored \nsputum \nNot reported \nNgangue 2022 \nLJ &",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 2, giving characteristics of the Truenat studies, conducted mainly in India and several African and Asia-Pacific countries, including rates of non-determinate results.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-002",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 2. Truenat assays: Characteristics of included studies",
      "pageStart": 20,
      "pageEnd": 20,
      "target": "../tables/table-2.csv",
      "narrative": "draft",
      "content": "Characteristics of the studies included in the review of Truenat assays. Each row is a study; columns give the reference standard, study design, proportion of people living with HIV, clinical setting, age group, country, high TB burden status, specimen type, and the proportion of non-determinate Truenat results.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-021",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 21",
      "pageStart": 21,
      "pageEnd": 21,
      "target": "../sections/page-021.md",
      "narrative": null,
      "content": "Methodological quality \nFigure 4 summarises the results of the risk of bias and applicability assessment for each of the \nincluded studies.  \nThe following studies were assessed as “high” risk of bias and/or applicability concern: \n● Aguilera-Alonso 2022 - Patient selection, applicability concern; Reference standard \napplicability concern based on recruitment of participants exclusively from tertiary referral \nhospitals and lack of speciation of mycobacteria isolated in culture. \n● Jose 2024 - Reference standard, high risk of bias; Patient selection, applicability concern \nbased on the referen",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarizes the risk of bias and applicability assessment for the included LC-aNAAT studies, listing those judged high risk or of concern and why.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p022-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 22,
      "pageEnd": 22,
      "target": "../images/img-p022-1.webp",
      "narrative": "draft",
      "content": "Risk-of-bias and applicability-concerns chart for 36 studies of LC-aNAATs on respiratory specimens with a microbiological reference standard. Each study is rated high (red), unclear (yellow) or low (green) for patient selection, index test (LC-aNAATs), reference standard and flow and timing; most ratings are low. High or unclear ratings occur mainly for patient selection.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-022",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 22",
      "pageStart": 22,
      "pageEnd": 22,
      "target": "../sections/page-022.md",
      "narrative": null,
      "content": "Figure 5. Risk of bias and applicability summary: respiratory specimens and microbiological reference",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A caption for the summary figure of risk of bias and applicability across studies of respiratory specimens.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p023-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 23,
      "pageEnd": 23,
      "target": "../images/img-p023-1.webp",
      "narrative": "draft",
      "content": "Stacked horizontal bar chart summarizing risk of bias and applicability concerns across 36 studies, as percentages of studies rated high (red), unclear (yellow) or low (green). For risk of bias, patient selection was 3 high, 5 unclear and 28 low; index test 1 unclear and 35 low; reference standard 1 high, 4 unclear and 31 low; flow and timing 2 unclear and 34 low. For applicability, patient selection was 8 high, 4 unclear and 24 low; index test 1 unclear and 35 low; reference standard 1 high, 6 unclear and 29 low.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-023",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 23",
      "pageStart": 23,
      "pageEnd": 23,
      "target": "../sections/page-023.md",
      "narrative": null,
      "content": "standard. Review authors' judgements about each domain presented as percentages across included \nstudies. The numbers shown on each bar indicate the number of studies in each judgement category. \n \n \nFindings  \n \nSummary estimates of diagnostic accuracy \nTable 3 summarises the results of the meta-analyses for LC-aNAATs and each index test in the class, \nincluding sensitivity and subgroup analyses. Main results are presented below with summary of findings \ntables in Appendix 2 (Appendix Tables 1 to 3) and additional forest plots provided in Appendix 3. \n \nTable 3. Diagnostic accuracy of LC-aNAA",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Begins the findings of the LC-aNAAT review, introducing Table 3 of summary sensitivity and specificity estimates for the class and individual tests. The table rows cover analyses by reference standard, specimen and smear status.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-003",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 3. Diagnostic accuracy of LC-aNAATs for pulmonary TB and rifampicin resistance in adults and adolescents",
      "pageStart": 23,
      "pageEnd": 24,
      "target": "../tables/table-3.csv",
      "narrative": "draft",
      "content": "Summary results of meta-analyses of the diagnostic accuracy of low-complexity automated NAATs (LC-aNAATs), Xpert Ultra and Truenat assays for detecting pulmonary TB and rifampicin resistance in adults and adolescents. Each numbered row is an analysis (test, reference standard and subgroup), with the number of studies and participants (cases), summary sensitivity and specificity, and PPV and NPV with 95% CIs.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-024",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 24",
      "pageStart": 24,
      "pageEnd": 24,
      "target": "../sections/page-024.md",
      "narrative": null,
      "content": "No. Analysis \nNumber \nof \nstudies \nNumber of \nparticipants \n(cases) \nSummary \nsensitivity \n(95% CI) \nSummary \nspecificity (95% \nCI) \nPPV (95% CI)1 \nNPV (95% CI)1 \n14    Xpert Ultra, MRS, \nhistory of TB \n9 \n809 (152) \n84.8  \n(78.2 to 89.7) \n86.2  \n(78.9 to 91.3) \n40.7  \n(29.1 to 53.5) \n98.1  \n(97.0 to 98.8) \n15 \nXpert Ultra, MRS, \nno history of TB \n9 \n3423 (1134) \n91.5  \n(87.9 to 94.1) \n95.1  \n(91.6 to 97.1) \n67.4  \n(53.9 to 78.5) \n99.0  \n(98.6 to 99.3) \n16 \nXpert Ultra, MRS, \nadults only \n21 \n9585 (2277) \n90.7 \n(87.5 to 93.1) \n95.0 \n(93.1 to 96.4) \n66.9 \n(58.5 to 74.2) \n98.9 \n(98.5 to 99.2) \n1",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues Table 3 of LC-aNAAT diagnostic accuracy, giving pooled sensitivity, specificity and predictive values by history of TB, age group, HIV status and for rifampicin resistance. Predictive values are calculated at an assumed 10% pre-test probability.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-025",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 25",
      "pageStart": 25,
      "pageEnd": 25,
      "target": "../sections/page-025.md",
      "narrative": null,
      "content": "sensitivity and specificity and their 95% CIs. \n2We performed meta-analysis of sensitivity using univariate fixed effect logistic regression because \nthere was little variation in estimates of sensitivity between studies and in many studies, there were \nzero false positives and true negatives (i.e. zero non-cases) so a bivariate meta-analysis was not \nfeasible. \n3Same results for LC-aNAATs because no adolescent only data for Truenat MTB Plus. \n4Specificity was not estimable for one study, 0% for one study and 75% for the third study. Meta-\nanalysis was not performed. \n5Excluding Gomathi 2020 d",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives footnotes to the accuracy table explaining the statistical models used, then discusses studies with lower sensitivity. One example is a Malawi trial enrolling only people with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p026-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 26,
      "pageEnd": 26,
      "target": "../images/img-p026-1.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals for 36 studies of LC-aNAATs for pulmonary TB, listing true and false positives and negatives and the brand (4 Truenat MTB Plus, 32 Xpert Ultra). Most sensitivities lie between about 0.8 and 1.0 (lowest 0.54, Mukoka 2023) and most specificities above 0.9; one study (Zar 2019) has sensitivity not estimable.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-026",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 26",
      "pageStart": 26,
      "pageEnd": 26,
      "target": "../sections/page-026.md",
      "narrative": null,
      "content": "performed on the second sample. Both liquid and solid media were used for culture. \n \nDescription of studies with lower specificity:  \nMishra 2020b: Participants were exclusively persons with a history of treated TB in the past two years. \nLikely, many of these participants would have had residual DNA from M. tuberculosis leading to a \npositive NAAT and negative TB culture.  \n \nWang 2019: This study had a high rate of Xpert Ultra positive/culture-negative results. Participants with \nsuspected PTB had one sputum that underwent MGIT testing. This study exclusively recruited persons \nwho were sme",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Discusses studies with lower specificity, such as those enrolling people previously treated for TB, and introduces the analysis against a composite reference standard, done only for Xpert Ultra.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-027",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 27",
      "pageStart": 27,
      "pageEnd": 27,
      "target": "../sections/page-027.md",
      "narrative": null,
      "content": "_(no text on this page)_",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A blank page with no extracted text.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-vfig-p028",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 28,
      "pageEnd": 28,
      "target": "../figures/vfig-p028.png",
      "narrative": "draft",
      "content": "Table of composite reference standard (CRS) definitions with columns 'Study' and 'CRS Definition'. It lists ten studies (Aguilera-Alonso 2022, Berhanu 2018, Chien 2020, Jaganath 2021, Opota 2019, Park 2019, Sabi 2018, Wang 2022, Zar 2019, Zhang 2021), each with the population (adults and/or adolescents) and how confirmed and unconfirmed or probable TB were defined, for example by culture, clinical and radiological findings, or response to TB treatment.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-028",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 28",
      "pageStart": 28,
      "pageEnd": 28,
      "target": "../sections/page-028.md",
      "narrative": null,
      "content": "Table 4. CRS Definitions \nStudy \nCRS Definition \nAguilera-Alonso \n2022 \nAdolescents; CRS definition: (1) Confirmed TB: >1 respiratory specimen positive for MTB \n(culture or PCR); (2) Unconfirmed TB: without bacteriological evidence who met >1 (if \nimmunological evidence of MTB) or 2 (immunological evidence absent) criteria: \nsymptoms/signs suggestive of TB, CXR consistent with TB, close TB exposure, response to \nTB treatment. Immunological evidence of MTB infection defined as positive TST or IGRA. \nBerhanu 2018 \nAdult patients with clinical diagnosis of TB by X ray or other method. \nChien 2020",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 4, listing the composite reference standard definitions used in each Xpert Ultra study, with a caption for the related risk-of-bias figure.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-004",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 4. CRS Definitions",
      "pageStart": 28,
      "pageEnd": 28,
      "target": "../tables/table-4.csv",
      "narrative": "draft",
      "content": "Definitions of the composite reference standard (CRS) used in each included study. Each row names a study and describes the population and the criteria it used to classify participants as having TB under the CRS.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p029-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 29,
      "pageEnd": 29,
      "target": "../images/img-p029-1.webp",
      "narrative": "draft",
      "content": "Risk-of-bias and applicability-concerns chart for 10 studies of LC-aNAATs on respiratory specimens with a composite reference standard. Each study is rated high (red), unclear (yellow) or low (green) for patient selection, index test (LC-aNAATs), reference standard and flow and timing; most ratings are low.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p029-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 29,
      "pageEnd": 29,
      "target": "../images/img-p029-2.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals for 10 studies of LC-aNAATs for pulmonary TB against a composite reference standard. Sensitivity varies widely, from 0.00 (Zar 2019) to 0.96 (Opota 2019), while specificity is 0.96–1.00 in all studies.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-029",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 29",
      "pageStart": 29,
      "pageEnd": 29,
      "target": "../sections/page-029.md",
      "narrative": null,
      "content": "The sensitivities in the 10 studies (2127 participants) were between 0% and 96% and the specificities \nwere between 96% and 100% (Figure 8). The summary sensitivity (95% CI) was 69.9% (52.7 to 82.9) and \nthe summary specificity was 98.8% (96.2 to 99.7).  \n \nFor a population of 1000 adults and adolescents where 100 have pulmonary tuberculosis on a composite \nreference standard, 80 would be LC-aNAAT-positive and 10 (13%) would not have tuberculosis (false-\npositives); 920 would be LC-aNAAT-negative and 30 (3%) would have tuberculosis (false-negatives). \n \n \nFigure 8. Forest plot of LC-aNAATs for",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that against a composite reference standard, LC-aNAATs had a summary sensitivity of about 70% and specificity of about 99%, with illustrative numbers per 1000 people tested. It then introduces the analysis for people living with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p030-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 30,
      "pageEnd": 30,
      "target": "../images/img-p030-1.webp",
      "narrative": "draft",
      "content": "Risk-of-bias and applicability-concerns chart for 12 studies of LC-aNAATs in people living with HIV. Each study is rated high (red), unclear (yellow) or low (green) for patient selection, index test (LC-aNAATs), reference standard and flow and timing; most ratings are low.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p030-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 30,
      "pageEnd": 30,
      "target": "../images/img-p030-2.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals for 12 studies of LC-aNAATs for pulmonary TB in people living with HIV (2 Truenat MTB Plus, 10 Xpert Ultra). Sensitivity ranges from 0.54 (Mukoka 2023) to 1.00 and specificity from 0.78 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-030",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 30",
      "pageStart": 30,
      "pageEnd": 30,
      "target": "../sections/page-030.md",
      "narrative": null,
      "content": "included studies. As previously noted, there are applicability concerns in patient selection for Park 2019 \nbased on recruitment of inpatients exclusively from a tertiary referral hospital.  \nFigure 9. Risk of bias and applicability concern for each study: respiratory specimens \nand microbiological reference standard in PLHIV \n \n \nTwelve studies (2016 participants) evaluated sputum specimens from PLHIV (Figure 10). The \nsensitivities ranged between 54% and 100% and the specificities between 78% and 100%. The summary \nsensitivity (95% CI) was 87.4% (83.8 to 90.3) and the summary specificity was",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarizes the twelve studies of sputum testing in people living with HIV, giving the range of sensitivities and specificities, with captions for the related figures.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p031-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 31,
      "pageEnd": 31,
      "target": "../images/img-p031-1.webp",
      "narrative": "draft",
      "content": "Risk-of-bias and applicability-concerns chart for 6 studies of LC-aNAATs on sputum in HIV-negative people. Each study is rated high (red), unclear (yellow) or low (green) for patient selection, index test (LC-aNAATs), reference standard and flow and timing; most ratings are low. The only high rating is applicability of patient selection for Park 2019.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p031-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 31,
      "pageEnd": 31,
      "target": "../images/img-p031-2.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals for 6 studies of LC-aNAATs for pulmonary TB in HIV-negative people (2 Truenat MTB Plus, 4 Xpert Ultra). Sensitivities range from 0.88 to 0.97 and specificities from 0.91 to 0.99.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-031",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 31",
      "pageStart": 31,
      "pageEnd": 31,
      "target": "../sections/page-031.md",
      "narrative": null,
      "content": "Should LC-aNAATs on respiratory samples be used to diagnose pulmonary TB \nin HIV negative adults and adolescents with signs and symptoms or screened \npositive for pulmonary TB, against a microbiological reference standard? \nFigure 11 summarises the results of the risk of bias and applicability assessment for each of the six \nincluded studies. As previously noted, there are applicability concerns in patient selection for Park 2019 \nbased on recruitment of inpatients exclusively from a tertiary referral hospital.  \nFigure 11. Risk of bias and applicability concern for each study: sputum specimen",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces the analysis of LC-aNAATs in HIV-negative adults and adolescents and then describes the frequency of Xpert Ultra trace results across studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p032-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 32,
      "pageEnd": 32,
      "target": "../images/img-p032-1.webp",
      "narrative": "draft",
      "content": "Risk-of-bias and applicability-concerns chart for 13 studies of LC-aNAATs for detection of rifampicin resistance. Each study is rated high (red), unclear (yellow) or low (green) for patient selection, index test (LC-aNAATs), reference standard and flow and timing; most ratings are low.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-032",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 32",
      "pageStart": 32,
      "pageEnd": 32,
      "target": "../sections/page-032.md",
      "narrative": null,
      "content": "ranged between 0% and 81.8%. Trace results were considered test positives in the primary analysis of \n34 studies of LC-aNAATs for pulmonary TB giving a summary sensitivity (95% CI) of 90.3% (87.7 to 92.4) \nand summary specificity of 94.9% (93.1 to 96.3). In a sensitivity analysis with trace results reclassified \nas test negatives, the summary estimates (95% CI) of sensitivity and specificity from the 34 studies were \n87.6% (84.9 to 90.0) and 96.4% (94.9 to 97.5) (Appendix 4, Appendix Table 5). \n \n \nDetection of rifampicin resistance \nShould LC-aNAATs on respiratory samples be used to diagnose",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports how classifying Xpert Ultra trace results as positive or negative changes pooled sensitivity and specificity, and introduces the rifampicin resistance analysis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p033-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 33,
      "pageEnd": 33,
      "target": "../images/img-p033-1.webp",
      "narrative": "draft",
      "content": "Risk-of-bias and applicability-concerns chart for 13 studies of LC-aNAATs for detection of rifampicin resistance. Each study is rated high (red), unclear (yellow) or low (green) for patient selection, index test (LC-aNAATs), reference standard and flow and timing; most ratings are low. It is nearly identical to the chart on the previous page.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p033-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 33,
      "pageEnd": 33,
      "target": "../images/img-p033-2.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals for detection of rifampicin resistance in 13 studies (2 Truenat MTB-RIF Dx, 11 Xpert Ultra). Truenat MTB-RIF Dx sensitivities are 0.53 and 0.85; Xpert Ultra sensitivities are 0.93–1.00 where estimable, and specificities are 0.97–1.00 across studies.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-033",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 33",
      "pageStart": 33,
      "pageEnd": 33,
      "target": "../sections/page-033.md",
      "narrative": null,
      "content": "Of the 13 (2553 participants) studies that evaluated sputum specimens, sensitivity was not estimable \nfor two studies (Figure 14). The sensitivities in the remaining 11 studies (2540 participants) included in \nthe meta-analysis were between 53% and 100% and the specificities were between 97% and 100%. The \nsummary sensitivity (95% CI) was 95.1% (83.1 to 98.7) and the summary specificity was 98.1% (97.0 to \n98.7). The certainty of evidence for both sensitivity and specificity were graded as “High”.  \nFor a population of 1000 adults and adolescents with pulmonary tuberculosis where 100 have rifa",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports a pooled sensitivity of about 95% and specificity of about 98% for LC-aNAAT detection of rifampicin resistance, with evidence graded high certainty.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-034",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 34",
      "pageStart": 34,
      "pageEnd": 34,
      "target": "../sections/page-034.md",
      "narrative": null,
      "content": "true positive. \n \n \n \n \n \n \nShould LC-aNAATs on respiratory samples be used to diagnose rifampicin \nresistance in PLHIV (adults and adolescents) with signs and symptoms or \nscreened positive for pulmonary TB against a microbiological reference \nstandard? \nWe identified no studies on this topic that met inclusion criteria. Several studies in Figure 14 either did \nnot include PLHIV or did not report participants’ HIV status (Aguilera-Alonso 2022, Chakravorty 2017, \nChien 2020, Gomathi 2020, Li 2023, Piersimoni 2019, Wang 2019, Wang 2022), did not have rifampicin \nresistant isolates in PLHIV (Par",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes that no eligible studies assessed rifampicin resistance detection in people living with HIV, then begins the review's discussion and main conclusions, including the effect of how trace results are interpreted.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-035",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 35",
      "pageStart": 35,
      "pageEnd": 35,
      "target": "../sections/page-035.md",
      "narrative": null,
      "content": "o Consistent with summary estimates in this meta-analysis, if Xpert Ultra trace positives are \nconsidered “positive”: summary sensitivity 90.7% (88.2 to 92.7) and summary specificity \n94.8% (92.8 to 96.3) \no If Xpert Ultra trace positives are considered “negative”, summary sensitivity 87.9% (85.2 to \n90.3) and summary specificity 96.5% (94.8 to 97.6).   \n● \nLC-aNAAT for the detection of rifampicin resistance had a summary sensitivity of 95.1% (83.1 to \n98.7) and summary specificity of 98.1% (97.0 to 98.7) based on 11 studies. Two of the included \nstudies evaluated Truenat MTB-RIF Dx as the ind",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Concludes the main findings on LC-aNAAT accuracy for TB and rifampicin resistance, discusses strengths, limitations and implications, and states authors' contributions.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-036",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 36",
      "pageStart": 36,
      "pageEnd": 36,
      "target": "../sections/page-036.md",
      "narrative": null,
      "content": "References \nIncluded Xpert Ultra Studies \n1. Aguilera-Alonso, D.; Solis-Garcia, G.; Noguera-Julian, A.; Gonzalez-Martin, J.; Roman Cobena, A.; \nBaquero-Artigao, F.; Toro Rueda, C.; Rodriguez-Molino, P.; Bloise Sanchez, I.; Vallmanya, T.; \nBernet-Sanchez, A.; Minguell Domingo, L.; Rubio, A.; Saavedra-Lozano, J.; Ruiz-Serrano, M. J.; \nBlazquez-Gamero, D.; Lopez-Roa, P.; Gomez-Pastrana, D.; Lopez Prieto, M. D.; Lopez Medina, E. \nM.; Gil-Brusola, A.; Martin Nalda, A.; Soriano-Arandes, A.; Tortola, T.; Falcon-Neyra, L.; Gonzalez \nGalan, V.; Tebruegge, M.; Santiago-Garcia, B.. Accuracy of Xpert Ultr",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The start of the reference list of included Xpert Ultra studies for the LC-aNAAT review.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-037",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 37",
      "pageStart": 37,
      "pageEnd": 37,
      "target": "../sections/page-037.md",
      "narrative": null,
      "content": "MTB/RIF Ultra and digital chest radiography for the diagnosis and treatment of TB in people living \nwith HIV: a randomised controlled trial (XACT-TB). Trans R Soc Trop Med Hyg. 2023;117(1):28-37. \n16. Nicol MP, Workman L, Prins M, Bateman L, Ghebrekristos Y, Mbhele S, et al. Accuracy of Xpert \nMtb/Rif Ultra for the Diagnosis of Pulmonary Tuberculosis in Children. Pediatr Infect Dis J. \n2018;37(10):e261-e3. \n17. Opota O, Zakham F, Mazza-Stalder J, Nicod L, Greub G, Jaton K. Added Value of Xpert MTB/RIF \nUltra for Diagnosis of Pulmonary Tuberculosis in a Low-Prevalence Setting. J Clin Microbiol.",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list of included Xpert Ultra studies, citing diagnostic accuracy studies from various countries.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-038",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 38",
      "pageStart": 38,
      "pageEnd": 38,
      "target": "../sections/page-038.md",
      "narrative": null,
      "content": "Children Using Xpert Ultra on Different Respiratory Specimens. Am J Respir Crit Care Med. \n2019;200(12):1531-8. \n33. Zhang P, Liu H, Wang H, Wu Y, Sun L, Rao M, et al. Performance of Xpert MTB/RIF Ultra for the \nDiagnosis of Pulmonary Tuberculosis Using Bronchoalveolar Lavage Samples in People Living with \nHIV/AIDS (PLWHA) in China: A Prospective Study. HIV AIDS (Auckl). 2021;13:905-16. \nIncluded Truenat MTB Plus and Truenat MTB-RIF Dx studies  \n1. Gomathi NS, Singh M, Myneedu VP, Chauhan DS, Tripathy SP, Sarin R, et al. Validation of an \nindigenous assay for rapid molecular detection of rifam",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Ends the Xpert Ultra references and lists included Truenat studies, among them an unpublished South African evaluation.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-039",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 39",
      "pageStart": 39,
      "pageEnd": 39,
      "target": "../sections/page-039.md",
      "narrative": null,
      "content": "Appendix \nAppendix 1. Detailed search strategies \n \nOvid MEDLINE(R) ALL <1946 to October 13, 2023> \n1       \npulmonary Tuberculosis/        \n2       \nmycobacterium tuberculosis.mp. or Mycobacterium tuberculosis/  \n3       \nTuberculosis, Multidrug-Resistant/    \n4       \nExtensively Drug-Resistant Tuberculosis/       \n5       \n(((tuberculosis or TB) adj3 (lung* or pulmonic or bronchial or pulmonary)) or \n((tuberculosis or TB) adj3 (respiratory or respirational))).mp. \n6       \n(tuberculosis adj3 (drug resistan* or multidrug resistan* or mdr or xdr)).mp.         \n7       \n(((isoniazid adj3 resis",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Appendix 1, giving the detailed MEDLINE search strategy for the LC-aNAAT review.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-040",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 40",
      "pageStart": 40,
      "pageEnd": 40,
      "target": "../sections/page-040.md",
      "narrative": null,
      "content": "15     \n(Xpert* or GeneXpert or cepheid or near* patient).tw.         \n16     \nultra.tw.           \n17     \n15 and 16        \n18     \n14 and 17        \n19     \n(Truenat* or Molbio or TruRif).mp. \n20     \n14 and 19        \n21     \n(STANDARD M10 or SD Biosensor or IRON-qPCR or BIONEER).mp.   \n22     \n14 and 21        \n \n Web of Science Search Strategy (Science Citation index-Expanded and Biosis Previews) \n#1 Search: (tuberculosis or TB) and (lung or pulmonary)  (Topic) OR multidrug resistant \ntuberculosis or extensively drug resistant tuberculosis  (Topic) OR MDR-TB or XDR-TB  \n(Topic)",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the search strategies, covering MEDLINE terms for the index tests and searches in Web of Science and Scopus.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-041",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 41",
      "pageStart": 41,
      "pageEnd": 41,
      "target": "../sections/page-041.md",
      "narrative": null,
      "content": "Search Name: Cochrane Central Register of Controlled Trials \nIssue 10 of 12, October 2023 \n#1     \nMeSH descriptor: [Tuberculosis] explode all trees     \n#2     \nMeSH descriptor: [Mycobacterium tuberculosis] explode all trees   \n#3     \n((tuberculosis or TB or MDR-TB or XDR-TB)):ti,ab,kw            \n#4     \nMeSH descriptor: [Tuberculosis, Multidrug-Resistant] explode all trees       \n#5     \nextensively drug resistant tuberculosis          \n#6     \nMeSH descriptor: [Extensively Drug-Resistant Tuberculosis] explode all trees         \n#7     \nmultidrug resistant tuberculosis         \n#8     \n#1",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives the search strategies used in the Cochrane Central Register and ProQuest dissertations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-042",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 42",
      "pageStart": 42,
      "pageEnd": 42,
      "target": "../sections/page-042.md",
      "narrative": null,
      "content": "Appendix 2. Summary of findings tables  \n \nAppendix Table 1. Should LC-aNAATs on respiratory samples be used to diagnose pulmonary TB in \nadults and adolescents with signs and symptoms or screened positive for pulmonary TB, against a \nmicrobiological reference standard? In the 34 studies (14840 participants) included in the meta-\nanalysis, sensitivity ranged between 54% and 100% and specificity between 71% and 100%. The \nsummary sensitivity (95% CI) was 90.4% (88.0 to 92.4) and the summary specificity was 94.9 % (93.0 \nto 96.3). The certainty of evidence for both sensitivity and specificity we",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Appendix 2, summary of findings tables for LC-aNAATs in diagnosing pulmonary TB and rifampicin resistance, giving pooled accuracy and certainty of evidence.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-043",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 43",
      "pageStart": 43,
      "pageEnd": 43,
      "target": "../sections/page-043.md",
      "narrative": null,
      "content": "evidence for both sensitivity and specificity was graded as high. \n \nTest result \nNumber of results per 1,000 patients tested (95% CI) \nNumber of \nparticipants  \n(studies) \nCertainty of the \nEvidence \n(GRADE) \nPrevalence2% \nTypically seen in \nPrevalence10% \nTypically seen in \nPrevalence15% \nTypically seen in \nTrue positives \n19 (17 to 20) \n95 (83 to 99) \n143 (125 to 148) \n491 \n(11) \n⨁⨁⨁⨁ \nHigh \nFalse negatives \n1 (0 to 3) \n5 (1 to 17) \n7 (2 to 25) \nTrue negatives \n961 (951 to 967) \n883 (873 to 888) \n834 (825 to 839) \n2026 \n(11) \n⨁⨁⨁⨁ \nHigh \nFalse positives \n19 (13 to 29) \n17 (12 to 27) \n16 (11",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Ends a summary of findings table with expected true and false results per 1,000 people tested at several prevalences, graded high certainty.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p044-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 44,
      "pageEnd": 44,
      "target": "../images/img-p044-1.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals, grouped by brand, for pulmonary TB on sputum with a microbiological reference standard (MRS): 32 Xpert Ultra studies sorted by sensitivity (0.54 to 1.00) and 4 Truenat MTB Plus studies (0.85 to 0.96). Most specificities are above 0.9.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-044",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 44",
      "pageStart": 44,
      "pageEnd": 44,
      "target": "../sections/page-044.md",
      "narrative": null,
      "content": "Appendix 3. Additional forest plots  \nAppendix Figures 1 to 11 summarise currently available data for LC-aNAATs by brand and subgroups. \nTwo studies (Penn-Nicholson 2021 and Theron 2024) compared the accuracy of Xpert Ultra and \nTruenat MTB Plus. Note that for the class level meta-analyses of LC-aNAATs presented in Table 3, to \navoid double counting, we selected the data for Truenat MTB Plus for inclusion in the meta-analyses \nbecause Truenat assays were the index tests evaluated in the studies.          \n \nAppendix Figure 1. Forest plot of LC-aNAATs for pulmonary TB by brand: sputum \nspecimen",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces Appendix 3, additional forest plots of LC-aNAATs by brand and subgroup, and explains how head-to-head studies were handled to avoid double counting.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p045-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 45,
      "pageEnd": 45,
      "target": "../images/img-p045-1.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals, grouped by brand, for pulmonary TB in people living with HIV against a microbiological reference standard: 11 Xpert Ultra studies (sensitivity 0.54 to 1.00) and 2 Truenat MTB Plus studies (0.92 and 0.86).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p045-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 45,
      "pageEnd": 45,
      "target": "../images/img-p045-2.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals, grouped by brand, for pulmonary TB in HIV-negative people against a microbiological reference standard: 4 Xpert Ultra studies (sensitivity 0.88 to 0.91) and 2 Truenat MTB Plus studies (0.97 and 0.88).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-045",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 45",
      "pageStart": 45,
      "pageEnd": 45,
      "target": "../sections/page-045.md",
      "narrative": null,
      "content": "Appendix Figure 2. Forest plot of LC-aNAATs for pulmonary TB by brand: PLHIV, \nsputum specimen and microbiological reference standard \n \n \nAppendix Figure 3. Forest plot of LC-aNAATs for pulmonary TB by brand: HIV \nnegative, sputum specimen and microbiological reference standard",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Captions for forest plots of LC-aNAAT accuracy by brand among people living with HIV and HIV-negative people.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p046-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 46,
      "pageEnd": 46,
      "target": "../images/img-p046-1.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals, grouped by brand, for smear-positive pulmonary TB against a microbiological reference standard: 16 Xpert Ultra and 3 Truenat MTB Plus studies. Sensitivities are 0.96–1.00, while specificities are often not estimable or have very wide intervals because of few true negatives.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p046-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 46,
      "pageEnd": 46,
      "target": "../images/img-p046-2.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals, grouped by brand, for smear-negative pulmonary TB against a microbiological reference standard: 17 Xpert Ultra studies (sensitivity 0.00 in Sabi 2018, otherwise 0.60 to 0.93) and 3 Truenat MTB Plus studies (0.72 to 0.81). Specificities are mostly 0.89–0.99.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-046",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 46",
      "pageStart": 46,
      "pageEnd": 46,
      "target": "../sections/page-046.md",
      "narrative": null,
      "content": "Appendix Figure 4. Forest plot of LC-aNAATs for pulmonary TB by brand: smear \npositive, sputum specimen and microbiological reference standard \nAppendix Figure 5. Forest plot of LC-aNAATs for pulmonary TB by brand: smear \nnegative, sputum specimen and microbiological reference standard",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Captions for forest plots of LC-aNAAT accuracy by brand in smear-positive and smear-negative people.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p047-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 47,
      "pageEnd": 47,
      "target": "../images/img-p047-1.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals, grouped by brand, for pulmonary TB in people with a history of TB against a microbiological reference standard: 11 Xpert Ultra studies with wide confidence intervals and 1 Truenat MTB Plus study (Ngangue 2022, sensitivity 0.96, specificity 0.93).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p047-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 47,
      "pageEnd": 47,
      "target": "../images/img-p047-2.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals, grouped by brand, for pulmonary TB in people with no history of TB against a microbiological reference standard: 9 Xpert Ultra studies (sensitivity 0.89 to 1.00, specificity 0.85 to 1.00) and 1 Truenat MTB Plus study (Ngangue 2022, 0.96 and 0.95).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-047",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 47",
      "pageStart": 47,
      "pageEnd": 47,
      "target": "../sections/page-047.md",
      "narrative": null,
      "content": "Appendix Figure 6. Forest plot of LC-aNAATs for pulmonary TB by brand: history of \nTB, sputum specimen and microbiological reference standard \n \n      \n \nAppendix Figure 7. Forest plot of LC-aNAATs for pulmonary TB by brand: no history \nof TB, sputum specimen and microbiological reference standard",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Captions for forest plots of LC-aNAAT accuracy by brand in people with and without a history of TB.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p048-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 48,
      "pageEnd": 48,
      "target": "../images/img-p048-1.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals, grouped by brand, for detection of rifampicin resistance: 13 Xpert Ultra studies (sensitivity 0.93–1.00 where estimable) and 2 Truenat MTB-RIF Dx studies (0.53 and 0.85). Specificities are 0.97–1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p048-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 48,
      "pageEnd": 48,
      "target": "../images/img-p048-2.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals, for 8 studies of Xpert Ultra (adolescents only, per the caption). Sensitivities range from 0.67 to 1.00 with wide intervals (one not estimable) and specificities from 0.80 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-048",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 48",
      "pageStart": 48,
      "pageEnd": 48,
      "target": "../sections/page-048.md",
      "narrative": null,
      "content": "Appendix Figure 8. Forest plot of LC-aNAATs for rifampicin resistance by brand: \nsputum specimen \n \n     \nAppendix Figure 9. Forest plot of Xpert Ultra: adolescents only \nSee characteristics of these studies in Table 1 and the summary estimates in Table 3. Adolescent only \ndata were not available for Truenat MTB Plus.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Captions for forest plots of rifampicin resistance detection by brand and of Xpert Ultra accuracy in adolescents.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p049-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 49,
      "pageEnd": 49,
      "target": "../images/img-p049-1.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals, grouped by brand, for pulmonary TB against a microbiological reference standard (adults only, per the caption): 22 Xpert Ultra studies (sensitivity 0.54 to 1.00) and 3 Truenat MTB Plus studies (0.85 to 0.96).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p049-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 49,
      "pageEnd": 49,
      "target": "../images/img-p049-2.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals, for detection of rifampicin resistance by LC-aNAATs, in two panels: smear positive and smear negative, each with 6 studies (1 Truenat MTB-RIF Dx, 5 Xpert Ultra). Truenat MTB-RIF Dx (Gomathi 2020) has the lowest sensitivity in both panels (0.57 and 0.17).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-049",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 49",
      "pageStart": 49,
      "pageEnd": 49,
      "target": "../sections/page-049.md",
      "narrative": null,
      "content": "Appendix Figure 10. Forest plot of LC-aNAATs: adult only \n \n \nAppendix Figure 11. Forest plot of LC-aNAATs for rifampicin resistance by smear \nstatus",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Captions for forest plots of LC-aNAAT accuracy in adults and of rifampicin resistance detection by smear status.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-050",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 50",
      "pageStart": 50,
      "pageEnd": 50,
      "target": "../sections/page-050.md",
      "narrative": null,
      "content": "Appendix 4. Analyses of Xpert Ultra results \nAppendix Table 4. Summary of Xpert Ultra trace results \nStudy ID \nTP trace \nresult \n(a) \nFP trace \nresult \n(b) \nNumber of \ntrace results \n(a + b) \nNumber of \nparticipants \n(n) \nPercentage of true \npositive trace \nresults [a/(a + b)] \nPercentage of \ntrace results \n[(a + b)/n] \nAguilera-Alonso \n2022 \n0 \n0 \n0 \n5 \n− \n0.0 \nBerhanu 2018 \n2 \n4 \n6 \n237 \n33.3 \n2.5 \nBoyles 2020 \n1 \n1 \n2 \n207 \n50.0 \n1.0 \nCalderwood 2023 \n13 \n21 \n34 \n896 \n38.2 \n3.8 \nChilukutu 2022 \n3 \n34 \n37 \n739 \n8.1 \n5.0 \nDorman 2018 \n13 \n19 \n32 \n1439 \n40.6 \n2.2 \nLi 2023 \n9 \n13 \n22 \n368 \n40.9",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Appendix 4, a table of Xpert Ultra trace results per study and the share that were true positives, with summary estimates when trace results are treated as negative. The page ends with the title of the next annex, B.2.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-051",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 51",
      "pageStart": 51,
      "pageEnd": 51,
      "target": "../sections/page-051.md",
      "narrative": null,
      "content": "pulmonary and extrapulmonary tuberculosis and rifampicin \nresistance in children: a systematic review of diagnostic accuracy \n \n \nAlexander W Kay1,2,3*, Maia Madison1,2,3,4, Katie Scandrett5, Tara Ness1,2, Pauline Amuge6, \nLeeberk Raja Inbaraj7, Jefferson Daniel8, Mukesh Kumar Sathyanarayanan7, Vignes Anand \nSrinivasalu7, Adhin Bhaskar7, Anna M Mandalakas1,2, Karen R Steingart9, Yemisi Takwoingi5 \n \n1. The Global Tuberculosis Program, Texas Children’s Hospital, Section of Global and \nImmigrant Health, Department of Pediatrics, Baylor College of Medicine, Houston, Texas, \nUSA \n2. Department of",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title, author list and affiliations for Web Annex B.2, a systematic review of the diagnostic accuracy of low complexity automated NAATs for pulmonary and extrapulmonary TB and rifampicin resistance in children.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-052",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 52",
      "pageStart": 52,
      "pageEnd": 52,
      "target": "../sections/page-052.md",
      "narrative": null,
      "content": "BACKGROUND \nTuberculosis is one of the top 10 causes of death and a leading cause from a single infectious agent, \nwith an estimated 1.1 million deaths in 2022 (WHO Global Tuberculosis Report 2023). Globally \nduring that year, an estimated 10.6 million people developed tuberculosis disease, including around \n1.1 million children younger than 15 years and 214,000 children who died of the disease (WHO \nGlobal Tuberculosis Report 2023). Recent models that have been accepted and supported by the \nWorld Health Organization (WHO) suggest that there is substantial underreporting as well as under-\ndia",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives background on the paediatric TB burden, citing 2022 estimates of cases and deaths among children under 15, and the challenges of diagnosing TB in children. It then begins the methods, listing the databases searched in October 2023.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-053",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 53",
      "pageStart": 53,
      "pageEnd": 53,
      "target": "../sections/page-053.md",
      "narrative": null,
      "content": "Index Medicus; and SCOPUS (Elsevier). VL also searched ClinicalTrials.gov (clinicaltrials.gov/) and \nthe WHO Clinical Trials Registry Platform (www.who.int/trialsearch) to identify ongoing trials. The \nreview authors examined the reference lists of included articles and relevant review articles identified \nthrough electronic searches. VL also searched for relevant dissertations in ProQuest Dissertations & \nTheses A&I.  The review authors searched for information on ongoing and unpublished studies from \nmanufacturers through a WHO public call; and from experts working on new diagnostics for TB",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the search methods of the paediatric LC-aNAAT review, including trial registries, reference lists and dissertations, and describes study selection, data extraction and risk-of-bias assessment. It also explains the meta-analysis approach and the use of GRADE to rate certainty separately for sensitivity and specificity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p054-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 54,
      "pageEnd": 54,
      "target": "../images/img-p054-1.webp",
      "narrative": "draft",
      "content": "PRISMA flow diagram for the review update in children. 14 studies came from the previous review; of 431 new database records (234 duplicates removed), 197 were screened, 63 reports assessed and 54 excluded with reasons listed, and 2 records came from organizations. Nine new studies were included, giving 23 studies in total.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-vfig-p054",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 54,
      "pageEnd": 54,
      "target": "../figures/vfig-p054.png",
      "narrative": "draft",
      "content": "PRISMA flow diagram (labelled 'Figure 1. PRISMA Diagram') with columns for previous studies, new studies via databases and registers, and new studies via other methods. 14 studies were included in the previous review; 431 database records, 234 duplicates removed, 197 screened, 134 excluded, 63 reports assessed, with exclusions for listed reasons (e.g. adult study population, n = 15; prior review, n = 12); 2 reports from organisations were assessed with none excluded. 9 new studies were included, for 23 studies in total.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-054",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 54",
      "pageStart": 54,
      "pageEnd": 54,
      "target": "../sections/page-054.md",
      "narrative": null,
      "content": "Study Identification and Characteristics \nFor pulmonary and extrapulmonary tuberculosis detection, we identified 23 studies that evaluated the \ndiagnostic accuracy of Xpert Ultra in sputum, gastric aspirate, nasopharyngeal aspirate, stool or \nextrapulmonary specimens. Table 1 describes the study characteristics, and the PRISMA diagram is \ndisplayed in Figure 1 (Page 2021). \n \nFigure 1. PRISMA Diagram \n \nOf the 23 studies included, 14 studies (61%) took place in high tuberculosis burden and 14 (61%) in \nhigh TB/HIV burden countries, with two studies (9%) performed in multiple settings of varyin",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that 23 studies evaluated Xpert Ultra in children using sputum, gastric aspirate, nasopharyngeal aspirate, stool or extrapulmonary specimens, and summarizes their settings and reference standards.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-055",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 55",
      "pageStart": 55,
      "pageEnd": 55,
      "target": "../sections/page-055.md",
      "narrative": null,
      "content": "Study \nCountries \nHigh TB \nburden/high \nTB/HIV \nburden \nClinical \nsetting \nType of \nspecimen \nStudy design \nPatient \nselection \nNumber \nof \ncultures \nComposi\nte \nreference \nstandard  \nAguilera-\nAlonso 2022 \nSpain \nNo/No \nUndefined \nSputum, gastric \naspirate \nProspective \ncohort \nConsecutive \nUnclear \nYes \nAurilio 2022 \nBrazil \nYes/Yes \nUndefined \nExtrapulmonary \nProspective \ncohort \nConsecutive \nUnclear \nYes \nBabo 2023 \nEthiopia \nYes/Yes \nOutpatient \nStool  \nProspective \ncohort \nConsecutive \nNone \nNo \nBarcellini \n2019 \nItaly \nNo/No \nOutpatient \nSputum \nCross-\nsectional \nConsecutive \nMultiple",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The first page of Table 1, listing key characteristics of included paediatric Xpert Ultra studies, such as country, TB burden, setting, specimen type, design and reference standard.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-056",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 56",
      "pageStart": 56,
      "pageEnd": 56,
      "target": "../sections/page-056.md",
      "narrative": null,
      "content": "Study \nCountries \nHigh TB \nburden/high \nTB/HIV \nburden \nClinical \nsetting \nType of \nspecimen \nStudy design \nPatient \nselection \nNumber \nof \ncultures \nComposi\nte \nreference \nstandard  \nParigi 2021 \nItaly \nNo/No \nInpatient  \nGastric aspirate \n  \nProspective \ncohort \nUnclear \nMultiple \nYes \nPradhan 2022 \nIndia \nYes/No \nInpatient \nCSF \nProspective \ncohort \nConsecutive \nSingle \nYes \nSabi 2018 \nTanzania \nNo/Yes \nInpatient \nand \noutpatient \nSputum \nProspective \ncohort \nConsecutive \nMultiple \nYes \nSabi 2022 \nTanzania, \nGambia, Mali \nVariable \nInpatient \nand \noutpatient \nSputum, gastric \naspirate \nPros",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the table of paediatric study characteristics and summarizes the risk of bias and applicability judgements for studies using sputum specimens.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-057",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 57",
      "pageStart": 57,
      "pageEnd": 57,
      "target": "../sections/page-057.md",
      "narrative": null,
      "content": "and a high risk of bias for one study (1/15, 6.7%) because in this study many children were excluded \nfrom the analysis (Liu 2021). \n \nStudies evaluating Xpert Ultra on gastric aspirate specimens (12 studies) \nIn the patient selection domain, for MRS, we judged the risk of bias as low for most studies (10/12, \n83.3%) and high-risk two studies (16.6%) because the manner of participant selection was a \nretrospective convenience sample and due to unclear inclusion criteria in the other. Regarding \napplicability, we judged low concern for most studies (9/12, 75%), unclear for one study (1/12, 8.3%",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarizes risk-of-bias and applicability judgements for paediatric Xpert Ultra studies using sputum, gastric aspirate and stool specimens, domain by domain.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-058",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 58",
      "pageStart": 58,
      "pageEnd": 58,
      "target": "../sections/page-058.md",
      "narrative": null,
      "content": "high concern for one study (10%) because participants with a definite diagnosis of tuberculosis were \nexcluded (Liu 2021). Regarding applicability, we judged low concern for most studies (8/10, 80%). \nWe judged high concern for one study (10%)  because the patients were exclusively inpatients in a \ntertiary/specialized hospital, and unclear for one study (10%) because the setting was undefined.  \n  \nIn the index test domain, we judged the risk of bias as low for all (100%) because Xpert Ultra test \nresults are automatically generated, and the user is provided with printable test results (thus",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the quality assessment for stool and other specimens, then begins a table of summary diagnostic accuracy of Xpert Ultra in children by specimen type and reference standard. The table rows cover sputum and gastric aspirate against culture or a composite standard.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p059-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 59,
      "pageEnd": 59,
      "target": "../images/img-p059-1.webp",
      "narrative": "draft",
      "content": "Risk-of-bias and applicability chart for 15 studies in children: sputum with a microbiological reference standard. Each study is rated high (red), unclear (yellow) or low (green) for patient selection, index test, reference test and flow and timing, and for applicability.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-vfig-p059",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 59,
      "pageEnd": 59,
      "target": "../figures/vfig-p059.png",
      "narrative": "draft",
      "content": "Fragment of a diagnostic accuracy table (continuation of Table 2 in this review) showing rows 8 to 11: nasopharyngeal aspirate with CRS (4 studies, 676 (465) participants), stool with MRS (10 studies, 2885 (255)), stool with CRS (6 studies, 1778 (510)), and, under 'Detection of rifampicin resistance', sputum specimens (3 studies, 76 (2)) with no estimates. Columns give pooled values with 95% intervals, e.g. stool with MRS 68.0 (50.3 to 81.7) and 98.2 (96.3 to 99.1).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-059",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 59",
      "pageStart": 59,
      "pageEnd": 59,
      "target": "../sections/page-059.md",
      "narrative": null,
      "content": "7 \nNasopharyngeal \naspirate; MRS \n6 \n1353 (94) \n46.2 (34.9 to \n57.9) \n97.5 (95.1 to 98.7) \n67.2 (44.3 to \n83.5) \n94.2 (92.9 to \n95.5) \n8 \nNasopharyngeal \naspirate; CRS \n4 \n676 (465) \n12.7 (09.2 to \n17.3) \n100.0 (98.3 to \n100.0) \n100.0 (37.2 to \n100.0) \n91.2 (90.7 to \n91.6) \n9 \nStool; MRS \n10 \n2885 \n(255)  \n68.0 (50.3 to \n81.7) \n98.2 (96.3 to 99.1) \n80.9 (60.4 to \n91.3) \n96.5 (94.6 to \n98.0) \n10 \nStool; CRS \n6 \n1778 \n(510) \n31.4 (17.2 to \n50.2) \n99.0 (97.4 to 99.6) \n77.3 (42.3 to \n93.4) \n92.8 (91.4 to \n94.7) \nDetection of rifampicin resistance \n11 \nSputum specimens \n3 \n76 (2) \n- \n- \n- \n- \n* Pre",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the paediatric accuracy table for nasopharyngeal aspirate and stool, then begins reporting results for sputum against culture, with eleven studies meta-analysed.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p060-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 60,
      "pageEnd": 60,
      "target": "../images/img-p060-1.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals for 15 studies in children: sputum with a microbiological reference standard. Sensitivity ranges from 0.57 to 0.91 where estimable and specificity from 0.82 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-060",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 60",
      "pageStart": 60,
      "pageEnd": 60,
      "target": "../sections/page-060.md",
      "narrative": null,
      "content": "(68.9 to 80.8) and the summary specificity (95% CI) was 95.9% (92.3 to 97.9) (high-certainty \n(sensitivity and specificity) evidence).  \nFor a population of 1,000 children where 100 have tuberculosis on culture, 112 would be Xpert Ultra \npositive; of these, 75 would have pulmonary tuberculosis (true positives) and 37 would not have \npulmonary tuberculosis (false positives); 888 would be Xpert Ultra negative; of these 863 would not \nhave pulmonary tuberculosis (true negatives) and 25 would have pulmonary tuberculosis (false-\nnegatives). \n \nFigure 3. LC-aNAATs for pulmonary tuberculosis using sp",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that Xpert Ultra on children's sputum had a summary sensitivity of about 75% and specificity of about 96% against culture, with high-certainty evidence and illustrative figures per 1,000 children.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p061-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 61,
      "pageEnd": 61,
      "target": "../images/img-p061-1.webp",
      "narrative": "draft",
      "content": "Risk-of-bias and applicability chart for 11 studies in children: sputum with a composite reference standard. Each study is rated high (red), unclear (yellow) or low (green) for patient selection, index test, reference test and flow and timing, and for applicability.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p061-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 61,
      "pageEnd": 61,
      "target": "../images/img-p061-2.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals for 11 studies in children: sputum with a composite reference standard. Sensitivity is low, mostly 0.17–0.27 (highest 0.63, Yenew 2023), while specificity is 1.00 in all studies.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-061",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 61",
      "pageStart": 61,
      "pageEnd": 61,
      "target": "../sections/page-061.md",
      "narrative": null,
      "content": "Figure 4. Risk of bias and applicability summary for pulmonary tuberculosis: sputum specimen and \ncomposite reference standard  \n \nEleven studies (2,824 participants) evaluating sputum against a CRS were included with sensitivities \nbetween 17% and 63% and specificities all 100% (Figure 5). Ten studies (2,823 participants) were \nincluded in the meta-analysis. The summary sensitivity (95% CI) was 23.5% (20.3 to 27.0) and the \nsummary specificity (95% CI) was 100.0% (99.7 to 100.0) (moderate-certainty (sensitivity and \nspecificity) evidence). \nFor a population of 1,000 children where 100 have tu",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports a much lower summary sensitivity of about 24% for Xpert Ultra on children's sputum against a composite reference standard, with specificity of 100%.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p062-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 62,
      "pageEnd": 62,
      "target": "../images/img-p062-1.webp",
      "narrative": "draft",
      "content": "Risk-of-bias and applicability chart for 11 studies in children: sputum in children living with HIV, microbiological reference standard. Each study is rated high (red), unclear (yellow) or low (green) for patient selection, index test, reference test and flow and timing, and for applicability.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-062",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 62",
      "pageStart": 62,
      "pageEnd": 62,
      "target": "../sections/page-062.md",
      "narrative": null,
      "content": "The studies are sorted on the plot by decreasing sensitivity. FN: false negative; FP: false positive; TN: \ntrue negative; TP: true positive. \n \n3. Should LC-aNAATs on respiratory samples be used to diagnose \npulmonary tuberculosis in children living with HIV (CLHIV) with signs \nand symptoms or screened positive for pulmonary tuberculosis, against a \nmicrobiological reference standard? \n \nFigure 6. Risk of bias and applicability summary for pulmonary tuberculosis in CLHIV: sputum \nspecimen and microbiological reference standard  \n \nEleven studies (445 participants) evaluating sputum against a M",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces the analysis of Xpert Ultra on sputum in children living with HIV against culture, with illustrative numbers of true and false results per 1,000 children.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p063-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 63,
      "pageEnd": 63,
      "target": "../images/img-p063-1.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals for 11 studies in children: sputum in children living with HIV, microbiological reference standard. Sensitivity ranges from 0.40 to 1.00 where estimable and specificity from 0.77 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p063-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 63,
      "pageEnd": 63,
      "target": "../images/img-p063-2.webp",
      "narrative": "draft",
      "content": "Risk-of-bias and applicability chart for 7 studies in children: sputum in children living with HIV, composite reference standard. Each study is rated high (red), unclear (yellow) or low (green) for patient selection, index test, reference test and flow and timing, and for applicability.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-063",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 63",
      "pageStart": 63,
      "pageEnd": 63,
      "target": "../sections/page-063.md",
      "narrative": null,
      "content": "Figure 7. LC-aNAATs for pulmonary tuberculosis in CLHIV using sputum and a microbiological \nreference standard \nThe studies are sorted on the plot by decreasing sensitivity. FN: false negative; FP: false positive; TN: \ntrue negative; TP: true positive. \n \n4. Should LC-aNAATs on respiratory samples be used to diagnose \npulmonary tuberculosis in CLHIV with signs and symptoms or screened \npositive for pulmonary tuberculosis, against a composite reference \nstandard? \n \nFigure 8. Risk of bias and applicability summary for pulmonary tuberculosis in CLHIV: sputum \nspecimen and composite reference sta",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that, against a composite reference standard, Xpert Ultra on sputum from children living with HIV had a summary sensitivity of about 19% and specificity of 100%.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p064-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 64,
      "pageEnd": 64,
      "target": "../images/img-p064-1.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals for 7 studies in children: sputum in children living with HIV, composite reference standard. Sensitivity ranges from 0.00 to 0.25, while specificity is 1.00 in all studies.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p064-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 64,
      "pageEnd": 64,
      "target": "../images/img-p064-2.webp",
      "narrative": "draft",
      "content": "Risk-of-bias and applicability chart for 12 studies in children: gastric aspirate with a microbiological reference standard. Each study is rated high (red), unclear (yellow) or low (green) for patient selection, index test, reference test and flow and timing, and for applicability.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-064",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 64",
      "pageStart": 64,
      "pageEnd": 64,
      "target": "../sections/page-064.md",
      "narrative": null,
      "content": "For a population of 1,000 CLHIV where 400 have tuberculosis on CRS, 74 would be Xpert Ultra \npositive; of these, 74 would have pulmonary tuberculosis (true positives) and 0 would not have \npulmonary tuberculosis (false positives); 926 would be Xpert Ultra negative; of these 600 would not \nhave pulmonary tuberculosis (true negatives) and 326 would have pulmonary tuberculosis (false-\nnegatives). \n \nFigure 9. LC-aNAATs for pulmonary tuberculosis in CLHIV using sputum and a composite \nreference standard \nThe studies are sorted on the plot by decreasing sensitivity. FN: false negative; FP: false po",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives illustrative results per 1,000 children living with HIV against the composite standard and introduces the analysis of gastric aspirate specimens.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p065-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 65,
      "pageEnd": 65,
      "target": "../images/img-p065-1.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals for 12 studies in children: gastric aspirate with a microbiological reference standard. Sensitivity ranges from 0.00 to 1.00 where estimable and specificity from 0.67 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-065",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 65",
      "pageStart": 65,
      "pageEnd": 65,
      "target": "../sections/page-065.md",
      "narrative": null,
      "content": "microbiological reference standard \nTwelve studies (1959 participants) evaluating gastric aspirate against a MRS were included with \nsensitivities between 0% and 100% and specificities between 67% and 100% (Figure 11). All 12 \nstudies were included in the meta-analysis. The summary sensitivity (95% CI) was 69.6% (60.3 to \n77.6) and the summary specificity (95% CI) was 91.0% (82.5 to 95.6) (moderate-certainty (sensitivity \nand specificity) evidence). \nFor a population of 1,000 children where 100 have tuberculosis on culture, 151 would be Xpert Ultra \npositive; of these, 70 would have pulmonary",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports a summary sensitivity of about 70% and specificity of about 91% for Xpert Ultra on gastric aspirate against culture in children, with moderate-certainty evidence.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p066-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 66,
      "pageEnd": 66,
      "target": "../images/img-p066-1.webp",
      "narrative": "draft",
      "content": "Risk-of-bias and applicability chart for 7 studies in children: gastric aspirate with a composite reference standard. Each study is rated high (red), unclear (yellow) or low (green) for patient selection, index test, reference test and flow and timing, and for applicability.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p066-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 66,
      "pageEnd": 66,
      "target": "../images/img-p066-2.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals for 7 studies in children: gastric aspirate with a composite reference standard. Sensitivity ranges from 0.19 to 0.83 where estimable and specificity from 0.91 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-066",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 66",
      "pageStart": 66,
      "pageEnd": 66,
      "target": "../sections/page-066.md",
      "narrative": null,
      "content": "Figure 12. Risk of bias and applicability summary for pulmonary tuberculosis: gastric aspirate and \ncomposite reference standard \nSeven studies (1,057 participants) evaluating gastric aspirate against a CRS were included with \nsensitivities between 19% and 83% and specificities 91% to 100% (Figure 13). All seven studies were \nincluded in the meta-analysis. The summary sensitivity (95% CI) was 48.8% (31.4 to 63.7) and the \nsummary specificity (95% CI) was 98.7% (97.7 to 99.3) (low-certainty (sensitivity) and high-certainty \n(specificity) evidence). \nFor a population of 1,000 children where 100",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports a summary sensitivity of about 49% and specificity of about 99% for Xpert Ultra on gastric aspirate against a composite reference standard in children.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p067-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 67,
      "pageEnd": 67,
      "target": "../images/img-p067-1.webp",
      "narrative": "draft",
      "content": "Risk-of-bias and applicability chart for 7 studies in children: nasopharyngeal aspirate with a microbiological reference standard. Each study is rated high (red), unclear (yellow) or low (green) for patient selection, index test, reference test and flow and timing, and for applicability.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p067-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 67,
      "pageEnd": 67,
      "target": "../images/img-p067-2.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals for 7 studies in children: nasopharyngeal aspirate with a microbiological reference standard. Sensitivity ranges from 0.33 to 0.67 where estimable and specificity from 0.50 to 0.99.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-067",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 67",
      "pageStart": 67,
      "pageEnd": 67,
      "target": "../sections/page-067.md",
      "narrative": null,
      "content": "screened positive for pulmonary tuberculosis, against a microbiological \nreference standard? \n \n \nFigure 14. Risk of bias and applicability summary for pulmonary tuberculosis: nasopharyngeal \naspirate and microbiological reference standard \n \nSeven studies (1,355 participants) evaluating nasopharyngeal aspirate against a MRS were included \nwith sensitivities between 33% and 67% and specificities between 50% and 99% (Figure 15). Six \nstudies (1,353) were included in a meta-analysis. The summary sensitivity (95% CI) was 46.2% (34.9 \nto 57.9) and the summary specificity (95% CI) was 97.5% (95.1 t",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that Xpert Ultra on nasopharyngeal aspirate had a summary sensitivity of about 46% against culture in children, with illustrative numbers per 1,000 tested.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p068-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 68,
      "pageEnd": 68,
      "target": "../images/img-p068-1.webp",
      "narrative": "draft",
      "content": "Risk-of-bias and applicability chart for 4 studies in children: nasopharyngeal aspirate with a composite reference standard. Each study is rated high (red), unclear (yellow) or low (green) for patient selection, index test, reference test and flow and timing, and for applicability.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p068-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 68,
      "pageEnd": 68,
      "target": "../images/img-p068-2.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals for 4 studies in children: nasopharyngeal aspirate with a composite reference standard. Sensitivity ranges from 0.09 to 0.36, while specificity is 1.00 in all studies.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-068",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 68",
      "pageStart": 68,
      "pageEnd": 68,
      "target": "../sections/page-068.md",
      "narrative": null,
      "content": "8. Should LC-aNAATs on nasopharyngeal aspirate specimens be used to \ndiagnose pulmonary tuberculosis in children with signs and symptoms or \nscreened positive for pulmonary tuberculosis, against a composite \nreference standard? \n \nFigure 16. Risk of bias and applicability summary for pulmonary tuberculosis: nasopharyngeal \naspirate and composite reference standard \n \nFour studies (676 participants) evaluating nasopharyngeal aspirate against a CRS were included with \nsensitivities between 9% and 36% and specificities of 100% (Figure 17). All four studies were \nincluded in the meta-analysis. The",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports low sensitivity for Xpert Ultra on nasopharyngeal aspirate against a composite reference standard and introduces the analysis of stool specimens against culture.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p069-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 69,
      "pageEnd": 69,
      "target": "../images/img-p069-1.webp",
      "narrative": "draft",
      "content": "Risk-of-bias and applicability chart for 10 studies in children: stool with a microbiological reference standard. Each study is rated high (red), unclear (yellow) or low (green) for patient selection, index test, reference test and flow and timing, and for applicability. The column headings are printed upside down in the image.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-069",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 69",
      "pageStart": 69,
      "pageEnd": 69,
      "target": "../sections/page-069.md",
      "narrative": null,
      "content": "Figure 18. Risk of bias and applicability summary for pulmonary tuberculosis: stool and \nmicrobiological reference standard \n \nTen studies (2885 participants) evaluating stool against a MRS of culture or Xpert Ultra on sputum or \ngastric aspirate were included with sensitivities between 26% and 100% and specificities between \n89% and 100% (Figure 19). All ten studies were included in the meta-analysis. The summary \nsensitivity (95% CI) was 68.0% (50.3 to 81.7) and the summary specificity (95% CI) was 98.2% (96.3 \nto 99.1) (moderate-certainty (sensitivity) and high-certainty (specificity) evide",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports a summary sensitivity of about 68% and specificity of about 98% for Xpert Ultra on stool in children against a microbiological reference standard.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p070-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 70,
      "pageEnd": 70,
      "target": "../images/img-p070-1.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals for 10 studies in children: stool with a microbiological reference standard. Sensitivity ranges from 0.26 to 1.00 where estimable and specificity from 0.89 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p070-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 70,
      "pageEnd": 70,
      "target": "../images/img-p070-2.webp",
      "narrative": "draft",
      "content": "Risk-of-bias and applicability chart for 6 studies in children: stool with a composite reference standard. Each study is rated high (red), unclear (yellow) or low (green) for patient selection, index test, reference test and flow and timing, and for applicability.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-070",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 70",
      "pageStart": 70,
      "pageEnd": 70,
      "target": "../sections/page-070.md",
      "narrative": null,
      "content": "Figure 19. Xpert Ultra for pulmonary tuberculosis using stool and a microbiological reference \nstandard. The studies are sorted on the plot by decreasing sensitivity. FN: false negative; FP: false \npositive; TN: true negative; TP: true positive. \n10. Should LC-aNAATs on stool be used to diagnose pulmonary \ntuberculosis in children with signs and symptoms or screened positive for \npulmonary tuberculosis, against a composite reference standard? \n \n \n \nFigure 20. Risk of bias and applicability summary for pulmonary tuberculosis: stool and composite \nreference standard  \nSix studies (1778 particip",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces the analysis of Xpert Ultra on stool against a composite reference standard, with illustrative numbers of true and false results per 1,000 children.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p071-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 71,
      "pageEnd": 71,
      "target": "../images/img-p071-1.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals for 6 studies in children: stool with a composite reference standard. Sensitivity ranges from 0.09 to 0.58 where estimable and specificity from 0.97 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p071-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 71,
      "pageEnd": 71,
      "target": "../images/img-p071-2.webp",
      "narrative": "draft",
      "content": "Risk-of-bias and applicability chart for 3 studies in children: detection of rifampicin resistance. Each study is rated high (red), unclear (yellow) or low (green) for patient selection, index test, reference test and flow and timing, and for applicability.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-071",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 71",
      "pageStart": 71,
      "pageEnd": 71,
      "target": "../sections/page-071.md",
      "narrative": null,
      "content": "Figure 21. Xpert Ultra for pulmonary tuberculosis using stool and a composite reference standard. \nThe studies are sorted on the plot by decreasing sensitivity. FN: false negative; FP: false positive; TN: \ntrue negative; TP: true positive. \n \nSubgroup analyses \nAppendix 3 summarizes the sensitivity and specificity of the studies included for different specimen \ntypes by HIV status (living with HIV and HIV negative), age group and nutritional status. Appendix 4 \nsummarizes the results of sensitivity analyses where we performed separate analyses excluding \nstudies without multiple cultures for a",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes subgroup analyses by HIV status, age and nutritional status and reports sparse data on Xpert Ultra detection of rifampicin resistance in children's sputum.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p072-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 72,
      "pageEnd": 72,
      "target": "../images/img-p072-1.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals for 3 studies in children: detection of rifampicin resistance. Sensitivity is not estimable in two studies and 1.00 [0.16, 1.00] in Zar 2018-2023; specificity is 1.00 in all three.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p072-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 72,
      "pageEnd": 72,
      "target": "../images/img-p072-2.webp",
      "narrative": "draft",
      "content": "Risk-of-bias and applicability chart for 3 studies in children: cerebrospinal fluid (CSF) with a microbiological reference standard. Each study is rated high (red), unclear (yellow) or low (green) for patient selection, index test, reference test and flow and timing, and for applicability.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p072-3",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 72,
      "pageEnd": 72,
      "target": "../images/img-p072-3.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals for 3 studies in children: cerebrospinal fluid (CSF) with a microbiological reference standard. Sensitivity ranges from 0.67 to 1.00 where estimable and specificity from 0.91 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-072",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 72",
      "pageStart": 72,
      "pageEnd": 72,
      "target": "../sections/page-072.md",
      "narrative": null,
      "content": "Figure 23. Xpert Ultra for detection of rifampicin resistance. FN: false negative; FP: false positive; \nTN: true negative; TP: true positive. \n  \nDetection of extrapulmonary tuberculosis \n12. Should LC-aNAATs on cerebrospinal fluid (CSF) be used to diagnose \ntuberculous meningitis in children with signs and symptoms of \ntuberculous meningitis, against a microbiological reference standard? \n \n \nFigure 24. Risk of bias and applicability summary for extrapulmonary tuberculosis using CSF and a \nmicrobiological reference standard  \nThree studies (215 participants) were included with sensitivities o",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Begins the results on extrapulmonary TB in children, reporting on Xpert Ultra accuracy in cerebrospinal fluid for tuberculous meningitis against culture.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p073-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 73,
      "pageEnd": 73,
      "target": "../images/img-p073-1.webp",
      "narrative": "draft",
      "content": "Risk-of-bias and applicability chart for 2 studies in children: cerebrospinal fluid (CSF) with a composite reference standard. Each study is rated high (red), unclear (yellow) or low (green) for patient selection, index test, reference test and flow and timing, and for applicability.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p073-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 73,
      "pageEnd": 73,
      "target": "../images/img-p073-2.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals for 2 studies in children: cerebrospinal fluid (CSF) with a composite reference standard. Sensitivity ranges from 0.12 to 0.19, while specificity is 1.00 in all studies.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-073",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 73",
      "pageStart": 73,
      "pageEnd": 73,
      "target": "../sections/page-073.md",
      "narrative": null,
      "content": "13. Should LC-aNAATs on CSF be used to diagnose tuberculous \nmeningitis in children with signs and symptoms of tuberculous meningitis, \nagainst a composite reference standard? \n \n \nFigure 26. Risk of bias and applicability summary for extrapulmonary tuberculosis using CSF and a \ncomposite reference standard  \nTwo studies (67 participants) were included with sensitivities of 12% to 19% and all had a specificity \nof 100% (Figure 27). Meta-analysis was not performed due to the small number of studies. \nFor a population of 1,000 children where 200 have tuberculosis on CRS, 24 to 46 would be Xpert",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports two small studies of Xpert Ultra on cerebrospinal fluid against a composite reference standard, with low sensitivity, and introduces the lymph node analysis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p074-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 74,
      "pageEnd": 74,
      "target": "../images/img-p074-1.webp",
      "narrative": "draft",
      "content": "Risk-of-bias and applicability chart for 2 studies in children: lymph node specimens with a microbiological reference standard. Each study is rated high (red), unclear (yellow) or low (green) for patient selection, index test, reference test and flow and timing, and for applicability.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p074-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 74,
      "pageEnd": 74,
      "target": "../images/img-p074-2.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals for 2 studies in children: lymph node specimens with a microbiological reference standard. Both studies show sensitivity 1.00; specificity is 0.82 (Zar 2018-2023) and 0.33 (Aurilio 2022).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-074",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 74",
      "pageStart": 74,
      "pageEnd": 74,
      "target": "../sections/page-074.md",
      "narrative": null,
      "content": "Figure 28. Risk of bias and applicability summary for extrapulmonary tuberculosis using lymph node \ntissue or aspirates and a microbiological reference standard  \nTwo studies (58 participants) were included all with a sensitivity of 100% and specificities between \n33% and 82% (Figure 29). Meta-analysis was not performed due to the small number of cases and \nstudies. \nFor a population of 1,000 children where 200 have tuberculosis on culture, 342 to 736 would be \nXpert Ultra positive; of these, 198 to 200  would have lymph node tuberculosis (true positives) and \n144 to 536 would not have lymph n",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports two small studies of Xpert Ultra on lymph node tissue and aspirates against culture, with high sensitivity and variable specificity; no meta-analysis was performed.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p075-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 75,
      "pageEnd": 75,
      "target": "../images/img-p075-2.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals for 2 studies in children: lymph node specimens with a composite reference standard. Sensitivity is 0.80 (Aurilio 2022) and 0.50 (Zar 2018-2023); specificity is not estimable in one and 1.00 in the other.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-075",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 75",
      "pageStart": 75,
      "pageEnd": 75,
      "target": "../sections/page-075.md",
      "narrative": null,
      "content": "Figure 30. Risk of bias and applicability summary for extrapulmonary tuberculosis using lymph node \ntissue and aspirates and a composite reference standard  \nTwo studies (58 participants) were included with sensitivities of 50% to 805 and a sensitivity of 100% \n(Figure 31). Meta-analysis was not performed due to the small number of studies and participants. \nFor a population of 1,000 children where 200 have tuberculosis on CRS, 100 to 168 would be Xpert \nUltra positive; of these, 100 to 160  would have lymph node tuberculosis (true positives) and 0 to 8 \nwould not have lymph node tuberculosis",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports small studies of Xpert Ultra on lymph node specimens against a composite reference standard and on other extrapulmonary specimens, too few for meta-analysis. The text contains a typographical error in the sensitivity range.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p076-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 76,
      "pageEnd": 76,
      "target": "../images/img-p076-1.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity with 95% confidence intervals for Xpert Ultra for extrapulmonary TB in children aged 0–9 years against culture, in four panels: pleural (3 studies), bone and joint (3), urine (1) and tissue/biopsy (1). Few cases are present and most estimates are not estimable or have very wide intervals.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-076",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 76",
      "pageStart": 76,
      "pageEnd": 76,
      "target": "../sections/page-076.md",
      "narrative": null,
      "content": "Figure 32. LC-aNAATs for extrapulmonary tuberculosis using a microbiological reference standard. \nEPTB: extrapulmonary tuberculosis; FN: false negative; FP: false positive; TN: true negative; TP: \ntrue positive. \nSUMMARY OF MAIN RESULTS \n  \n● For sputum, sensitivity was 75.3% (95% CI 68.9 to 80.8) and specificity was 95.9% (95% CI \n92.3 to 97.9) in children 0 to 9 years, against MRS. \n● For sputum, sensitivity was 70.4% (95% CI 56.0 to 81.7) and specificity was 95.6%  (95% CI \n89.9 to 98.2) in children 0 to 9 years living with HIV, against MRS. \n● For gastric aspirate, sensitivity was 69.6% (9",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarizes the main results of the paediatric review, giving Xpert Ultra sensitivity and specificity by specimen type and reference standard. It notes that sensitivity was similar for sputum, gastric aspirate and stool.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-077",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 77",
      "pageStart": 77,
      "pageEnd": 77,
      "target": "../sections/page-077.md",
      "narrative": null,
      "content": "formal comparisons were performed, and it is difficult to draw conclusions about comparative \naccuracy. The Xpert Ultra specificity for each specimen type was greater than 90%.  \nACKNOWLEDGEMENTS \nWe would like to thank the authorship team of the Cochrane Review that accompanies this update: \nSabine Verkuijl, Kerri Viney, Tiziana Mazini, Annemieke BRANDS, Lucia González Fernández, \nMichael Eisenhut, Karen Steingart, and Anna Mandalakas. We are grateful to Dr Sayera Banu, \nProgramme on Emerging Infections, Infectious Diseases Division, International Centre for Diarrhoeal \nDisease Research, Bang",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Concludes that Xpert Ultra specificity exceeded 90% for each specimen type, though formal comparisons were not made, then gives acknowledgements and begins the list of included study references.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-078",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 78",
      "pageStart": 78,
      "pageEnd": 78,
      "target": "../sections/page-078.md",
      "narrative": null,
      "content": "Enimil 2022 \nEnimil, A. K.; Nuttall, J. J. C.; Centner, C. M.; Beylis, N.; Eley, B. S.. Xpert MTB/RIF Ultra and \nmycobacterial culture in routine clinical care at a paediatric hospital. Southern African Journal of \nInfectious Diseases 20 June 2022;2313-1810. [DOI: https://doi.org/10.4102/sajid.v37i1.398] \n \nJaganath 2021   \nJaganath D, Wambi P, Reza TF, Nakafeero J, Aben EO, Kiconco E, et al. A prospective evaluation of \nXpert MTB/RIF Ultra for childhood pulmonary tuberculosis in Uganda. Journal of the Pediatric \nInfectious Diseases Society 2021;piaa159. \n  \nKabir 2020   \nKabir S, Rahman SM, A",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the references of included paediatric Xpert Ultra studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-079",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 79",
      "pageStart": 79,
      "pageEnd": 79,
      "target": "../sections/page-079.md",
      "narrative": null,
      "content": "MTB/RIF Ultra for the diagnosis of pulmonary tuberculosis in children. Paediatric Infectious Disease \nJournal 2018;37(10):e261-3. [CRS ID: 14396201] \n \nParigi 2021   \nParigi S, Venturini E, Galli L, Chiappini E. Xpert MTB/RIF Ultra performance in diagnosing \npaediatric pulmonary TB in gastric aspirates. International Journal of Tuberculosis and Lung Disease \n2021;25(1):75-7. \n \nPradhan 2022 \nPradhan, N. N.; Paradkar, M. S.; Kagal, A.; Valvi, C.; Kinikar, A.; Khwaja, S.; Dhage, R.; Chandane, \nJ.; Ithape, M.; Bendre, M.; Madewar, R.; Nadgeri, V.; Nijampurkar, A.; Jain, D.; Gupte, N.; Gupta, \nA.;",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Ends the included study references and begins the list of excluded studies, each with its reason for exclusion.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-080",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 80",
      "pageStart": 80,
      "pageEnd": 80,
      "target": "../sections/page-080.md",
      "narrative": null,
      "content": "10.1016/j.ijid.2021.11.010; Unable to disaggregate data to children 0-9 years of age \n \nChiCTR1800015075 \nDiagnositc accuracy of Xpert MTB/RIF Ultra assay on diagnosing pediatric pulmonary tuberculosis \n2018;():2018; Data is unpublished and incomplete \n \nChristopher 2021 \nChristopher, D. J.; Coelho, V.; Ebby, G. S.; Shankar, D.; Gupta, R.; Thangakunam, B. Incremental \nyield of Xpert® MTB/RIF Ultra over Xpert® MTB/RIF in the diagnosis of extrapulmonary TB. \nInternational Journal of Tuberculosis and Lung Disease 2021;25(11):939-944; 2021; DOI: \n10.5588/ijtld.21.0280; Adult study population \n \nCo",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the list of excluded studies, citing reasons such as adult populations, inability to separate data for children under 10, or unpublished data.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-081",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 81",
      "pageStart": 81,
      "pageEnd": 81,
      "target": "../sections/page-081.md",
      "narrative": null,
      "content": "Gutierrez, N.; Cherif, G. F.; Williams, C. M.; Wisniewska, A.; Barer, M. R.; Rigouts, L.; de Jong, B. \nC. Annals of Clinical Microbiology and Antimicrobials 2023;22(1) (no pagination)(81): 2023; DOI: \n10.1186/s12941-023-00633-8; Adult study population \n \nKagujje 2023 \nKagujje, M.; Nyangu, S.; Maimbolwa, M. M.; Shuma, B.; Mutti, L.; Somwe, P.; Sanjase, N.; Chungu, \nC.; Kerkhoff, A. D.; Muyoyeta, M. Face mask sampling (FMS) for tuberculosis shows lower \ndiagnostic sensitivity than sputum sampling in Guinea. PLoS ONE 2023;18(7 July) (no \npagination)(e0288643): 2023; DOI: 10.1371/journal.pone.0288",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the excluded studies list with reasons, including registered trials whose data were unpublished or incomplete.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-082",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 82",
      "pageStart": 82,
      "pageEnd": 82,
      "target": "../sections/page-082.md",
      "narrative": null,
      "content": "NCT03831906. TB-Speed Pneumonia. clinicaltrials.gov/show/NCT03831906 (first received 6 \nFebruary 2019); data were unpublished and incomplete \n \nNCT04038632   \nNCT04038632. TB-Speed Decentralisation Study. clinicaltrials.gov/show/NCT04038632 (first \nreceived 2 August 2019); data were unpublished and incomplete. \n \nNCT05818059 \nhttps://clinicaltrials.gov/show/NCT05818059 2023;(): 2023; Not a diagnostic accuracy study \n \nNicol 2020   \nNicol M, Schumacher S, Workman L, Broger T, Baard C, Prins M, et al. Accuracy of a novel urine \ntest, Fujifilm SILVAMP TB LAM, for the diagnosis of pulmonary tuberc",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the excluded studies list, including TB-Speed trials and studies not designed to assess diagnostic accuracy.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-083",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 83",
      "pageStart": 83,
      "pageEnd": 83,
      "target": "../sections/page-083.md",
      "narrative": null,
      "content": "Quan 2023 \nQuan, S.; Zou, T.; Duan, L.; Tian, X.; Wang, Y.; Zhu, Y.; Fang, M.; Shi, Y.; Wan, C.; Sun, L.; Shen, \nA.; Clinical Characteristics of Pulmonary Tuberculosis in Children Tested by Xpert MTB/RIF Ultra. \nPediatr Infect Dis J ;42(5):389-395; DOI: 10.1097/INF.0000000000003866; Unable to disaggregate \ndata to children 0-9 years of age \n \nRekart 2023 \nRekart, M. L.; Aung, A.; Cullip, T.; Mulanda, W.; Mun, L.; Pirmahmadzoda, B.; Kliescokova, J.; \nAchar, J.; Alvarez, J. L.; Sitali, N.; Sinha, A. Household drug-resistant TB contact tracing in \nTajikistan. Int J Tuberc Lung Dis ;27(10):748-753",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the excluded studies list, mostly studies that could not separate data for young children or enrolled adults.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-084",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 84",
      "pageStart": 84,
      "pageEnd": 84,
      "target": "../sections/page-084.md",
      "narrative": null,
      "content": "Sun 2022 \nSun, L.; Liu, Y.; Fang, M.; Chen, Y.; Zhu, Y.; Xia, C.; Jia, J.; Quan, S.; Wang, Y.; Tian, X.; Shi, Y.; \nDuan, L.; Shi, X.; Liao, Q.; Wan, C.; Shen, A.; Use of Xpert MTB/RIF Ultra assay on stool and \ngastric aspirate samples to diagnose pulmonary tuberculosis in children in a high-tuberculosis-burden \nbut resource-limited area of China. Int J Infect Dis 2022;114():236-243; 2022; DOI: \n10.1016/j.ijid.2021.11.012; Unable to disaggregate data to children 0-9 years of age \n \nSun 2019 \nSun, L.; Qi, X.; Liu, F.; Wu, X.; Yin, Q.; Guo, Y.; Xu, B.; Jiao, A.; Jiao, W.; Shen, C.; Xiao, J.; Shen",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the excluded studies list with reasons for each exclusion.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-085",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 85",
      "pageStart": 85,
      "pageEnd": 85,
      "target": "../sections/page-085.md",
      "narrative": null,
      "content": "Yao 2023 \nYao, L.; Chen, S.; Sha, W.; Gu, Y. The diagnostic performance of endobronchial ultrasound with \nXpert MTB/RIF Ultra in smear-negative pulmonary tuberculosis. BMC Infectious Diseases \n2023;23(1) (no pagination)(107): 2023; DOI: 10.1186/s12879-023-08073-7; Adult study population \n \nYenew 2023 \nYenew, B.; de Haas, P.; Diriba, G.; Kebede, A.; Sherefdin, B.; Demissie, Y.; Bedru, A.; Sahile, M.; \nMengesha, E.; Dememew, Z. G.; Tegegn, B.; Slyzkyi, A.; Amare, M.; Getahun, M.; Abdella, S.; \nJerene, D.; Tiemersma, E. Optimization of the Simple One-Step Stool Processing Method to Diagnose \nTube",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Ends the excluded studies list and begins the other references, citing GRADE guidance on certainty of test accuracy evidence.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-086",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 86",
      "pageStart": 86,
      "pageEnd": 86,
      "target": "../sections/page-086.md",
      "narrative": null,
      "content": "Schünemann HJ, Mustafa R, Brozek J, Steingart KR, Leeflang M, Murad MH, et al. GRADE \nguidelines: 21 part 2. Inconsistency, Imprecision, publication bias and other domains for rating the \ncertainty of evidence for test accuracy and presenting it in evidence profiles and summary of findings \ntables. Journal of Clinical Epidemiology 2020;122:142-52. \n  \nWHO Consolidated Guidelines (Module 3) 2020   \nWorld Health Organization. WHO consolidated guidelines on tuberculosis. Module 3: diagnosis – \nrapid diagnostics for tuberculosis detection. Licence: CC BY-NC-SA 3.0 IGO. \nwho.int/publications/i/item",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Ends the other references, citing GRADE guidance and WHO consolidated guidelines and handbooks on TB diagnosis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-087",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 87",
      "pageStart": 87,
      "pageEnd": 87,
      "target": "../sections/page-087.md",
      "narrative": null,
      "content": "APPENDIX  \n \nAppendix 1: Detailed search strategies \n \nOvid MEDLINE(R) ALL <1946 to October 03, 2023> \n1 \nMycobacterium tuberculosis/ \n \n2 \nTuberculosis, Pulmonary/ \n \n3 \nExtensively Drug-Resistant Tuberculosis/ \n \n4 \nTuberculosis, Multidrug-Resistant/ \n \n5 \n(tuberculosis or TB).tw.  \n6 \n((extrapulmonary or lymph node* or mening* or pulmonary) and TB).ti. or ((extrapulmonary \nor lymph node* or mening* or pulmonary) and TB).ab.  \n7 \nexp Tuberculosis, Extrapulmonary/ \n \n8 \n5 and 6  \n9 \n1 or 2 or 3 or 4 or 7 or 8  \n10 \n(child or children* or childhood or infan* or newborn or neonat* or toddler* o",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Appendix 1, the MEDLINE search strategy for the paediatric LC-aNAAT review, combining TB, child and test name terms.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-088",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 88",
      "pageStart": 88,
      "pageEnd": 88,
      "target": "../sections/page-088.md",
      "narrative": null,
      "content": "1 \nMycobacterium tuberculosis/ \n \n2 \nLung Tuberculosis/ \n \n3 \nextensively drug resistant tuberculosis/  \n4 \ndrug resistant tuberculosis/ \n \n5 \n(tuberculosis or TB).tw.  \n6 \n((extrapulmonary or lymph node* or mening* or pulmonary) and TB).ti. or ((extrapulmonary \nor lymph node* or mening* or pulmonary) and TB).ab.  \n7 \nexp extrapulmonary tuberculosis/ \n \n8 \n5 and 6  \n9 \n1 or 2 or 3 or 4 or 7 or 8  \n10 \n(child or children* or childhood or infan* or newborn or neonat* or toddler* or \nadolescen*).mp. \n \n11 \nexp Child/ or exp Infant/ or Adolescent/ or exp Pediatrics/ \n \n12 \n10 or 11 \n \n13 \n9 and 12",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives the Embase and Web of Science search strategies for the paediatric review.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-089",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 89",
      "pageStart": 89,
      "pageEnd": 89,
      "target": "../sections/page-089.md",
      "narrative": null,
      "content": "#5 Search: #3 AND #4  \n#6 Search: (Xpert* or GeneXpert* or cepheid) and Ultra  (Topic) \n#7 Search: #5 AND #6  \n#8 Search: (Truenat* or Molbio or TruRif)  (Topic) \n \n#9 Search: #5 AND #8  \n#10 Search: (STANDARD M10 or SD Biosensor or IRON-qPCR or BIONEER)  (Topic)  \nSearch: #5 AND #10 \n \n \nScopus (Elseviers) \n( TITLE-ABS-KEY ( ( standard AND m10 OR sd AND biosensor OR iron-qpcr OR bioneer ) ) ) \nAND ( ( TITLE-ABS-KEY ( ( child OR children* OR childhood OR infan* OR newborn OR neonat* \nOR toddler* OR adolescen* ) ) ) AND ( ( TITLE-ABS-KEY ( ( tuberculosis OR tb ) AND ( \nextrapulmonary OR lymph A",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives the Web of Science, Scopus and regional database search strategies for the paediatric review.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-090",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 90",
      "pageStart": 90,
      "pageEnd": 90,
      "target": "../sections/page-090.md",
      "narrative": null,
      "content": "Search Name: Cochrane Database of Systematic Reviews \nIssue 10 of 12, October 2023 \nID \nSearch Hits \n#1 \nMeSH descriptor: [Tuberculosis] explode all trees \n \n#2 \nMeSH descriptor: [Mycobacterium tuberculosis] explode all trees \n \n#3 \n((tuberculosis or TB or MDR-TB or XDR-TB)):ti,ab,kw  \n#4 \n#1 or #2 or #3  \n#5 \nextrapulmonary tuberculosis \n \n#6 \nMeSH descriptor: [Tuberculosis, Extrapulmonary] explode all trees \n \n#7 \nPTB or EPTB  \n#8 \n#4 or #5 or #6 or #7  \n#9 \nchild or children* or childhood or infan* or newborn or neonat* or toddler* or adolescen* \n#10 \n#8 and #9 \n \n#11 \n(Xpert* or GeneXpert*",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives the Cochrane library and trial registry search strategies for the paediatric review.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p091-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 91,
      "pageEnd": 91,
      "target": "../images/img-p091-1.webp",
      "narrative": "draft",
      "content": "PRISMA flow diagram of study selection for Truenat assays (Appendix 2A). From 1175 database records (524 duplicates removed), 651 were screened and 34 reports assessed, all excluded (including 8 for wrong age group); 4 further reports from other methods were also excluded. No studies were included.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-091",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 91",
      "pageStart": 91,
      "pageEnd": 91,
      "target": "../sections/page-091.md",
      "narrative": null,
      "content": "Appendix 2. PRISMA diagram for other LC-aNAATs \nA. Truenat ( Truenat MTB Plus, and Truenat MTB‐RIF Dx) \n \n \n \n \n \n \n \n \n \n \n \n \nB. STANDARD M10, SD Biosensor; IRON-qPCR, BIONEER",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Appendix 2, PRISMA diagram captions for Truenat and for STANDARD M10 and IRON-qPCR in the paediatric review.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p092-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 92,
      "pageEnd": 92,
      "target": "../images/img-p092-1.webp",
      "narrative": "draft",
      "content": "PRISMA flow diagram of study selection for STANDARD M10 and IRON-qPCR (Appendix 2B). From 37 database records (3 duplicates removed), 34 were screened, 32 excluded and 2 reports assessed; both were excluded (an ongoing trial and a conference abstract/case-control study). No studies were included.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-092",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 92",
      "pageStart": 92,
      "pageEnd": 92,
      "target": "../sections/page-092.md",
      "narrative": null,
      "content": "_(no text on this page)_",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A blank page with no extracted text.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-093",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 93",
      "pageStart": 93,
      "pageEnd": 93,
      "target": "../sections/page-093.md",
      "narrative": null,
      "content": "Appendix 3. Diagnostic accuracy of LC-aNAATs (Xpert Ultra) for detection of \npulmonary tuberculosis: subgroup analyses \n \nNo. \nSubgroup \nSpecimen, \nreference \nstandard \nStudies \nTotal \n(cases) \nSummary \nsensitivity (95% \nCI) \nSummary \nspecificity (95% \nCI) \nPPV (95% CI)* \nNPV (95% CI)* \n1 \nCLHIV \nSputum, MRS \n7 \n400 (57) \n70.4 (56.0 to 81.7) \n95.6 (89.9 to 98.2) \n64.2 (38.0 to 83.4) 96.7 (94.8 to 98.0) \n2 \nCLHIV \nSputum, CRS \n7 \n317 (242) \n18.6 (14.2 to 24.0) \n100.0 (95.2 to \n100.0) \n100.0 (24.7 to \n100.0) 91.7 (90.9 to 92.2) \n3 \nHIV(-) \nSputum, MRS \n10 2081 (267) \n78.3 (67.8 to 86.1) \n95.3 (9",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Appendix 3, the start of a table of Xpert Ultra subgroup accuracy for paediatric pulmonary TB by HIV status, age group and specimen type. It also covers reference standard, giving pooled sensitivity, specificity and predictive values.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-094",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 94",
      "pageStart": 94,
      "pageEnd": 94,
      "target": "../sections/page-094.md",
      "narrative": null,
      "content": "28 \n1-4 yrs \nNPA, MRS \n5 \n799 (57) \n49.3 (36.2 to 62.5) \n98.7 (96.6 to 99.5) \n81.0 (54.5 to 93.5) 94.6 (93.2 to 96.0) \n29 \n1-4 yrs \nNPA, CRS \n4 446 (310) \n11.0 (06.9 to 17.1) \n100.0 (97.3 to \n100.0) \n100.0 (22.4 to \n100.0) 91.0 (90.4 to 91.6) \n30 \n1-4 yrs \nStool, MRS \n7 1420 (154) \n74.6 (63.1 to 83.4) \n95.3 (92.3 to 97.2) \n63.8 (47.6 to 76.6) 97.1 (95.7 to 98.1) \n31 \n1-4 yrs \nStool, CRS \n5 746 (199) \n28.3 (12.0 to 53.3) \n99.3 (97.4 to 99.8) \n81.0 (33.5 to 96.7) 92.6 (90.9 to 95.1) \n32 \n5-9 yrs \nSputum, MRS \n11 890 (112) \n81.9 (65.3 to 91.6) \n95.4 (91.6 to 97.6) \n66.6 (46.3 to 80.6) 97.9 (96.0",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the subgroup accuracy table for children aged 1–4 years and other subgroups by specimen and reference standard.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-095",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 95",
      "pageStart": 95,
      "pageEnd": 95,
      "target": "../sections/page-095.md",
      "narrative": null,
      "content": "12 \nNPA, CRS \n3 \n534 (370) 13.4 (07.0 to 24.2) \n100.0 (97.8 to 100.0) \n100.0 (25.9 to \n100.0) \n91.2 (90.4 to 92.2) \n13 \nStool, MRS \n4 \n2242 \n(100) \n62.1 (48.7 to 48.7) \n99.0 (97.7 to 99.6) \n87.8 (70.5 to 93.1) \n95.9 (94.5 to 94.6) \n14 \nStool, CRS \n2 \n916 (193) 26.8 (05.4 to 70.3) \n99.3 (98.3 to 99.7) \n81.1 (26.5 to 96.4) \n92.4 (90.3 to 96.8) \n* Predictive values were calculated at a pre-test probability of 10%. \nSensitivity, specificity and predictive values are expressed as percentages. \nCI: confidence interval; CRS: composite reference standard; GA: gastric aspirate; MRS: microbiological ref",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Ends the subgroup accuracy table with nasopharyngeal aspirate and stool results and abbreviations; some confidence intervals in the stool row appear misprinted in the source.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-096",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 96",
      "pageStart": 96,
      "pageEnd": 96,
      "target": "../sections/page-096.md",
      "narrative": null,
      "content": "Web Annex B.3. Accuracy of LC-aNAATs for extrapulmonary \ntuberculosis and rifampicin resistance in adults and adolescents: \na systematic review \n \nMikashmi Kohli1, Leeberk Raja Inbaraj2, Angela Salomon3, Jefferson Daniel4, Mukesh \nKumar Sathya Narayanan5, Vignes Anand Srinivasalu2, Adhin Bhaskar6, Katie Scandrett7, \nKaren R Steingart8, Yemisi Takwoingi7 \n \n1Xpert Ultra review lead \n2Truenat review lead \n \n1 FIND, Geneva, Switzerland \n2 Dept. of Clinical Research, ICMR- National Institute for Research in Tuberculosis, Chennai, \nIndia \n3 McGill University, Montreal, Canada \n4 Christian Medical C",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title, authors and affiliations of Web Annex B.3, a systematic review of LC-aNAAT accuracy for extrapulmonary TB and rifampicin resistance in adults and adolescents.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-097",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 97",
      "pageStart": 97,
      "pageEnd": 97,
      "target": "../sections/page-097.md",
      "narrative": null,
      "content": "Background \nGlobally, extrapulmonary tuberculosis accounted for 20% of the 7.4 million cases of \ntuberculosis notified in 2022, ranging from 8% in the WHO Western Pacific Region to 24% in \nthe Eastern Mediterranean Region (WHO Global Tuberculosis Report 2023). In 2022, there \nwere about 520,000 new cases of rifampicin-resistant tuberculosis; of these cases, 78% had \nmultidrug-resistant tuberculosis (WHO Global Tuberculosis Report 2023). Xpert Ultra and \nTruenat assays are WHO-recommended rapid tests that simultaneously detect tuberculosis and \nrifampicin resistance in people with signs and sym",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives background on extrapulmonary TB, which made up a fifth of notified TB cases in 2022, and on rifampicin-resistant TB. It describes Xpert Ultra's improved detection limit compared with Xpert MTB/RIF and its new trace result category.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-098",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 98",
      "pageStart": 98,
      "pageEnd": 98,
      "target": "../sections/page-098.md",
      "narrative": null,
      "content": "that corresponds to the lowest bacillary load for M. tuberculosis detection. This new category \nis reported as 'MTB trace DETECTED'. Interpreting a trace call result requires a reassessment \nof clinical symptoms and history of prior tuberculosis. No rifampicin resistance results are \navailable (indeterminate) for people with trace results. As with Xpert MTB/RIF, Xpert Ultra \ndetects both live and dead bacteria. \nTo address limitations in rifampicin resistance detection, Xpert Ultra uses melting temperature-\nbased analysis, in lieu of real-time PCR analysis with Xpert MTB/RIF. Melting temperatu",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains how Xpert Ultra trace results should be interpreted and describes the other index tests, including Truenat, STANDARD M10 and the multiplex IRON-qPCR platform.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-099",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 99",
      "pageStart": 99,
      "pageEnd": 99,
      "target": "../sections/page-099.md",
      "narrative": null,
      "content": "in 40 minutes. The IRON qPCR RFIA kit is a cartridge-based assay that detects M. tuberculosis \nand resistance to 11 drugs: rifampicin, isoniazid, fluoroquinolones (levofloxacin, \nmoxifloxacin, gatifloxacin and ofloxacin), aminoglycosides (kanamycin, amikacin, \ncapreomycin), prothionamide and ethionamide. It uses IRON qPCR, which is a single \ninstrument for extraction and amplification and detection. Two modules are available per \ninstrument (Bioneer 2023). IRON qPCR RFIA is not commercially available. \n \nMethods \n \nSearch methods \n \nElectronic searches \nThe Cochrane Infectious Diseases Group I",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the IRON-qPCR kit for detecting TB and resistance to many drugs, then sets out the review's methods for analysing results per participant and by type of extrapulmonary TB.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-100",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 100",
      "pageStart": 100,
      "pageEnd": 100,
      "target": "../sections/page-100.md",
      "narrative": null,
      "content": "diagnosis), and rifampicin resistance. We also performed analyses separately by type of \nreference standard (microbiological or composite). Sensitivity and specificity were estimated \nfor each included study with 95% confidence intervals (CIs) and these estimates were presented \non forest plots and in receiver operating characteristic (ROC) space on summary ROC (SROC) \nplots.  \n \nWe used a bivariate random-effects model to determine summary estimates for sensitivity and \nspecificity (Chu 2006; Reitsma 2005). If bivariate meta‐analysis was not possible due to sparse \ndata, or if there was littl",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains the meta-analysis approach for the extrapulmonary review, including bivariate models and analyses by reference standard, and introduces the search results.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-vfig-p101",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 101,
      "pageEnd": 101,
      "target": "../figures/vfig-p101.png",
      "narrative": "draft",
      "content": "PRISMA 2020 study flow diagram for LC-aNAATs. 11 studies were included in the previous review; 1776 database records, 871 duplicates removed, 905 screened, 794 excluded, 111 reports assessed and 86 excluded (reasons in supplementary figures); 4 records from public calls were assessed with 1 excluded (inappropriate reference standard). 25 new studies (3 reports) were included, giving 36 studies in total (3 reports).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-101",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 101",
      "pageStart": 101,
      "pageEnd": 101,
      "target": "../sections/page-101.md",
      "narrative": null,
      "content": "96\n \nFigure 1. Study flow diagram for LC-aNAATs \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \nRecords identified from: \nDatabases (n =1776) \n \nRecords removed before \nscreening: \nDuplicate records removed  \n(n =871 ) \n \nRecords screened \n(n = 905) \nRecords excluded \n(n =  794) \nReports sought for retrieval \n(n =  111) \nReports not retrieved \n(n = 0) \nReports assessed for eligibility \n(n = 111) \nReports excluded:86 \n(See reasons for each test in \nsupplementary figures 1-3) \nRecords identified from: \nPublic calls (n = 4) \nReports assessed for eligibility \n(n = 4)",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The study flow diagram for the extrapulmonary LC-aNAAT review, showing records identified, screened, excluded and included, with a PRISMA reference.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-102",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 102",
      "pageStart": 102,
      "pageEnd": 102,
      "target": "../sections/page-102.md",
      "narrative": null,
      "content": "97\nTable 1. Key characteristics of included studies  \nStudy \nCountry \nReference \nstandard \nPLHIV % \nClinical setting \nHigh \nTB \nburden \nHigh \nHIV-TB \nburden \ncountry \nSpecimen type \nXpert Ultra \nAlomatu 2023 \nSouth Africa \nMGIT \n52.6 \nTertiary care centre \nYes \nYes \nPericardial fluid \nAntel 2020 \nSouth Africa \nMGIT \n51 \nTertiary care centre \nYes \nYes \nLymph node aspirate and biopsy \nBahr 2017 \nUganda \nMGIT \n100 \nTertiary care centre \nYes \nYes \nCerebrospinal fluid \nBoloko 2022 \nSouth Africa \nMGIT \n100 \nTertiary care centre \nYes \nYes \nBlood \nChin 2019 \nUganda \nMGIT \n18 \nTertiary care centre \nYes",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 1 of a systematic review on low-complexity automated nucleic acid amplification tests (LC-aNAATs) for extrapulmonary TB, listing included Xpert Ultra studies by country, reference standard, share of people with HIV, setting and specimen type. Most studies are from tertiary centres in high-burden countries such as South Africa, Uganda and India.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-103",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 103",
      "pageStart": 103,
      "pageEnd": 103,
      "target": "../sections/page-103.md",
      "narrative": null,
      "content": "98\nStudy \nCountry \nReference \nstandard \nPLHIV % \nClinical setting \nHigh \nTB \nburden \nHigh \nHIV-TB \nburden \ncountry \nSpecimen type \nPenata 2021 \nColombia \nMGIT and \nLJ \n26 \nUniversity hospital \nNo \nNo \nCerebrospinal fluid, pleural fluid, bone or \njoint aspirate, pericardial fluid, peritoneal \nfluid, urine, lymph node biopsy \nPerez-Risco 2018 \nSpain \nMGIT and \nLJ \nUnknown \nReferral laboratory \nNo \nNo \nCerebrospinal fluid, pleural fluid, bone or \njoint aspirate, peritoneal fluid, urine \nQuinn 2021 \nUganda \nMGIT \n94 \nTertiary care centre \nYes \nYes \nCerebrospinal fluid \nShao 2020 \nChina \nMGIT \n0 \nT",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues Table 1 of included studies, adding studies from Colombia, Spain, Uganda, China and India with their reference standards and specimen types. Footnotes define the culture methods and abbreviations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p104-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 104,
      "pageEnd": 104,
      "target": "../images/img-p104-1.webp",
      "narrative": "draft",
      "content": "Stacked horizontal bar chart summarizing risk of bias and applicability concerns for 36 Xpert Ultra studies (high red, unclear yellow, low green). Risk of bias: patient selection 1 high, 5 unclear, 30 low; index test 36 low; reference standard 9 unclear, 27 low; flow and timing 36 low. Applicability: patient selection 4 high, 21 unclear, 11 low; index test 3 high, 1 unclear, 32 low; reference standard 5 unclear, 31 low.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-104",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 104",
      "pageStart": 104,
      "pageEnd": 104,
      "target": "../sections/page-104.md",
      "narrative": null,
      "content": "Findings \nWe included 36 studies for Xpert Ultra and three studies for Truenat MTB plus. We found no \neligible studies for the other index tests. For a class level analysis for each specimen type \nacross different technologies, we could only include studies from Xpert Ultra based on the \ncriteria in Box 1. Therefore, the summary estimates in Table 2 were obtained from only \nstudies of Xpert Ultra. We present the findings in detail below for each specimen type \ntogether with statements about the certainty of the evidence. Appendix 2 contains the \nsummary of findings tables from the assessment d",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that 36 Xpert Ultra studies and three Truenat MTB Plus studies were included, with no eligible studies for other index tests, so pooled estimates come from Xpert Ultra only. It notes that GRADE summary of findings tables are in an appendix and presents a risk of bias and applicability summary figure.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p105-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 105,
      "pageEnd": 105,
      "target": "../images/img-p105-1.webp",
      "narrative": "draft",
      "content": "Risk-of-bias and applicability chart for three Truenat MTB Plus studies (Jose 2024, Sharma 2021, Sharma 2023). Jose 2024 is rated high risk for the reference standard; both Sharma studies are unclear for the reference standard in both risk of bias and applicability; all other domains are low.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-105",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 105",
      "pageStart": 105,
      "pageEnd": 105,
      "target": "../sections/page-105.md",
      "narrative": null,
      "content": "Figure 2b. Summary of risk of bias and applicability assessment for each \nTruenat MTB plus study \n \nThe estimates of the diagnostic accuracy of LC-aNAATs for each form of extrapulmonary \ntuberculosis is given in Table 2. \n \nTable 2. Diagnostic accuracy of LC-aNAATs for detection of extrapulmonary \ntuberculosis in adults and adolescents \nEPTB \nspecimen  \nReference \nstandard \nNumber \nof \nstudies \nNumber of \nparticipants \n(cases) \nSummary \nsensitivity  \n(95% CI) \nSummary \nspecificity \n(95% CI) \nPositive \npredictive \nvalue  \n(95% CI)* \nNegative \npredictive \nvalue  \n(95% \nCI)* \nCerebrospinal \nfluid",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 2 gives pooled sensitivity, specificity and predictive values of LC-aNAATs for extrapulmonary TB by specimen and reference standard, with cerebrospinal fluid showing 88.2% sensitivity against microbiological reference. Pleural fluid and urine show lower sensitivity against a composite reference standard.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-106",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 106",
      "pageStart": 106,
      "pageEnd": 106,
      "target": "../sections/page-106.md",
      "narrative": null,
      "content": "EPTB \nspecimen  \nReference \nstandard \nNumber \nof \nstudies \nNumber of \nparticipants \n(cases) \nSummary \nsensitivity  \n(95% CI) \nSummary \nspecificity \n(95% CI) \nPositive \npredictive \nvalue  \n(95% CI)* \nNegative \npredictive \nvalue  \n(95% \nCI)* \nUrine \nCRS \n3 \n693 (159) \n \n23.0 (14.7 \nto 34.1) \n98.9 (89.7 \nto 99.9) \n70.7 (13.7 \nto 97.4) \n92.0 (90.4 \nto 93.2) \nBone or joint \nfluid \nMRS \n3 \n126 (58) \n \n96.6 (87.2 \nto 99.1) \n91.1 (80.8 \nto 96.2) \n54.8 (33.6 \nto 74.2) \n99.6 (98.3 \nto 99.9) \nBone or joint \nfluid \nCRS \n1 \n145 (111) \n \n- \n- \n- \n- \nPericardial \nfluid \nMRS \n3 \n202 (75) \n \n84.0 (73.9 \nto 90.",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes Table 2 with estimates for urine, bone or joint fluid, pericardial, peritoneal and blood specimens, where bone or joint fluid shows high sensitivity. It then introduces the results for tuberculous meningitis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p107-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 107,
      "pageEnd": 107,
      "target": "../images/img-p107-1.webp",
      "narrative": "draft",
      "content": "Risk of bias and applicability chart for tuberculous meningitis in cerebrospinal fluid, with one row per study (18 studies). Each study is rated for risk of bias (patient selection, index test, reference standard, flow and timing) and applicability (patient selection, index test, reference standard) using green (low), yellow (unclear) and red (high) circles. Most ratings are low; high ratings are given for Perez-Risco 2018 (risk of bias, patient selection); Mekkaoui 2021, Slail 2023 (applicability, patient selection); Chin 2019, Shao 2020 (applicability, index test).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-107",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 107",
      "pageStart": 107,
      "pageEnd": 107,
      "target": "../sections/page-107.md",
      "narrative": null,
      "content": "Figure 3. Risk of bias and applicability summary for each study: tuberculous \nmeningitis in cerebrospinal fluid \n \nShould LC-aNAATs be used to diagnose tuberculous meningitis in cerebrospinal fluid in \nadults with signs and symptoms of tuberculous meningitis, against a microbiological \nreference standard? \nLC-aNAAT summary sensitivity and specificity (95%) were 88.2% (83.7 to 91.6) and 96.0% \n(86.8 to 98.9) based on 16 Xpert Ultra studies (1684 participants); high certainty (sensitivity) \nand moderate certainty (specificity) evidence. \n \nShould LC-aNAATs be used to diagnose tuberculous meningi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that Xpert Ultra on cerebrospinal fluid had 88.2% sensitivity and 96.0% specificity against a microbiological reference and 60.3% sensitivity against a composite reference for tuberculous meningitis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p108-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 108,
      "pageEnd": 108,
      "target": "../images/img-p108-1.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Xpert Ultra; specimen: cerebrospinal fluid; reference standard: culture; 18 studies. For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity ranges from 0.50 to 1.00. Sensitivity is not estimable for Mekkaoui 2021, Osei 2019. Specificity ranges from 0.50 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p108-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 108,
      "pageEnd": 108,
      "target": "../images/img-p108-2.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Xpert Ultra; specimen: cerebrospinal fluid; reference standard: composite reference standard; 10 studies. For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity ranges from 0.18 to 0.76. Specificity ranges from 0.95 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p108-3",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 108,
      "pageEnd": 108,
      "target": "../images/img-p108-3.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Truenat MTB plus; specimen: cerebrospinal fluid; reference standard: culture; 2 studies (Jose 2024, Sharma 2021). For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity ranges from 0.95 to 1.00. Specificity ranges from 0.55 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p108-4",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 108,
      "pageEnd": 108,
      "target": "../images/img-p108-4.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Truenat MTB plus; specimen: cerebrospinal fluid; reference standard: composite reference standard; 1 study (Sharma 2021). For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity is 0.79 (95% CI 0.70–0.86). Specificity is 1.00 (95% CI 0.91–1.00).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-108",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 108",
      "pageStart": 108,
      "pageEnd": 108,
      "target": "../sections/page-108.md",
      "narrative": null,
      "content": "Figure 4. LC-aNAATs for tuberculous meningitis in cerebrospinal fluid \nThe studies are ordered by decreasing sensitivity. CRS: composite reference standard; CSF: cerebrospinal fluid; \nFN: false negative; FP: false positive; TB: tuberculosis; TN: true negative; TP: true positive. \n \nDetection of lymph node tuberculosis \nThe methodological quality assessment results are summarised in Figure 5 and study \nestimates of sensitivity and specificity are presented in Figure 6 for lymph node tuberculosis \nin lymph node aspirate.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A figure caption for forest plots of LC-aNAAT accuracy in cerebrospinal fluid, followed by an introduction to results for lymph node TB.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p109-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 109,
      "pageEnd": 109,
      "target": "../images/img-p109-1.webp",
      "narrative": "draft",
      "content": "Risk of bias and applicability chart for lymph node tuberculosis in lymph node aspirate, with one row per study (9 studies). Each study is rated for risk of bias (patient selection, index test, reference standard, flow and timing) and applicability (patient selection, index test, reference standard) using green (low), yellow (unclear) and red (high) circles. Most ratings are low; high ratings are given for Hoel 2020, Hoel 2020a, Slail 2023 (applicability, patient selection).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-109",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 109",
      "pageStart": 109,
      "pageEnd": 109,
      "target": "../sections/page-109.md",
      "narrative": null,
      "content": "Figure 5. Risk of bias and applicability summary for each study: lymph node \ntuberculosis in lymph node aspirate \n \nShould LC-aNAATs be used to diagnose lymph node tuberculosis in lymph node aspirate in \nadults and adolescents with signs and symptoms of lymph node tuberculosis, against a \nmicrobiological reference standard? \nLC-aNAATs summary sensitivity and specificity (95% CI) were 85.3% (73.4 to 92.4) and \n74.1% (63.5 to 82.5) from nine Xpert Ultra studies (445 participants); low certainty \n(sensitivity) and very low certainty (specificity) evidence. \n \nShould LC-aNAATs be used to diagnose",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports pooled Xpert Ultra accuracy on lymph node aspirate, with 85.3% sensitivity and 74.1% specificity against a microbiological reference and 71.3% sensitivity against a composite reference, with low to very low certainty.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p110-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 110,
      "pageEnd": 110,
      "target": "../images/img-p110-1.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Xpert Ultra; specimen: lymph node aspirate; reference standard: culture; 9 studies. For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity ranges from 0.50 to 1.00. Specificity ranges from 0.33 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p110-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 110,
      "pageEnd": 110,
      "target": "../images/img-p110-2.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Xpert Ultra; specimen: lymph node aspirate; reference standard: composite reference standard; 6 studies. For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity ranges from 0.60 to 0.81. Specificity ranges from 0.71 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p110-3",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 110,
      "pageEnd": 110,
      "target": "../images/img-p110-3.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Truenat MTB plus; specimen: lymph node aspirate; reference standard: composite reference standard (upper panel) and culture (lower panel); 1 study (Sharma 2023). For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Against the composite reference standard, sensitivity is 0.77 and specificity 1.00; against culture, sensitivity is 1.00 and specificity 0.56.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-110",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 110",
      "pageStart": 110,
      "pageEnd": 110,
      "target": "../sections/page-110.md",
      "narrative": null,
      "content": "Figure 6. LC-aNAATs for lymph node tuberculosis in lymph node aspirate \nThe studies are ordered by decreasing sensitivity. CRS: composite reference standard; FN: false negative; FP: \nfalse positive; LN: lymph node; TB: tuberculosis; TN: true negative; TP: true positive. \n \nThe methodological quality assessment results are summarised in Figure 7 and study \nestimates of sensitivity and specificity are presented in Figure 8 for lymph node tuberculosis \nin lymph node tissue.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A figure caption for lymph node aspirate forest plots, introducing results for lymph node tissue.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p111-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 111,
      "pageEnd": 111,
      "target": "../images/img-p111-1.webp",
      "narrative": "draft",
      "content": "Risk of bias and applicability chart for lymph node tuberculosis in lymph node tissue, with one row per study (9 studies). Each study is rated for risk of bias (patient selection, index test, reference standard, flow and timing) and applicability (patient selection, index test, reference standard) using green (low), yellow (unclear) and red (high) circles. Most ratings are low; high ratings are given for Hoel 2020, Hoel 2020a, Mekkaoui 2021, Slail 2023 (applicability, patient selection); Wu 2019 (applicability, index test).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-111",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 111",
      "pageStart": 111,
      "pageEnd": 111,
      "target": "../sections/page-111.md",
      "narrative": null,
      "content": "Figure 7. Risk of bias and applicability summary for each study: lymph node \ntuberculosis in lymph node tissue \n \nShould LC-aNAATs be used to diagnose lymph node tuberculosis in lymph node tissue in \nadults and adolescents with signs and symptoms of lymph node tuberculosis, against a \nmicrobiological reference standard? \nLC-aNAATs summary sensitivity and specificity (95% CI) were 96.5% (84.7 to 99.3) and \n79.4% (65.4 to 88.8) from eight Xpert Ultra studies (578 participants); moderate certainty \n(sensitivity) and very low certainty (specificity) evidence. \n \nShould LC-aNAATs be used to diagnos",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports pooled Xpert Ultra accuracy on lymph node tissue, with 96.5% sensitivity and 79.4% specificity against a microbiological reference and lower sensitivity against a composite reference.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p112-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 112,
      "pageEnd": 112,
      "target": "../images/img-p112-1.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Xpert Ultra; specimen: lymph node biopsy; reference standard: culture; 9 studies. For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity ranges from 0.88 to 1.00. Sensitivity is not estimable for Hoel 2020. Specificity ranges from 0.38 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p112-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 112,
      "pageEnd": 112,
      "target": "../images/img-p112-2.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Xpert Ultra; specimen: lymph node biopsy; reference standard: composite reference standard; 4 studies. For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity ranges from 0.47 to 0.73. Sensitivity is not estimable for Hoel 2020. Specificity ranges from 0.96 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p112-3",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 112,
      "pageEnd": 112,
      "target": "../images/img-p112-3.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Truenat MTB plus; specimen: lymph node biopsy; reference standard: culture; 1 study (Jose 2024). For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity is 0.86 (95% CI 0.42–1.00). Specificity is 0.98 (95% CI 0.91–1.00).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p112-4",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 112,
      "pageEnd": 112,
      "target": "../images/img-p112-4.webp",
      "narrative": "draft",
      "content": "Risk of bias and applicability chart for pleural tuberculosis in pleural fluid, with one row per study (14 studies). Each study is rated for risk of bias (patient selection, index test, reference standard, flow and timing) and applicability (patient selection, index test, reference standard) using green (low), yellow (unclear) and red (high) circles. Most ratings are low; high ratings are given for Perez-Risco 2018 (risk of bias, patient selection); Hoel 2020, Mekkaoui 2021, Slail 2023 (applicability, patient selection); Wu 2019 (applicability, index test).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-112",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 112",
      "pageStart": 112,
      "pageEnd": 112,
      "target": "../sections/page-112.md",
      "narrative": null,
      "content": "Figure 8. LC-aNAATs for lymph node tuberculosis in lymph node tissue \nThe studies are ordered by decreasing sensitivity. FN: false negative; FP: false positive; TB: tuberculosis; TN: \ntrue negative; TP: true positive \n \nDetection of pleural tuberculosis  \nThe methodological quality assessment results are summarised in Figure 9 and study \nestimates of sensitivity and specificity are presented in Figure 10 for pleural tuberculosis in \npleural fluid. \n \n \nFigure 9. Risk of bias and applicability summary for each study: pleural \ntuberculosis in pleural fluid",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A figure caption for lymph node tissue results, introducing the risk of bias summary for pleural TB in pleural fluid.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p113-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 113,
      "pageEnd": 113,
      "target": "../images/img-p113-1.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity by study (13 studies), part of Figure 10 (LC-aNAATs for pleural tuberculosis in pleural fluid); the panel heading naming the test and reference standard is not included in the image. For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity ranges from 0.00 to 1.00. Specificity ranges from 0.57 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p113-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 113,
      "pageEnd": 113,
      "target": "../images/img-p113-2.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Xpert Ultra; specimen: pleural fluid; reference standard: composite reference standard; 7 studies. For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity ranges from 0.00 to 0.61. Specificity ranges from 0.94 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p113-3",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 113,
      "pageEnd": 113,
      "target": "../images/img-p113-3.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Truenat MTB plus; specimen: pleural fluid; reference standard: culture; 1 study (Jose 2024). For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity is 1.00 (95% CI 0.03–1.00). Specificity is 1.00 (95% CI 0.95–1.00).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-113",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 113",
      "pageStart": 113,
      "pageEnd": 113,
      "target": "../sections/page-113.md",
      "narrative": null,
      "content": "Should LC-aNAATs be used to diagnose pleural tuberculosis in pleural fluid in adults and \nadolescents with signs and symptoms of pleural tuberculosis, against a microbiological \nreference standard? \nLC-aNAATs summary sensitivity and specificity (95% CI) were 74.0% (60.8 to 83.9) and \n88.1% (78.8 to 93.6) from 13 Xpert Ultra studies (1041 participants); low certainty \n(sensitivity) and very low certainty (specificity) evidence.  \n \nShould LC-aNAATs be used to diagnose pleural tuberculosis in pleural fluid in adults and \nadolescents with signs and symptoms of pleural tuberculosis, against a comp",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports Xpert Ultra accuracy for pleural TB in pleural fluid, with 74.0% sensitivity and 88.1% specificity from 13 studies against a microbiological reference. It also gives estimates against a composite reference and introduces results for pleural tissue.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p114-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 114,
      "pageEnd": 114,
      "target": "../images/img-p114-1.webp",
      "narrative": "draft",
      "content": "Risk of bias and applicability chart for pleural tuberculosis in pleural tissue, with one row per study (2 studies (Christopher 2021, Gao 2021)). Each study is rated for risk of bias (patient selection, index test, reference standard, flow and timing) and applicability (patient selection, index test, reference standard) using green (low), yellow (unclear) and red (high) circles. No study is rated high; both are rated unclear for applicability of patient selection, and Gao 2021 is also unclear for the reference standard.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p114-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 114,
      "pageEnd": 114,
      "target": "../images/img-p114-2.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Xpert Ultra; specimen: pleural tissue; reference standard: culture; 2 studies (Gao 2021, Christopher 2021). For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity ranges from 0.80 to 1.00. Specificity ranges from 0.75 to 0.86.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p114-3",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 114,
      "pageEnd": 114,
      "target": "../images/img-p114-3.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Xpert Ultra; specimen: pleural tissue; reference standard: composite reference standard; 2 studies (Gao 2021, Christopher 2021). For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity ranges from 0.54 to 0.81. Specificity ranges from 0.94 to 0.97.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-114",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 114",
      "pageStart": 114,
      "pageEnd": 114,
      "target": "../sections/page-114.md",
      "narrative": null,
      "content": "Figure 11. Risk of bias and applicability summary for each study: pleural tuberculosis in \npleural tissue \n \nShould LC-aNAATs be used to diagnose pleural tuberculosis in pleural tissue in adults and \nadolescents with signs and symptoms of pleural tuberculosis, against a microbiological \nreference standard? \nLC-aNAATs sensitivities were 80% and 100% and specificities were 75% and 86% from two \nXpert Ultra studies (105 participants); low certainty (sensitivity) and very low certainty \n(specificity) evidence.  \n \nShould LC-aNAATs be used to diagnose pleural tuberculosis in pleural tissue in adult",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that two studies on pleural tissue showed sensitivities of 80–100% and specificities of 75–86% against a microbiological reference, with low or very low certainty. It also covers composite reference results and introduces genitourinary TB in urine.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p115-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 115,
      "pageEnd": 115,
      "target": "../images/img-p115-1.webp",
      "narrative": "draft",
      "content": "Risk of bias and applicability chart for genitourinary tuberculosis in urine, with one row per study (6 studies). Each study is rated for risk of bias (patient selection, index test, reference standard, flow and timing) and applicability (patient selection, index test, reference standard) using green (low), yellow (unclear) and red (high) circles. Most ratings are low; high ratings are given for Perez-Risco 2018 (risk of bias, patient selection); Mekkaoui 2021 (applicability, patient selection). The final study row and the legend continue in a separate image.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p115-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 115,
      "pageEnd": 115,
      "target": "../images/img-p115-2.webp",
      "narrative": "draft",
      "content": "Final row of the risk of bias and applicability chart for genitourinary tuberculosis in urine (Figure 13): Spener-Gomes 2021 is rated low on all four risk-of-bias domains, unclear on applicability of patient selection and index test, and low on applicability of the reference standard. Below it is the legend: red (high), green (low), yellow (unclear).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p115-3",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 115,
      "pageEnd": 115,
      "target": "../images/img-p115-3.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Xpert Ultra; specimen: urine; reference standard: culture; 7 studies. For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity ranges from 0.25 to 1.00. Sensitivity is not estimable for Huerga 2023, Mekkaoui 2021, Osei 2019. Specificity ranges from 0.28 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p115-4",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 115,
      "pageEnd": 115,
      "target": "../images/img-p115-4.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Xpert Ultra; specimen: urine; reference standard: composite reference standard; 3 studies (Minnies 2021, Huerga 2023, Cresswell 2020a). For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity ranges from 0.13 to 0.34. Specificity ranges from 0.95 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p115-5",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 115,
      "pageEnd": 115,
      "target": "../images/img-p115-5.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Truenat MTB plus; specimen: urine; reference standard: culture; 1 study (Jose 2024). For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity is not estimable for Jose 2024. Specificity is 0.90 (95% CI 0.55–1.00).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-115",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 115",
      "pageStart": 115,
      "pageEnd": 115,
      "target": "../sections/page-115.md",
      "narrative": null,
      "content": "Figure 13. Risk of bias and applicability summary for each study: \ngenitourinary tuberculosis in urine \n \nShould LC-aNAATs be used to diagnose genitourinary tuberculosis in urine in adults and \nadolescents with signs and symptoms of genitourinary tuberculosis, against a \nmicrobiological reference standard? \nLC-aNAATs sensitivities were all 100% (3 studies) and specificities were between 28% and \n100% (6 studies, 232 participants); very low certainty (sensitivity and specificity) evidence.  \n \nShould LC-aNAATs be used to diagnose genitourinary tuberculosis in urine in adults and \nadolescents wi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that Xpert Ultra on urine had 100% sensitivity in three studies but widely varying specificity against a microbiological reference, with low sensitivity against a composite reference. It introduces results for bone or joint TB.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p116-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 116,
      "pageEnd": 116,
      "target": "../images/img-p116-1.webp",
      "narrative": "draft",
      "content": "Risk of bias and applicability chart for bone or joint tuberculosis in bone or joint fluid, with one row per study (5 studies). Each study is rated for risk of bias (patient selection, index test, reference standard, flow and timing) and applicability (patient selection, index test, reference standard) using green (low), yellow (unclear) and red (high) circles. Most ratings are low; high ratings are given for Perez-Risco 2018 (risk of bias, patient selection); Mekkaoui 2021, Slail 2023 (applicability, patient selection).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-116",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 116",
      "pageStart": 116,
      "pageEnd": 116,
      "target": "../sections/page-116.md",
      "narrative": null,
      "content": "The methodological quality assessment results are summarised in Figure 15 and study \nestimates of sensitivity and specificity are presented in Figure 16 for detection of bone or \njoint tuberculosis in bone or joint fluid. \n \nFigure 15. Risk of bias and applicability summary for each study: bone or joint \ntuberculosis in bone of joint fluid \n \nShould LC-aNAATs be used to diagnose bone or joint tuberculosis in bone or joint aspirate \nin adults and adolescents with signs and symptoms of bone or joint tuberculosis, against a \nmicrobiological reference standard? \nLC-aNAATs summary sensitivity and s",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports Xpert Ultra accuracy for bone or joint TB, with 96.6% sensitivity and 91.1% specificity from three studies against a microbiological reference. A single study against a composite reference gave 90% sensitivity and 97% specificity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p117-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 117,
      "pageEnd": 117,
      "target": "../images/img-p117-1.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Xpert Ultra; specimen: bone or joint aspirate; reference standard: culture; 5 studies. For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity ranges from 0.88 to 1.00. Sensitivity is not estimable for Peñata 2021. Specificity ranges from 0.80 to 1.00. Specificity is not estimable for Perez-Risco 2018.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p117-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 117,
      "pageEnd": 117,
      "target": "../images/img-p117-2.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Xpert Ultra; specimen: bone or joint aspirate; reference standard: composite reference standard; 1 study (Sun 2019). For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity is 0.96 (95% CI 0.91–0.99). Specificity is 0.97 (95% CI 0.85–1.00).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p117-3",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 117,
      "pageEnd": 117,
      "target": "../images/img-p117-3.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Truenat MTB plus; specimen: synovial fluid (upper panel) and synovial/bone tissue (lower panel); reference standard: culture; 1 study (Jose 2024). For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. In synovial fluid, sensitivity is not estimable and specificity is 1.00 (95% CI 0.59–1.00); in synovial/bone tissue, sensitivity is 1.00 (0.29–1.00) and specificity 1.00 (0.88–1.00).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p117-4",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 117,
      "pageEnd": 117,
      "target": "../images/img-p117-4.webp",
      "narrative": "draft",
      "content": "Risk of bias and applicability chart for pericardial tuberculosis in pericardial fluid, with one row per study (5 studies). Each study is rated for risk of bias (patient selection, index test, reference standard, flow and timing) and applicability (patient selection, index test, reference standard) using green (low), yellow (unclear) and red (high) circles. Most ratings are low; high ratings are given for Mekkaoui 2021, Slail 2023 (applicability, patient selection).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-117",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 117",
      "pageStart": 117,
      "pageEnd": 117,
      "target": "../sections/page-117.md",
      "narrative": null,
      "content": "Figure 16. LC-aNAATs for bone or joint tuberculosis in bone or joint fluid \nThe studies are ordered by decreasing sensitivity. FN: false negative; FP: false positive; TB: tuberculosis; TN: \ntrue negative; TP: true positive. \n \nDetection of pericardial tuberculosis \nThe methodological quality assessment results are summarised in Figure 17 and study \nestimates of sensitivity and specificity are presented in Figure 18 for detection of pericardial \ntuberculosis in pericardial fluid. \n \nFigure 17. Risk of bias and applicability summary for each study: pericardial \ntuberculosis in pericardial fluid",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports Xpert Ultra accuracy for pericardial TB in pericardial fluid, with 84.0% sensitivity and 86.6% specificity from three studies, at low certainty.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p118-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 118,
      "pageEnd": 118,
      "target": "../images/img-p118-1.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Xpert Ultra; specimen: pericardial fluid; reference standard: culture; 5 studies. For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity ranges from 0.67 to 0.86. Sensitivity is not estimable for Mekkaoui 2021, Peñata 2021. Specificity ranges from 0.83 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p118-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 118,
      "pageEnd": 118,
      "target": "../images/img-p118-2.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Xpert Ultra; specimen: peritoneal fluid; reference standard: culture; 3 studies (Peñata 2021, Slail 2023, Perez-Risco 2018). For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity ranges from 0.33 to 0.67. Specificity ranges from 0.94 to 1.00. Specificity is not estimable for Perez-Risco 2018.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p118-3",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 118,
      "pageEnd": 118,
      "target": "../images/img-p118-3.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Truenat MTB plus; specimen: peritoneal fluid; reference standard: culture; 1 study (Jose 2024). For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity is 0.00 (95% CI 0.00–0.97). Specificity is 0.93 (95% CI 0.68–1.00).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p118-4",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 118,
      "pageEnd": 118,
      "target": "../images/img-p118-4.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity. Test: Xpert Ultra; specimen: blood; reference standard: culture; 1 study (Boloko 2022). For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity is 0.38 (95% CI 0.33–0.43). Specificity is 0.98 (95% CI 0.94–1.00).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p118-5",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 118,
      "pageEnd": 118,
      "target": "../images/img-p118-5.webp",
      "narrative": "draft",
      "content": "Risk of bias and applicability chart for peritoneal tuberculosis in peritoneal fluid, with one row per study (3 studies (Penata 2021, Slail 2023, Perez-Risco 2018)). Each study is rated for risk of bias (patient selection, index test, reference standard, flow and timing) and applicability (patient selection, index test, reference standard) using green (low), yellow (unclear) and red (high) circles. Most ratings are low; high ratings are given for Perez-Risco 2018 (risk of bias, patient selection); Slail 2023 (applicability, patient selection).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-118",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 118",
      "pageStart": 118,
      "pageEnd": 118,
      "target": "../sections/page-118.md",
      "narrative": null,
      "content": "Figure 18. LC-aNAATs for pericardial, peritoneal and disseminated \ntuberculosis \nThe studies are ordered by decreasing sensitivity. FN: false negative; FP: false positive; TN: true negative; TP: \ntrue positive. \n \nDetection of peritoneal tuberculosis \nThe methodological quality assessment results are summarised in Figure 19 and study \nestimates of sensitivity and specificity are presented in Figure 18 for detection of peritoneal \ntuberculosis in peritoneal fluid. \n \nFigure 19. Risk of bias and applicability summary for each study: peritoneal \ntuberculosis in peritoneal fluid \n \nShould LC-aNAAT",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that Xpert Ultra sensitivity for peritoneal TB in peritoneal fluid ranged from 33% to 67% with specificities of 94–100% in three small studies, at very low to low certainty.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p119-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 119,
      "pageEnd": 119,
      "target": "../images/img-p119-1.webp",
      "narrative": "draft",
      "content": "Risk of bias and applicability chart for disseminated tuberculosis in blood, with one row per study (1 study (Boloko 2022)). Each study is rated for risk of bias (patient selection, index test, reference standard, flow and timing) and applicability (patient selection, index test, reference standard) using green (low), yellow (unclear) and red (high) circles. Boloko 2022 is rated low on all four risk-of-bias domains and on applicability of the index test, and unclear on applicability of patient selection and reference standard.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p119-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 119,
      "pageEnd": 119,
      "target": "../images/img-p119-2.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity for 3 studies (Bahr 2017, Cresswell 2020, Quinn 2021), part of Figure 21 (LC-aNAAT for TB meningitis in HIV-positive adults and adolescents); the panel heading is not included in the image. For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity ranges from 0.87 to 0.90. Specificity ranges from 0.88 to 0.92.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-119",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 119",
      "pageStart": 119,
      "pageEnd": 119,
      "target": "../sections/page-119.md",
      "narrative": null,
      "content": "The methodological quality assessment results are summarised in Figure 20 and study \nestimates of sensitivity and specificity are presented in Figure 18 for detection of \ndisseminated tuberculosis in blood. \n \nFigure 20. Risk of bias and applicability summary for each study: disseminated \ntuberculosis in blood \n \nShould LC-aNAATs be used to diagnose disseminated tuberculosis in blood in adults and \nadolescents with signs and symptoms of disseminated tuberculosis, against a microbiological \nreference standard? \nLC-aNAAT sensitivity and specificity (95% CI) were 38% and 98.0% from one Xpert Ultr",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that one study of Xpert Ultra on blood for disseminated TB found 38% sensitivity and 98% specificity. It introduces subgroup analyses comparing concentrated and unconcentrated cerebrospinal fluid.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-120",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 120",
      "pageStart": 120,
      "pageEnd": 120,
      "target": "../sections/page-120.md",
      "narrative": null,
      "content": "Table 3. Impact of concentration on diagnostic accuracy of LC-aNAATs for \nCSF \nSubgroup \nStudies \n(specimens) \nSummary \nsensitivity (95% \nCI) \nSummary \nspecificity (95% \nCI) \nMicrobiological reference standard \nConcentrated CSF \n5 (781) \n92.8 (87.5 to 96.0) \n93.6 (70.8 to 98.9) \nUnconcentrated \nCSF  \n6 (470) \n81.0 (68.0 to 89.5) \n85.8 (68.9 to 94.3) \n \nComposite reference standard \nConcentrated CSF  \n5 (828) \n70.0 (64.7 to 74.7) \n99.9 (81.8 to 100.0) \nUnconcentrated \nCSF  \n3 (228) \n50.9 (41.7 to 60.0) \n100.0 (97.4 to 100.0) \nSummary estimates of sensitivity and specificity are presented as per",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 3 shows higher Xpert Ultra sensitivity on concentrated than unconcentrated cerebrospinal fluid. A further table lists starting volumes, concentration and accuracy by study.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-121",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 121",
      "pageStart": 121,
      "pageEnd": 121,
      "target": "../sections/page-121.md",
      "narrative": null,
      "content": "Impact of concentration and starting volume of pleural fluid on accuracy \nThere were insufficient data to perform meta-analyses for pleural fluid by concentration. \nTable 5 provides the available information for concentration and starting volume from \nstudies.  \n \nTable 5. Starting volumes and concentration for pleural fluid and diagnostic \naccuracy \nStudy, year \nPleural fluid \nvolume \nConcentration Sensitivity (95% \nCI) \nSpecificity (95% \nCI) \nMeldau 2019* \n1 mL  \nYes \nUses CRS* \nMinnies 2023 \n20 mL \nYes \n100 (74 – 100) \n87 (76 – 94) \nPeñata 2021 \n> 2 mL \nYes \n100 (29 – 100) \n93 (86 – 98) \nWa",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes there were too few data to meta-analyse pleural fluid by concentration and tabulates starting volumes and accuracy by study. It then introduces results for detection of rifampicin resistance.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-005",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 5. Starting volumes and concentration for pleural fluid and diagnostic accuracy",
      "pageStart": 121,
      "pageEnd": 121,
      "target": "../tables/table-5.csv",
      "narrative": "draft",
      "content": "Pleural fluid starting volume, use of a concentration step, and sensitivity and specificity (95% CI) reported by each study of LC-aNAATs on pleural fluid. One study used a composite reference standard and has no estimates in the table.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p122-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 122,
      "pageEnd": 122,
      "target": "../images/img-p122-1.webp",
      "narrative": "draft",
      "content": "Risk of bias and applicability chart for rifampicin resistance, with one row per study (16 studies). Each study is rated for risk of bias (patient selection, index test, reference standard, flow and timing) and applicability (patient selection, index test, reference standard) using green (low), yellow (unclear) and red (high) circles. Most ratings are low; high ratings are given for Hoel 2020a, Mekkaoui 2021, Slail 2023 (applicability, patient selection); Chin 2019, Wu 2019 (applicability, index test).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-vfig-p122",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 122,
      "pageEnd": 122,
      "target": "../figures/vfig-p122.png",
      "narrative": "draft",
      "content": "Table of pooled diagnostic accuracy for detection of rifampicin resistance: 13 studies, 446 (54) participants (cases), sensitivity 100.0 (93.4 to 100.0), specificity 99.4 (92.1 to 100.0), positive predictive value 94.6 (56.7 to 100.0) and negative predictive value 100.0 (99.2 to 100.0), all with 95% CI.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-122",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 122",
      "pageStart": 122,
      "pageEnd": 122,
      "target": "../sections/page-122.md",
      "narrative": null,
      "content": "Figure 22. Risk of bias and applicability summary for each study: rifampicin \nresistance \n \nShould LC-aNAAT in extrapulmonary specimens be used to diagnose rifampicin resistance in \nadults and adolescents with presumed extrapulmonary tuberculosis? \nLC-aNAAT summary sensitivity and specificity (95% CI) were 100.0% (93.4 to 100.0) and \n99.4% (92.1 to 100.0) from 13 Xpert Ultra studies (446 participants; Table 6); high certainty \n(sensitivity and specificity) evidence.  \n \nTable 6. Diagnostic accuracy of LC-aNAATs for detection of rifampicin \nresistance in adults and adolescents \nNumber \nof \nstud",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that Xpert Ultra on extrapulmonary specimens detected rifampicin resistance with 100% sensitivity and 99.4% specificity in 13 studies, at high certainty.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-006",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 6. Diagnostic accuracy of LC-aNAATs for detection of rifampicin resistance in adults and adolescents",
      "pageStart": 122,
      "pageEnd": 122,
      "target": "../tables/table-6.csv",
      "narrative": "draft",
      "content": "Pooled diagnostic accuracy of LC-aNAATs for detecting rifampicin resistance in extrapulmonary specimens from adults and adolescents: number of studies and participants (cases), sensitivity, specificity and positive and negative predictive values with 95% CIs.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p123-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 123,
      "pageEnd": 123,
      "target": "../images/img-p123-1.webp",
      "narrative": "draft",
      "content": "Paired forest plots of sensitivity and specificity for detection of rifampicin resistance by LC-aNAATs (Figure 23); 16 studies. For each study the table lists TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as squares with confidence-interval lines on scales from 0 to 1. Sensitivity is 1.00 in every study where it is estimable, with lower confidence limits ranging from 0.03 to 0.72. Sensitivity is not estimable for Chin 2019, Mekkaoui 2021, Spener-Gomes 2021. Specificity ranges from 0.98 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-123",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 123",
      "pageStart": 123,
      "pageEnd": 123,
      "target": "../sections/page-123.md",
      "narrative": null,
      "content": "Figure 23. LC-aNAATs for detection of rifampicin resistance \nFN: false negative; FP: false positive; TN: true negative; TP: true positive.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A figure caption for forest plots of LC-aNAAT accuracy for detecting rifampicin resistance.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-124",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 124",
      "pageStart": 124,
      "pageEnd": 124,
      "target": "../sections/page-124.md",
      "narrative": null,
      "content": "Discussion \nThe sensitivity and specificity of LC-aNAATs for detection of extrapulmonary tuberculosis \nvaried with site of disease and specimen type. Based on a microbiological reference standard, \nsensitivity was lowest for detection of peritoneal tuberculosis using peritoneal fluid with a \nsummary sensitivity (95% CI) of 50.0% (20.0 to 80.0) and summary specificity (95% CI) of \n98.4% (89.3 to 99.8). Conversely, sensitivity was highest for detection of lymph node \ntuberculosis using lymph node tissue with a summary sensitivity (95% CI) of 96.3% (89.1 to \n98.8) and summary specificity (95% CI)",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The discussion notes that accuracy varied by disease site and specimen, with sensitivity lowest for peritoneal fluid and highest for lymph node tissue against a microbiological reference. It calls for studies of other LC-aNAATs and notes high-certainty evidence of very high accuracy for rifampicin resistance.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-125",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 125",
      "pageStart": 125,
      "pageEnd": 125,
      "target": "../sections/page-125.md",
      "narrative": null,
      "content": "References \nBioneer Corporation. IRON-qPCR™. 2023. Available at: eng.bioneer.com/20-a-2600.html \n(accessed 19 September 2023). \nChu H, Cole SR. Bivariate meta-analysis of sensitivity and specificity with sparse data: a \ngeneralized linear mixed model approach. J Clin Epidemiol. 2006;59(12):1331-2. \nGlobal TB Report, 2023. WHO, Geneva, Switzerland  \nKohli M, Schiller I, Dendukuri N, Yao M, Dheda K, Denkinger CM, et al. Xpert MTB/RIF \nUltra and Xpert MTB/RIF assays for extrapulmonary tuberculosis and rifampicin resistance \nin adults. Cochrane Database Syst Rev. 2021;1:CD012768. \nReitsma JB, Glas",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Begins the review's reference list, citing manufacturer information, meta-analysis methods, the Global TB Report and a Cochrane review, followed by included studies of Xpert Ultra.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-126",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 126",
      "pageStart": 126,
      "pageEnd": 126,
      "target": "../sections/page-126.md",
      "narrative": null,
      "content": "tuberculosis to detect and measure Mycobacterium tuberculosis blood stream infection: a \ndiagnostic and disease biomarker cohort study. Lancet Microbe 2022;3(7):e521-e532. [DOI: \ndoi: 10.1016/S2666-5247(22)00062-3] \nChin JH, Musubire AK, Morgan N, Pellinen J, Grossman S, Bhatt JM, et al. Xpert MTB/RIF \nUltra for detection of Mycobacterium tuberculosis in cerebrospinal fluid. Journal of Clinical \nMicrobiology 2019;57(6):pii: e00249-19. [CRS ID: 16196410] \nChristopher DJ, Coelho V, Ebby GS, Shankar D, Gupta R, Thangakunam B. Incremental \nyield of Xpert® MTB/RIF Ultra over Xpert® MTB/RIF in the d",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list of included studies on Xpert Ultra and Truenat MTB Plus for extrapulmonary TB in specimens such as cerebrospinal fluid, blood and pleural fluid.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-127",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 127",
      "pageStart": 127,
      "pageEnd": 127,
      "target": "../sections/page-127.md",
      "narrative": null,
      "content": "tuberculosis. Afr J Thorac Crit Care Med 2019;31. [DOI: 10.7196/AJTCCM.2019.v25i2.011] \nMansfield M, McLaughlin AM, Roycroft E, Montgomery L, Keane J, Fitzgibbon MM, \nRogers TR. Diagnostic Performance of Xpert MTB/RIF Ultra Compared with Predecessor \nTest, Xpert MTB/RIF, in a Low TB Incidence Setting: a Retrospective Service Evaluation. . \nMicrobiol Spectr 2022;10(3):e0234521. [DOI: 10.1128/spectrum.02345-21] \nMekkaoui L, Hallin M, Mouchet M , Payen M, Maillart E, Clevenbergh P, GeorgalaI A,  Van \nden Wijngaert S. Performance of Xpert MTB/RIF Ultra for diagnosis of pulmonary and \nextra-pulmona",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list with included diagnostic accuracy studies of Xpert Ultra from various countries and settings.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-128",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 128",
      "pageStart": 128,
      "pageEnd": 128,
      "target": "../sections/page-128.md",
      "narrative": null,
      "content": "Sharma Aman. Determining the diagnostic potential of Truenat MTB Plus for Tubercular \nlymphadenitis and detection of drug resistance and a comparison with GeneXpert Ultra. \nTuberculosis 2023/09/01;142:102379. [DOI: https://doi.org/10.1016/j.tube.2023.102379] \nSlail Mousa J, Booq Rayan Y, Al-Ahmad Ibrahim H, Alharbi Arwa A, Alharbi Shafi F, \nAlotaibi Mutlaq Z, Aljubran Abdulaziz M, Aldossary Ahmad M, Memish Ziad A, Alyamani \nEssam J, Tawfik Essam A, Binjomah Abdulwahab Z. Evaluation of Xpert MTB/RIF Ultra for \nthe Diagnosis of Extrapulmonary Tuberculosis: A Retrospective Analysis in Saudi Arabi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Ends the reference list with studies comparing Truenat MTB Plus and Xpert Ultra and evaluating Xpert Ultra in Saudi Arabia, China and elsewhere.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-129",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 129",
      "pageStart": 129,
      "pageEnd": 129,
      "target": "../sections/page-129.md",
      "narrative": null,
      "content": "Appendix \nAppendix 1. Detailed search strategies \n \nOvid MEDLINE(R) ALL <1946 to October 10, 2023> \n1 \nMycobacterium tuberculosis/ \n \n2 \nTuberculosis/ \n \n3 \nTuberculosis, Multidrug-Resistant/ \n \n4 \nExtensively Drug-Resistant Tuberculosis/ \n \n5 \n(Tuberculosis or TB or MDR TB or XDR TB or \"Multidrug Resistant Tuberculosis\" or \n\"Extensively Drug Resistant Tuberculosis\" or tuberculous).tw. \n \n6 \n(extrapulmonary or extra pulmonary or EPTB).tw. \n \n7 \n(lymphadenitis or disseminated or miliary or pleur* or skeletal or spine or mening* or \nintracranial or intra ocular or ocular or abdominal or splenic",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Appendix 1 gives the detailed literature search strategy, beginning with Ovid MEDLINE and Embase search terms for extrapulmonary TB and the index tests.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-130",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 130",
      "pageStart": 130,
      "pageEnd": 130,
      "target": "../sections/page-130.md",
      "narrative": null,
      "content": "7 \n(lymphadenitis or disseminated or miliary or pleur* or skeletal or spine or mening* or \nintracranial or intra ocular or ocular or abdominal or splenic or genitourinary or pericardial).ti,ab.\n \n \n8 \nextrapulmonary tuberculosis/  \n9 \n1 or 2 or 3 or 4 or 5 \n \n10 \n6 or 7  \n11 \n9 and 10 \n \n12 \n8 or 11  \n13 \nXpert*.tw. and Ultra.mp. \n14 \n(GeneXpert or cepheid).tw. and ultra.mp.  \n15 \nnear* patient.tw. and Ultra.mp.  \n16 \n13 or 14 or 15  \n17 \n12 and 16 \n \n18 \n(Truenat* or Molbio or TruRif).mp. \n \n19 \n12 and 18 \n \n20 \n(STANDARD M10 or SD Biosensor or IRON-qPCR or BIONEER).mp. \n \n21 \n12 and 20",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the search strategies with remaining Embase lines and the Web of Science and Cochrane Library searches.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-131",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 131",
      "pageStart": 131,
      "pageEnd": 131,
      "target": "../sections/page-131.md",
      "narrative": null,
      "content": "#4 \n((lymphadenitis or disseminated or miliary or pleur* or skeletal or spine or mening* or \nintracranial or intra ocular or ocular or abdominal or splenic or genitourinary or pericardial)):ti,ab,kw \n#5 \n#3 and #4 \n#6 \nMeSH descriptor: [Tuberculosis, Extrapulmonary] explode all trees \n \n#7          EPTB or extrapulmonary tuberculosis \n \n#8 \n#5 or #6 or #7 \n#9 \n(Xpert* or GeneXpert* or cepheid) and Ultra \n \n#10 \n#8 and #9 \n \n#11 \nTruenat* or Molbio or TruRif \n \n#12 \n#8 and #11 \n \n#13 \n(STANDARD M10 or SD Biosensor or IRON-qPCR or BIONEER) \n \n#14 \n#8 and #13  \n \nGlobal Index Medicus \ntw:((tw:(ex",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the search strategies for the Cochrane Library, Global Index Medicus, Scopus and ClinicalTrials.gov.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-132",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 132",
      "pageStart": 132,
      "pageEnd": 132,
      "target": "../sections/page-132.md",
      "narrative": null,
      "content": "(STANDARD M10 or SD Biosensor or IRON-qPCR or BIONEER) | Extrapulmonary Tuberculosis   \n \nWHO ICTRP \ntuberculosis and extrapulmonary and  (Xpert* or truenat or STANDARD M10 or SD Biosensor or \nIRON-qPCR or NAAT*) \n \nProquest Dissertation and Thesis \n S1 title(tuberculosis) AND title (extrapulmonary) AND (STANDARD M10 or SD Biosensor or IRON-\nqPCR or BIONEER)  \n S2 title(tuberculosis) AND title(extrapulmonary) AND Truenat  \n S3 title(tuberculosis) AND title(extrapulmonary) AND (Xpert Ultra)",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Ends the search strategies with the WHO trials registry and ProQuest dissertation searches.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-133",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 133",
      "pageStart": 133,
      "pageEnd": 133,
      "target": "../sections/page-133.md",
      "narrative": null,
      "content": "128\nAppendix 2. Summary of findings tables \n \n1. Should LC-aNAAT on CSF be used to diagnose TB meningitis in adults and adolescents with signs and symptoms \nof TB meningitis, MRS? \nPatient or population: adults and adolescents with signs and symptoms of TB meningitis, MRS \nPooled sensitivity:0.88 (95% CI: 0.84 to 0.92)| Pooled specificity:0.96 (95% CI: 0.87 to 0.99) \nTest result \nNumber of results per 1,000 patients tested (95% CI) \nNumber of \nparticipants  \n(studies) \nCertainty of the \nEvidence (GRADE) \nPrevalence2.5% \nTypically seen in \nPrevalence10% \nTypically seen in \nPrevalence20% \nTypica",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Appendix 2 begins the GRADE summary of findings tables, showing Xpert Ultra on cerebrospinal fluid for TB meningitis against a microbiological reference, with results per 1,000 people at different prevalences. It explains downgrading for imprecision.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-134",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 134",
      "pageStart": 134,
      "pageEnd": 134,
      "target": "../sections/page-134.md",
      "narrative": null,
      "content": "129\n2. Should LC-aNAAT on CSF be used to diagnose TB meningitis in adults and adolescents with signs and symptoms \nof TB meningitis, CRS? \nPatient or population: adults and adolescents with signs and symptoms of TB meningitis, CRS \nPooled sensitivity:0.60 (95% CI: 0.51 to 0.69)| Pooled specificity:0.99 (95% CI: 0.98 to 1.00) \nTest result \nNumber of results per 1,000 patients tested (95% CI) \nNumber of \nparticipants  \n(studies) \nCertainty of the \nEvidence (GRADE) \nPrevalence2.5% \nTypically seen in \nPrevalence10% \nTypically seen in \nPrevalence20% \nTypically seen in \nTrue positives \n15 (13 to 17)",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A GRADE summary of findings table for Xpert Ultra on cerebrospinal fluid for TB meningitis against a composite reference standard, with pooled sensitivity of 0.60. It explains downgrading for inconsistency given widely varying sensitivities.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-135",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 135",
      "pageStart": 135,
      "pageEnd": 135,
      "target": "../sections/page-135.md",
      "narrative": null,
      "content": "130\n3. Should LC-aNAAT on CSF be used to diagnose TB meningitis in PLHIV with signs and symptoms of TB \nmeningitis, MRS? \nPatient or population: PLHIV with signs and symptoms of TB meningitis, MRS \nPooled sensitivity:0.89 (95% CI: 0.77 to 0.95) | Pooled specificity:0.91 (95% CI: 0.87 to 0.93) \nTest result \nNumber of results per 1,000 patients tested (95% CI) \nNumber of \nparticipants  \n(studies) \nCertainty of the \nEvidence (GRADE) \nPrevalence5% \nTypically seen in \nPrevalence20% \nTypically seen in \nPrevalence40% \nTypically seen in \nTrue positives \n44 (38 to 47) \n177 (153 to 189) \n354 (306 to 379",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A GRADE summary of findings table for Xpert Ultra on cerebrospinal fluid for TB meningitis in people living with HIV, with pooled sensitivity of 0.89 and specificity of 0.91. Footnotes explain judgements on imprecision.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-136",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 136",
      "pageStart": 136,
      "pageEnd": 136,
      "target": "../sections/page-136.md",
      "narrative": null,
      "content": "131\n4. Should LC-aNAAT on lymph node aspirate be used to diagnose lymph node TB in adults and adolescents with \nsigns and symptoms of lymph node TB, MRS? \nPatient or population: adults and adolescents with signs and symptoms of lymph node TB, MRS \nPooled sensitivity:0.85 (95% CI: 0.73 to 0.92)| Pooled specificity:0.74 (95% CI: 0.64 to 0.82) \nTest result \nNumber of results per 1,000 patients tested (95% CI) \nNumber of \nparticipants  \n(studies) \nCertainty of the \nEvidence (GRADE) \nPrevalence2.5% \nTypically seen in \nPrevalence10% \nTypically seen in \nPrevalence20% \nTypically seen in \nTrue positive",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A GRADE summary of findings table for Xpert Ultra on lymph node aspirate against a microbiological reference, with pooled sensitivity of 0.85 and specificity of 0.74. Footnotes explain downgrading for risk of bias and imprecision.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-137",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 137",
      "pageStart": 137,
      "pageEnd": 137,
      "target": "../sections/page-137.md",
      "narrative": null,
      "content": "132\nassumed. We downgraded two levels for imprecision.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation line from the previous table's footnotes on downgrading for imprecision.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-138",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 138",
      "pageStart": 138,
      "pageEnd": 138,
      "target": "../sections/page-138.md",
      "narrative": null,
      "content": "133\n5. Should LC-aNAAT on lymph node aspirate be used to diagnose lymph node TB in adults and adolescents with \nsigns and symptoms of lymph node TB, CRS? \nPatient or population: adults and adolescents with signs and symptoms of lymph node TB, CRS \nPooled sensitivity:0.71 (95% CI: 0.64 to 0.77)|Pooled specificity:0.97 (95% CI: 0.82 to 1.00) \nTest result \nNumber of results per 1,000 patients tested (95% CI) \nNumber of \nparticipants  \n(studies) \nCertainty of the \nEvidence (GRADE) \nPrevalence2.5% \nTypically seen in \nPrevalence10% \nTypically seen in \nPrevalence20% \nTypically seen in \nTrue positives",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A GRADE summary of findings table for Xpert Ultra on lymph node aspirate against a composite reference, with pooled sensitivity of 0.71 and specificity of 0.97. Footnotes explain downgrading for inconsistency and imprecision.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-139",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 139",
      "pageStart": 139,
      "pageEnd": 139,
      "target": "../sections/page-139.md",
      "narrative": null,
      "content": "134",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A page bearing only a page number.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-140",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 140",
      "pageStart": 140,
      "pageEnd": 140,
      "target": "../sections/page-140.md",
      "narrative": null,
      "content": "135\n6. Should LC-aNAAT on lymph node tissue be used to diagnose lymph node TB in adults and adolescents with signs \nand symptoms of lymph node TB, MRS? \nPatient or population: adults and adolescents with signs and symptoms of lymph node TB, MRS \nPooled sensitivity:0.96 (95% CI: 0.85 to 0.99)|Pooled specificity:0.79 (95% CI: 0.65 to 0.89) \nTest result \nNumber of results per 1,000 patients tested (95% CI) \nNumber of \nparticipants  \n(studies) \nCertainty of the \nEvidence (GRADE) \nPrevalence2.5% \nTypically seen in \nPrevalence10% \nTypically seen in \nPrevalence20% \nTypically seen in \nTrue positives",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A GRADE summary of findings table for Xpert Ultra on lymph node tissue against a microbiological reference, with pooled sensitivity of 0.96 and specificity of 0.79. Footnotes explain downgrading for inconsistency and imprecision.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-141",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 141",
      "pageStart": 141,
      "pageEnd": 141,
      "target": "../sections/page-141.md",
      "narrative": null,
      "content": "136\n7. Should LC-aNAAT on lymph node tissue be used to diagnose lymph node TB in adults and adolescents with signs \nand symptoms of lymph node TB, CRS? \nPatient or population: adults and adolescents with signs and symptoms of lymph node TB, CRS \nPooled sensitivity:0.61 (95% CI: 0.47 to 0.74)| Pooled specificity:0.97 (95% CI: 0.92 to 0.99) \nTest result \nNumber of results per 1,000 patients tested (95% CI) \nNumber of \nparticipants  \n(studies) \nCertainty of the \nEvidence (GRADE) \nPrevalence2.5% \nTypically seen in \nPrevalence10% \nTypically seen in \nPrevalence20% \nTypically seen in \nTrue positives",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A GRADE summary of findings table for Xpert Ultra on lymph node tissue against a composite reference, with pooled sensitivity of 0.61 and specificity of 0.97. It begins the next table on pleural fluid.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-142",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 142",
      "pageStart": 142,
      "pageEnd": 142,
      "target": "../sections/page-142.md",
      "narrative": null,
      "content": "137\nsymptoms of pleural TB, MRS? \nPatient or population: adults and adolescents with signs and symptoms of pleural TB, MRS \nPooled sensitivity:0.74 (95% CI: 0.61 to 0.84) |Pooled specificity:0.88 (95% CI: 0.79 to 0.94) \nTest result \nNumber of results per 1,000 patients tested (95% CI) \nNumber of \nparticipants  \n(studies) \nCertainty of the \nEvidence (GRADE) \nPrevalence2.5% \nTypically seen in \nPrevalence10% \nTypically seen in \nPrevalence20% \nTypically seen in \nTrue positives \n19 (15 to 21) \n74 (61 to 84) \n148 (122 to 168) \n264 \n(13) \n⨁⨁◯◯ \nLowa,b \nFalse negatives \n6 (4 to 10) \n26 (16 to 39) \n52",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A GRADE summary of findings table for Xpert Ultra on pleural fluid against a microbiological reference, with pooled sensitivity of 0.74 and specificity of 0.88. Footnotes explain downgrading, and the next table on the composite reference begins.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-143",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 143",
      "pageStart": 143,
      "pageEnd": 143,
      "target": "../sections/page-143.md",
      "narrative": null,
      "content": "138\nsymptoms of pleural TB, CRS? \nPatient or population: adults and adolescents with signs and symptoms of pleural TB, CRS  \nPooled sensitivity:0.44 (95% CI: 0.33 to 0.55)| Pooled specificity:0.99 (95% CI: 0.95 to 1.00) \nTest result \nNumber of results per 1,000 patients tested (95% CI) \nNumber of \nparticipants  \n(studies) \nCertainty of the \nEvidence (GRADE) \nPrevalence2.5% \nTypically seen in \nPrevalence10% \nTypically seen in \nPrevalence20% \nTypically seen in \nTrue positives \n11 (8 to 14) \n44 (33 to 55) \n87 (66 to 110) \n298 \n(7) \n⨁⨁◯◯ \nLowa,b \nFalse negatives \n14 (11 to 17) \n56 (45 to 67) \n113",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A GRADE summary of findings table for Xpert Ultra on pleural fluid against a composite reference, with pooled sensitivity of 0.44 and specificity of 0.99. Footnotes explain downgrading for imprecision and risk of bias.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-144",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 144",
      "pageStart": 144,
      "pageEnd": 144,
      "target": "../sections/page-144.md",
      "narrative": null,
      "content": "139\n10. Should LC-aNAAT on pleural tissue be used to diagnose pleural TB in adults and adolescents with signs and \nsymptoms of pleural TB, MRS? \nPatient or population: adults and adolescents with signs and symptoms of pleural TB, MRS \nRange of sensitivities:0.80 to 1.00| Range of specificities:0.75 to 0.86 \nTest result \nNumber of results per 1,000 patients tested (95% CI) \nNumber of \nparticipants  \n(studies) \nCertainty of the \nEvidence (GRADE) \nPrevalence2.5% \nTypically seen in \nPrevalence10% \nTypically seen in \nPrevalence20% \nTypically seen in \nTrue positives \n20 to 25 \n80 to 100 \n160 to 200",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A GRADE summary of findings table for Xpert Ultra on pleural tissue against a microbiological reference, with sensitivities of 0.80–1.00 from two studies. It begins the next table on the composite reference.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-145",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 145",
      "pageStart": 145,
      "pageEnd": 145,
      "target": "../sections/page-145.md",
      "narrative": null,
      "content": "140\nsymptoms of pleural TB, CRS? \nPatient or population: adults and adolescents with signs and symptoms of pleural TB, CRS \nRange of sensitivities:0.54 to 0.81| Range of specificities:0.94 to 0.97 \nTest result \nNumber of results per 1,000 patients tested (95% CI) \nNumber of \nparticipants  \n(studies) \nCertainty of the \nEvidence (GRADE) \nPrevalence2.5% \nTypically seen in \nPrevalence10% \nTypically seen in \nPrevalence20% \nTypically seen in \nTrue positives \n14 to 20 \n54 to 81 \n108 to 162 \n53 \n(2) \n⨁⨁◯◯ \nLowa,b \nFalse negatives \n5 to 11 \n19 to 46 \n38 to 92 \nTrue negatives \n917 to 946 \n846 to 873 \n75",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A GRADE summary of findings table for Xpert Ultra on pleural tissue against a composite reference, with sensitivities of 0.54–0.81. It begins the table on urine for genitourinary TB.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-146",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 146",
      "pageStart": 146,
      "pageEnd": 146,
      "target": "../sections/page-146.md",
      "narrative": null,
      "content": "141\nPatient or population: adults and adolescents with signs and symptoms of genitourinary TB, MRS \nRange of sensitivities:1.00 to 1.00| Range of specificities:0.28 to 1.00 \nTest result \nNumber of results per 1,000 patients tested (95% CI) \nNumber of \nparticipants  \n(studies) \nCertainty of the \nEvidence (GRADE) \nPrevalence2.5% \nTypically seen in \nPrevalence10% \nTypically seen in \nPrevalence20% \nTypically seen in \nTrue positives \n25 to 25 \n100 to 100 \n200 to 200 \n14 \n(3) \n⨁◯◯◯ \nVery lowa,b \nFalse negatives \n0 to 0 \n0 to 0 \n0 to 0 \nTrue negatives \n273 to 975 \n252 to 900 \n224 to 800 \n218 \n(6) \n⨁◯",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A GRADE summary of findings table for Xpert Ultra on urine against a microbiological reference, with 100% sensitivity but specificity ranging from 0.28 to 1.00, at very low certainty. It begins the composite reference table.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-147",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 147",
      "pageStart": 147,
      "pageEnd": 147,
      "target": "../sections/page-147.md",
      "narrative": null,
      "content": "142\nPooled sensitivity:0.23 (95% CI: 0.15 to 0.34)| Pooled specificity:0.99 (95% CI: 0.90 to 1.00) \nTest result \nNumber of results per 1,000 patients tested (95% CI) \nNumber of \nparticipants  \n(studies) \nCertainty of the \nEvidence (GRADE) \nPrevalence2.5% \nTypically seen in \nPrevalence10% \nTypically seen in \nPrevalence20% \nTypically seen in \nTrue positives \n6 (4 to 9) \n23 (15 to 34) \n46 (29 to 68) \n159 \n(3) \n⨁⨁◯◯ \nLowa,b \nFalse negatives \n19 (16 to 21) \n77 (66 to 85) \n154 (132 to 171) \nTrue negatives \n964 (875 to 974) \n890 (807 to 899) \n791 (718 to 799) \n534 \n(3) \n⨁⨁⨁◯ \nModerateb \nFalse positiv",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A GRADE summary of findings table for Xpert Ultra on urine against a composite reference, with pooled sensitivity of 0.23 and specificity of 0.99. It begins the table on bone or joint TB.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-148",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 148",
      "pageStart": 148,
      "pageEnd": 148,
      "target": "../sections/page-148.md",
      "narrative": null,
      "content": "143\nTest result \nNumber of results per 1,000 patients tested (95% CI) \nNumber of \nparticipants  \n(studies) \nCertainty of the \nEvidence (GRADE) \nPrevalence2.5% \nTypically seen in \nPrevalence10% \nTypically seen in \nPrevalence20% \nTypically seen in \nTrue positives \n24 (22 to 25) \n97 (87 to 99) \n193 (174 to 198) \n58 \n(3) \n⨁⨁◯◯ \nLowa,b \nFalse negatives \n1 (0 to 3) \n3 (1 to 13) \n7 (2 to 26) \nTrue negatives \n888 (788 to 938) \n820 (727 to 866) \n729 (646 to 770) \n68 \n(3) \n⨁⨁◯◯ \nLowa,c \nFalse positives \n87 (37 to 187) \n80 (34 to 173) \n71 (30 to 154) \nCI: confidence interval \nExplanations \na. For indirec",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A GRADE summary of findings table for Xpert Ultra on bone or joint fluid against a microbiological reference, with footnotes on indirectness and imprecision. It begins the composite reference table based on a single study.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-149",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 149",
      "pageStart": 149,
      "pageEnd": 149,
      "target": "../sections/page-149.md",
      "narrative": null,
      "content": "144\nTest result \nNumber of results per 1,000 patients tested (95% CI) \nNumber of \nparticipants  \n(studies) \nCertainty of the \nEvidence (GRADE) \nPrevalence2.5% \nTypically seen in \nPrevalence10% \nTypically seen in \nPrevalence20% \nTypically seen in \nTrue positives \n24 (23 to 25) \n96 (91 to 99) \n192 (182 to 198) \n111 \n(1) \n⨁⨁⨁◯ \nModeratea \nFalse negatives \n1 (0 to 2) \n4 (1 to 9) \n8 (2 to 18) \nTrue negatives \n946 (829 to 975) \n873 (765 to 900) \n776 (680 to 800) \n34 \n(1) \n⨁◯◯◯ \nVery lowa,b \nFalse positives \n29 (0 to 146) \n27 (0 to 135) \n24 (0 to 120) \nCI: confidence interval \nExplanations \na. For in",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the GRADE table for bone or joint TB against a composite reference from a single Chinese study, noting it may not be representative. Footnotes explain downgrading for imprecision.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-150",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 150",
      "pageStart": 150,
      "pageEnd": 150,
      "target": "../sections/page-150.md",
      "narrative": null,
      "content": "145\n16. Should LC-aNAAT on peritoneal fluid be used to diagnose peritoneal TB in adults and adolescents with signs \nand symptoms of peritoneal TB, MRS? \nPatient or population: adults and adolescents with signs and symptoms of peritoneal TB, MRS \nRange of sensitivities:0.33 to 0.67|Range of specificities:0.94 to 1.00 \nTest result \nNumber of results per 1,000 patients tested (95% CI) \nNumber of \nparticipants  \n(studies) \nCertainty of the \nEvidence (GRADE) \nPrevalence2.5% \nTypically seen in \nPrevalence10% \nTypically seen in \nPrevalence20% \nTypically seen in \nTrue positives \n8 to 17 \n33 to 67 \n66",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A GRADE summary of findings table for Xpert Ultra on peritoneal fluid, with sensitivities of 0.33–0.67 and very low certainty. Footnotes explain downgrading for indirectness, inconsistency and imprecision.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-151",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 151",
      "pageStart": 151,
      "pageEnd": 151,
      "target": "../sections/page-151.md",
      "narrative": null,
      "content": "146\nimprecision.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of the previous table's footnote on imprecision.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-152",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 152",
      "pageStart": 152,
      "pageEnd": 152,
      "target": "../sections/page-152.md",
      "narrative": null,
      "content": "147\n17. Should LC-aNAAT on pericardial fluid be used to diagnose pericardial TB in adults and adolescents with signs \nand symptoms of pericardial TB, MRS? \nPatient or population: adults and adolescents with signs and symptoms of pericardial TB, MRS \nPooled sensitivity:0.84 (95% CI: 0.74 to 0.91)| Pooled specificity:0.87 (95% CI: 0.80 to 0.92) \nTest result \nNumber of results per 1,000 patients tested (95% CI) \nNumber of \nparticipants  \n(studies) \nCertainty of the \nEvidence (GRADE) \nPrevalence2.5% \nTypically seen in \nPrevalence10% \nTypically seen in \nPrevalence20% \nTypically seen in \nTrue positi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A GRADE summary of findings table for Xpert Ultra on pericardial fluid, with pooled sensitivity of 0.84 and specificity of 0.87. Footnotes explain downgrading for indirectness and imprecision.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-153",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 153",
      "pageStart": 153,
      "pageEnd": 153,
      "target": "../sections/page-153.md",
      "narrative": null,
      "content": "148\n18. Should LC-aNAAT in blood be used to diagnose disseminated TB in adults and adolescents with signs and \nsymptoms of disseminated TB, MRS? \nPatient or population: adults and adolescents with signs and symptoms of disseminated TB, MRS \nSingle study sensitivity:0.38 (95% CI: 0.33 to 0.43)| Single study specificity:0.98 (95% CI: 0.94 to 1.00) \nTest result \nNumber of results per 1,000 patients tested (95% CI) \nNumber of \nparticipants  \n(studies) \nCertainty of the \nEvidence (GRADE) \nPrevalence2.5% \nTypically seen in \nPrevalence10% \nTypically seen in \nPrevalence20% \nTypically seen in \nTrue pos",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A GRADE summary of findings table for Xpert Ultra on blood for disseminated TB from a single study, with sensitivity of 0.38 and specificity of 0.98. It begins the table on rifampicin resistance.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-154",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 154",
      "pageStart": 154,
      "pageEnd": 154,
      "target": "../sections/page-154.md",
      "narrative": null,
      "content": "149\nsigns and symptoms of EPTB, MRS? \nPatient or population: adults and adolescents with signs and symptoms of EPTB, MRS \nPooled sensitivity:1.00 (95% CI: 0.93 to 1.00)|Pooled specificity:0.99 (95% CI: 0.92 to 1.00) \nTest result \nNumber of results per 1,000 patients tested (95% CI) \nNumber of \nparticipants  \n(studies) \nCertainty of the \nEvidence (GRADE) \nPrevalence2% \nTypically seen in \nPrevalence10% \nTypically seen in \nPrevalence15% \nTypically seen in \nTrue positives \n20 (19 to 20) \n100 (93 to 100) \n150 (140 to 150) \n54 \n(13) \n⨁⨁⨁⨁ \nHigh \nFalse negatives \n0 (0 to 1) \n0 (0 to 7) \n0 (0 to 10)",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A GRADE summary of findings table showing pooled sensitivity of 1.00 and specificity of 0.99 for detecting rifampicin resistance in extrapulmonary specimens, at high certainty.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-155",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 155",
      "pageStart": 155,
      "pageEnd": 155,
      "target": "../sections/page-155.md",
      "narrative": null,
      "content": "150\nSupplementary Figures and Tables \n \nSupplementary Figure 1. Study flow diagram for Xpert Ultra \n \nSupplementary Figure 2. Study flow diagram for Truenat MTB plus \n \na \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \nRecords identified from: \nDatabases (n =564) \n \nRecords removed before \nscreening: \nDuplicate records removed  \n(n = 344 ) \n \nRecords screened \n(n = 220) \nRecords excluded \n(n = 145) \nReports sought for retrieval \n(n =  75) \nReports not retrieved (n = 0) \nReports assessed for eligibility \n(n = 75) \nReports excluded: 52 \nDid not evaluate Xpert Ultra: 32 \nNo",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "PRISMA study flow diagrams for the Xpert Ultra and Truenat MTB Plus searches, showing records screened, excluded with reasons and included.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-156",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 156",
      "pageStart": 156,
      "pageEnd": 156,
      "target": "../sections/page-156.md",
      "narrative": null,
      "content": "151\n \nSupplementary Figure 3. Study flow diagram for STANDARD M10 and others \n* Manufacturer to discontinue Truenat MTB kits from the international market from 2024 \nFrom:  Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:n71. \ndoi: 10.1136/bmj.n71. For more information, visit: http://www.prisma-statement.org/ \nRecords identified from*: \nDatabases (n =1175) \n \nRecords removed before \nscreening: \nDuplicate records removed  \n(n = 524) \n \nRecords screened \n(n = 651) \nRecords e",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A PRISMA flow diagram for STANDARD M10 and other tests, with a note that the manufacturer will discontinue Truenat MTB kits internationally from 2024.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-157",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 157",
      "pageStart": 157,
      "pageEnd": 157,
      "target": "../sections/page-157.md",
      "narrative": null,
      "content": "152\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \nFrom:  Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:n71. \ndoi: 10.1136/bmj.n71. For more information, visit: http://www.prisma-statement.org/ \n \nRecords identified from*: \nDatabases (n =37) \n \nRecords removed before \nscreening: \nDuplicate records removed  \n(n = 3) \n \nRecords screened \n(n = 34) \nRecords excluded** \n(n =32) \nReports sought for retrieval \n(n = 2) \nReports not retrieved \n(n = 0) \nReports",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A PRISMA flow diagram showing that no studies were included for one index test search after two reports were excluded for addressing pulmonary TB.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-158",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 158",
      "pageStart": 158,
      "pageEnd": 158,
      "target": "../sections/page-158.md",
      "narrative": null,
      "content": "Supplementary Table 1: Certainty of evidence for all forms of EPTB.  \n \nEPTB \nspecimen \nReference \nstandard \n# studies \nParticipants \n(cases) \nSummary \nsensitivity \n(95% CI) \nSummary \nspecificity \n(95% CI) \nEffect Estimates \nLow \nMedium \nCerebrospinal \nfluid \nMRS \n16 \n1687  \n(287) \n  \n88.2  \n(83.7 to 91.6) \n96.0 \n(86.8 to 98.9) \nTP: 22(21-23) \nFN: 3 (2-4)  \nTN:  936 (846-\n964) \nFP:  39 (11-129) \nTP: 88 (84-92) \nFN: 12(8-16) \nTN: 864(781-\n890) \nFP: 36(10-119) \nT\nF\nT\n7\nF\nCerebrospinal \nfluid \nCRS \n10 \n1397 \n(753) \n60.3  \n(50.9 to 69.0) \n99.2  \n(98.1 to 99.7) \nTP: 15(13-17) \nFN: 10 (8-12) \nTN: 96",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Supplementary Table 1 gives certainty of evidence and effect estimates per 1,000 people for each specimen type, beginning with cerebrospinal fluid and lymph node fluid. The extracted table text is jumbled.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-159",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 159",
      "pageStart": 159,
      "pageEnd": 159,
      "target": "../sections/page-159.md",
      "narrative": null,
      "content": "Lymph node \nfluid \nCRS \n6 \n461 \n(243) \n71.3  \n(64.3 to 77.4) \n97.4  \n(82.2 to 99.7) \nTP: 18 (16-19) \nFN: 7 (6-9) \nTN: 950 (801- \n972) \nFP: 25 (3-174) \nTP: 71 (64-77) \nFN: 29 (23-36) \nTN: 877 (740-\n897) \nFP: 23 (3-160) \nT\nF\nT\n7\nF\nLymph node \ntissue \nMRS \n8 \n578 \n(97) \n96.5  \n(84.7 to 99.3) \n79.4  \n(65.4 to 88.8) \nTP: 24 (21-25) \nFN: 1 (0-4) \nTN: 774 (638-\n866) \nFP: 201 (109-\n337) \nTP: 97 (85-99) \nFN: 3 (1-15) \nTN: 715 (589-\n799) \nFP: 185 (101-\n311) \nT\n1\nF\nT\n7\nF\nLymph node \ntissue \nCRS \n3 \n229 \n(109) \n61.5  \n(47.1 to 74.2) \n96.7  \n(91.5 to 98.7) \nTP: 15 (12-19) \nFN: 10 (6-13) \nTN: 943 (892-\n962)",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the certainty of evidence table for lymph node, pleural fluid and pleural tissue specimens. The extracted table text is jumbled.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-160",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 160",
      "pageStart": 160,
      "pageEnd": 160,
      "target": "../sections/page-160.md",
      "narrative": null,
      "content": "105 \n(53) \nFN: 5-11 \nTN: 917- 946 \nFP: 29-58 \nFN: 19-46 \nTN: 846- 873 \nFP: 27-54 \nF\nT\nF\nUrine \nMRS \n6 \n232 \n(14) \n100% \n28% to 100% \nTP: 25-25 \nFN: 0- 0 \nTN: 273-975 \nFP: 0-702 \nTP: 100-100 \nFN: 0-0 \nTN: 252-900 \nFP: 0- 648 \nT\nF\nT\nF\nUrine \nCRS \n3 \n693 \n(159) \n23.0  \n(14.7 to 34.1) \n98.9  \n(89.7 to 99.9) \nTP: 6 (4-9) \nFN: 19 ( 16-21) \nTN: 964 (875-\n974) \nFP: 11 (1-100) \nTP: 23 ( 15-34) \nFN: 77 (66-85) \nTN: 890 (807- \n899) \nFP: 10 (1-93) \nT\nF\n1\nT\n7\nF\nSynovial fluid \nMRS \n3 \n126 \n(58) \n96.6  \n(87.2 to 99.1) \n91.1  \n(80.8 to 96.2) \nTP: 24 ( 22- 25) \nFN: 1 (0-3) \nTN: 888 (788- \n938) \nFP: 87 (37-187",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the certainty of evidence table for urine, synovial, peritoneal and pericardial fluid specimens. The extracted table text is jumbled.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-161",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 161",
      "pageStart": 161,
      "pageEnd": 161,
      "target": "../sections/page-161.md",
      "narrative": null,
      "content": "FP: 131 (83-\n200) \nF\nBlood \nMRS \n1 \n578 \n(423) \n \n38%  \n(33% to 43%) \n98%  \n(94% to \n100%) \nTP: 10 (8-11) \nFN: 15 (14-17) \nTN: 956 (917- \n975) \nFP: 19 (0-58) \nTP: 38 (33-43) \nFN: 62 (57-67) \nTN: 882 (846-\n900) \nFP: 18 (0-54)  \nT\nF\n1\nT\n8\nF\nRIF resistance MRS \n13  \n446 \n(54) \n100.0  \n(93.4 to \n100.0) \n99.4  \n(92.1 to \n100.0) \nTP: 20 (19-20) \nFN: 0 (0-1) \nTN: 974 (903-\n980) \nFP: 6 (0-77) \nTP: 100 (93-100) \nFN: 0 (0-7) \nTN: 895 (829-\n900) \nFP: 5 (0-71) \nT\n1\nF\nT\n8\nF",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the certainty of evidence table with blood and rifampicin resistance results.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-162",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 162",
      "pageStart": 162,
      "pageEnd": 162,
      "target": "../sections/page-162.md",
      "narrative": null,
      "content": "Web Annex B.4. Impact of Rapid Nucleic Acid Amplification \nTests for Tuberculosis on Patient Outcomes: A Systematic \nReview \n \nDate: 15 April 2024 \nReview authors and affiliations: \nEleanor Ochodo (Review Lead), Centre for Global Health Research, Kenya Medical Research \nInstitute (KEMRI), Kenya/Centre for Evidence-based Health Care, Stellenbosch University, \nSouth Africa \nLisa Were, Centre for Global Health Research, KEMRI, Kenya \nJenifer Otieno, Centre for Global Health Research, KEMRI, Kenya \nVittoria Lutje, Cochrane Infectious Diseases Group Information Specialist, LSTM, United \nKingdom \nKa",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title page of Web Annex B.4, a 2024 systematic review on the impact of rapid nucleic acid amplification tests for TB on patient outcomes, listing its authors and their affiliations. It was prepared for a WHO guideline meeting in May 2024.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-163",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 163",
      "pageStart": 163,
      "pageEnd": 163,
      "target": "../sections/page-163.md",
      "narrative": null,
      "content": "List of Abbreviations \ncRCT \nCI \nCluster Randomized Controlled Trial \nConfidence Interval \nDR-TB \nDrug-resistant Tuberculosis \nDS-TB \nDrug-susceptible Tuberculosis \nEPTB \nExtrapulmonary Tuberculosis \nFQ \nFluoroquinolone  \nGDG \nGuideline Development Group \nGDT \nGuideline Development Tool \nGRADE \nGrading of Recommendations, Assessment, Development, and Evaluations \nICC \nInterclass Correlation Coefficient  \niRCT \nIndividual Randomized Controlled Trial \nITT \nIntention to Treat \nLAMP \nLoop‐mediated isothermal amplification \nLC-aNAATs \nLow complexity-automated Nucleic Acid Amplification Tests \nLC-mN",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A list of abbreviations used in the review of rapid nucleic acid amplification tests and patient outcomes.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-164",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 164",
      "pageStart": 164,
      "pageEnd": 164,
      "target": "../sections/page-164.md",
      "narrative": null,
      "content": "Executive Summary \n \nBackground \nWHO recommends Rapid nucleic acid amplification tests (NAATs) as the initial diagnostic test \nfor tuberculosis (TB), some of which can detect drug resistance simultaneously. They have \nhigher diagnostic accuracy than smear microscopy and have fast turnaround times that may \nlead to a positive impact on patient outcomes. \n \nObjective \nTo assess the impact of low complexity (LC) and moderate complexity (MC) NAATs compared \nto smear microscopy or culture on patient outcomes in people investigated for TB. \n \nMethods \nWe searched several electronic databases from in",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The executive summary explains that WHO recommends rapid NAATs as initial TB tests and that the review assessed their impact on patient outcomes compared with smear microscopy or culture. It reports trial results on time to diagnosis and treatment, with moderate certainty.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-165",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 165",
      "pageStart": 165,
      "pageEnd": 165,
      "target": "../sections/page-165.md",
      "narrative": null,
      "content": "Conclusions \nThe use of Xpert Ultra may decrease the time to TB diagnosis and improve the time to TB \ntreatment initiation. However, we found that compared to smear microscopy or culture, \nthere is limited evidence with moderate certainty of evidence to confidently conclude \nthat Xpert Ultra has an impact on patient-important outcomes.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Concludes that Xpert Ultra may shorten time to diagnosis and treatment but that evidence of its effect on patient-important outcomes is limited.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-166",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 166",
      "pageStart": 166,
      "pageEnd": 166,
      "target": "../sections/page-166.md",
      "narrative": null,
      "content": "Background \nDescription of the condition \nTB is one of the leading causes of death worldwide and affects all ages (1). TB is caused by \nMycobacterium tuberculosis (M. tuberculosis) bacteria. TB primarily affects the respiratory \nsystem causing pulmonary TB (PTB) but can also affect other organ systems in the form of \nextrapulmonary TB (EPTB). While TB is preventable and largely curable, an estimated 10.6 \nmillion people fell ill with TB worldwide in 2022 and around 1.3 million people died of TB (2, \n3). \n \nPulmonary TB \nPTB is the most prevalent form of TB. Loss of appetite, weight loss, letha",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives background on TB, noting 10.6 million cases and about 1.3 million deaths in 2022, and describes pulmonary and extrapulmonary forms. It introduces drug-resistant TB as a major global health security threat.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-167",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 167",
      "pageStart": 167,
      "pageEnd": 167,
      "target": "../sections/page-167.md",
      "narrative": null,
      "content": "and testing for drug resistance using rapid molecular tests, culture methods or sequencing \ntechnologies. Resistance to rifampicin—the most effective first-line drug—is of greatest \nconcern. Resistance to rifampicin and isoniazid is defined as multidrug-resistant TB (MDR-TB).  \n \nBoth MDR-TB and rifampicin-resistant TB (RR-TB) require treatment with second-line drugs \n(3). WHO uses five categories to classify cases of DR-TB: isoniazid-resistant TB, RR-TB, MDR-\nTB, extensively drug-resistant TB (XDR-TB) and pre-XDR-TB. \nThe definitions for DR-TB have been revised in response to advances in TB t",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains detection and classification of drug-resistant TB, including rifampicin-resistant, MDR, pre-XDR and XDR-TB. It then lists the classes of NAATs reviewed, such as Xpert Ultra, Truenat, STANDARD M10 and IRON-qPCR.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-168",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 168",
      "pageStart": 168,
      "pageEnd": 168,
      "target": "../sections/page-168.md",
      "narrative": null,
      "content": "● LC-mNAATs: TB-LAMP \n● MC-aNAATs: FluoroType XDR-TB \nLC-aNAATs \nXpert MTB/RIF Ultra, Cepheid, USA \nFor ease of reference, we refer to Xpert MTB/RIF Ultra as Xpert Ultra. Xpert Ultra is a NAAT \nthat functions as an automated closed system that performs real-time polymerase chain \nreaction (PCR). Xpert Ultra can be used for diagnosing TB and RR-TB. Xpert Ultra can be \nperformed by operators with minimal technical expertise (11). For the detection of PTB, a \nCochrane review reported a summary sensitivity and specificity (95% credible interval (CrI)) \nof Xpert Ultra against culture to be 90.9% (8",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists further NAAT classes, including TB-LAMP and FluoroType XDR-TB, and describes Xpert Ultra as an automated PCR system usable with minimal expertise, citing its reported accuracy. It also describes Truenat assays and their possible results.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-169",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 169",
      "pageStart": 169,
      "pageEnd": 169,
      "target": "../sections/page-169.md",
      "narrative": null,
      "content": "• RIF Resistance DETECTED \n• Indeterminate or ERROR \n \nSTANDARD M10, SD Biosensor, Republic of Korea \nThe STANDARD M10 is an in vitro molecular diagnosis system that automatically extracts and \namplifies nucleic acids (DNA, RNA) from the collected samples, and analyses the data from \namplification in real-time, by combining the processes of isothermal amplification and real-\ntime PCR in a single platform using all-in-one cartridge systems. There are two variations \ncurrently available for TB: STANDARD™ M10 MTB/NTM and STANDARD™ M10 MDR-TB. \nThese assays are intended for use with the STANDARD M",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes result categories of rifampicin resistance testing and the STANDARD M10 and IRON-qPCR molecular platforms, as part of the index test descriptions for a review of NAAT impact on patient outcomes.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-170",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 170",
      "pageStart": 170,
      "pageEnd": 170,
      "target": "../sections/page-170.md",
      "narrative": null,
      "content": "qPCR ™ has a high multiplexing capacity of up to 40 targets in a single test and provides results \nin 40 minutes. The IRON qPCR RFIA kit is a cartridge-based assay that detects M. tuberculosis \nand resistance to 11 drugs: rifampicin, isoniazid, fluoroquinolones (levofloxacin, moxifloxacin, \ngatifloxacin and ofloxacin), aminoglycosides (kanamycin, amikacin, capreomycin), \nprothionamide and ethionamide. It uses IRON qPCR, which is a single instrument for \nextraction, amplification and detection. Two modules are available per instrument. IRON \nqPCR RFIA is not commercially available. \n \nLC-mNAATs",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the IRON-qPCR kit for detecting TB and resistance to 11 drugs, then explains how rapid near-patient NAATs could improve outcomes by cutting loss to follow-up before diagnosis and treatment.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-171",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 171",
      "pageStart": 171,
      "pageEnd": 171,
      "target": "../sections/page-171.md",
      "narrative": null,
      "content": "perhaps, reduce time to treatment. Pre‐treatment loss to follow‐up (LTFU) is associated with \nhigh mortality rates, and thus, its reduction may lead to reduced mortality (22). A larger \nproportion of people with true TB may receive effective therapy, and fewer individuals may \nbe falsely diagnosed with the disease on clinical grounds alone, and incorrectly treated. Rapid \ndetection of drug resistance may lead to more rapid initiation of effective treatment. These \nfactors could potentially improve important patient outcomes. \n \nWhy it was important to do this review \nA Cochrane Review assessed",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Argues that reducing pre-treatment loss to follow-up may lower mortality and that rapid drug resistance detection could speed appropriate therapy, and states the review's objective of assessing NAAT impact on patient outcomes.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-172",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 172",
      "pageStart": 172,
      "pageEnd": 172,
      "target": "../sections/page-172.md",
      "narrative": null,
      "content": "people being investigated for TB.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A fragment completing the review objective, referring to people being investigated for TB.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-173",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 173",
      "pageStart": 173,
      "pageEnd": 173,
      "target": "../sections/page-173.md",
      "narrative": null,
      "content": "Methods \nThis review was written following the recommended methodology described in the Cochrane \nHandbook for Systematic Reviews of Interventions (23) and the PRISMA 2020 statement (24). \nThe protocol for the review is available on the Open Source Repository. \n \nCriteria for considering studies for this review \n \nTypes of studies \nWe included cluster and individually randomised controlled trials (RCTs) that met our \ninclusion criteria and provided data on the outcomes of interest.  \n \nTypes of participants \nWe included people with presumptive TB. Based on the WHO definition of age groups, we",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Sets out the methods for the review of NAAT impact, following Cochrane and PRISMA guidance, including randomized trials and primary outcomes of all-cause mortality and cure, with loss to follow-up and time to diagnosis as secondary outcomes.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-174",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 174",
      "pageStart": 174,
      "pageEnd": 174,
      "target": "../sections/page-174.md",
      "narrative": null,
      "content": "• Time to treatment \nFor all outcomes, we used the WHO definitions outlined in Table 1 below. The WHO consulted \non the definitions of DR-TB treatment outcomes in 2020. Following this consultation, the \nWHO proposed new definitions of TB treatment outcomes that apply to the assessment of \ntreatment outcomes of all forms of TB (25). \n \nTable 1: TB treatment outcome deﬁnitions for both drug‐resistant and drug‐susceptible TB \nPatient-outcome \nDefinition \nCure \nA pulmonary TB patient with bacteriologically confirmed TB at the \nbeginning of treatment who completed treatment as recommended \nby the n",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Adds time to treatment as an outcome and presents WHO's updated TB treatment outcome definitions in a table, then begins describing the October 2023 literature searches.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-175",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 175",
      "pageStart": 175,
      "pageEnd": 175,
      "target": "../sections/page-175.md",
      "narrative": null,
      "content": "Cochrane Central Register of Controlled Trials (CENTRAL), using the Cochrane Library, issue \n10, October 2023; MEDLINE (Ovid); Embase (Ovid); Science Citation Index and Biosis previews \n(ISI Web of Knowledge); WHO Global Index Medicus; and SCOPUS (Elsevier). VL also searched \nClinicalTrials.gov (clinicaltrials.gov/) and the WHO Clinical Trials Registry Platform \n(www.who.int/trialsearch) to identify ongoing trials. \n \nSearching other resources \nWe examined the reference lists of included articles and relevant review articles identified \nthrough the electronic searches. VL also searched for rel",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists the databases and trial registries searched and describes study selection and independent data extraction by two review authors.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-176",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 176",
      "pageStart": 176,
      "pageEnd": 176,
      "target": "../sections/page-176.md",
      "narrative": null,
      "content": "participant characteristics, intervention, control, each outcome measured and how it was \nmeasured, covariates outcomes were adjusted for, length of follow‐up, and measure of \neffect, with standard errors or 95% CIs. For binary outcomes, we extracted the 2x2 data or \nrelative measures of effect (risk or odds ratios) if available. For time‐to‐event outcomes, we \nextracted hazard ratios if available. In addition, for cluster RCTs (cRCTs), we extracted the \nnumber of participants and clusters randomised to each diagnostic arm, the number of \nparticipants monitored for each outcome of interest, an",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the data extracted for binary and time-to-event outcomes, adjustments for cluster trials and risk-of-bias assessment, and introduces GRADE assessment of certainty.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-177",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 177",
      "pageStart": 177,
      "pageEnd": 177,
      "target": "../sections/page-177.md",
      "narrative": null,
      "content": "important outcome as described by Balshem 2011 (31). Drawing on valid reasons we \ndowngraded the level of certainty in the evidence within the following five domains: risk of \nbias, inconsistency, indirectness, imprecision and publication bias. The following criteria were \napplied to interpret the findings: \n● High: we were very confident that the true effect lies close to that of the estimate of \nthe effect. \n● Moderate: we were moderately confident in the effect estimate; the true effect is \nlikely to be close to the estimate of the effect, but there is a possibility that it is \nsubstantiall",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains the GRADE domains and certainty levels used, then reports that two trials, one individually randomized and one cluster randomized, met the inclusion criteria.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-vfig-p178",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 178,
      "pageEnd": 178,
      "target": "../figures/vfig-p178.png",
      "narrative": "draft",
      "content": "PRISMA 2020 flow chart of study selection. 1067 records from databases and registers; 6 duplicates and 1 record marked ineligible by automation tools removed; 1060 screened and 991 excluded; 69 reports assessed and 67 excluded (ineligible index test 21, study design 20, study type 9, population 7, duplicates 5, inaccessible full texts 3, ineligible comparator 2). One report from a WHO public call was assessed and excluded (ineligible comparator). 2 studies were included in the review.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-178",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 178",
      "pageStart": 178,
      "pageEnd": 178,
      "target": "../sections/page-178.md",
      "narrative": null,
      "content": "173\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \nFigure 1: The PRISMA 2020 ﬂow chart illustrating the study selection process (24) \n*Details are provided for these studies in Appendix 3. \nRecords identified from*: \nDatabases and registers (n \n=1067) \nRecords removed before \nscreening: \nDuplicate records removed \n(n = 6) \nRecords marked as \nineligible by automation \ntools (n = 1) \nRecords removed for other \nreasons (n = 0) \nRecords screened \n(n = 1060) \nRecords excluded** \n(n = 991) \nReports sought for retrieval \n(n = 69) \nReports not retrieved \n(n = 0) \nReports assessed for el",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The PRISMA flow chart for the patient outcomes review, showing records screened and two studies included.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-179",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 179",
      "pageStart": 179,
      "pageEnd": 179,
      "target": "../sections/page-179.md",
      "narrative": null,
      "content": "Characteristics of included studies \nThe characteristics of the two included RCTs are summarised in Table 2. Of note, no study \nreported all-cause mortality and cure as outcomes. \n \nCattamanchi 2021 (32) conducted a cRCT amongst adults aged ≥18 years in Ugandan \ncommunity health centres for TB diagnosis. They compared a multicomponent diagnostic \nstrategy (on-site Xpert Ultra testing, guided restructuring of clinic workflows, and monthly \nfeedback of quality metrics) to routine care (on-site sputum smear microscopy with referral-\nbased molecular testing). The trial included 10,644 patients acr",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the two included trials, a Ugandan cluster trial of a multicomponent Xpert Ultra strategy and a South African mobile-clinic trial comparing Xpert Ultra with smear microscopy, noting neither reported mortality or cure.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-180",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 180",
      "pageStart": 180,
      "pageEnd": 180,
      "target": "../sections/page-180.md",
      "narrative": null,
      "content": "175\nTable 2: Key characteristics of included studies \nAuthor, year  Study \ndesign \nCountry \n(WHO \nregion) \nFacility \n(geographical \narea) \nTarget population \nCondition \nIndex test \n(sample type) \nComparator test \n(sample type) \nOutcomes \nEsmail 2023 \n(33) \nParallel \niRCT with \n2 arms \n(single-\ncentre) \nSouth \nAfrica \n(African \nregion) \nCommunity-\nbased scalable \nmobile clinic \n(per-urban \narea) \n588 adults >18 \nyears, at least one \nTB symptom for 2 \nweeks or PLHIV \nirrespective of \nsymptoms \nTB \nXpert Ultra \n(sputum) \nSmear microscopy \n(TB culture was also \ndone on the second \nsputum sample)",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 2, giving key characteristics of the two included trials, such as design, country, setting, population, index and comparator tests, and outcomes.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-181",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 181",
      "pageStart": 181,
      "pageEnd": 181,
      "target": "../sections/page-181.md",
      "narrative": null,
      "content": "176\nTable 3: Outcome results \nAuthor, \nyear  \n(study \ndesign) \nType of \nanalysis & \nsample size \nLost-to-follow-up (n/N, %) \nPatient outcome results \nIntervention \narm \nComparison \narm  \nOverall (at \nstudy \nendpoint) \nPrimary \nSecondary \nEsmail 2023 \n(iRCT) (33) \nModified ITT \nN = 584 \nI = 288 \nC = 296 \nNot \nreported \nNot \nreported \nNot \nreported \nTime to TB treatment, days: \nSame-day and/or culture-based \ntreatment: \nHazard ratio: 2.3 (95% CI: 1.0−5.1), \nP=0.004 \nSame-day treatment (excluding \nculture-based treatment initiation): \nHazard ratio: 3.0 (95% CI: 1.2−7.7), \nP=0.002 \n \nNot reported",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 3, giving outcome results of the two trials, showing shorter times to TB diagnosis and treatment with Xpert Ultra.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p182-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 182,
      "pageEnd": 182,
      "target": "../images/img-p182-1.webp",
      "narrative": "draft",
      "content": "Risk of bias summary for an individually randomized trial (Figure 2): Esmail 2023, Xpert Ultra versus smear microscopy, outcome time to treatment, weight 596. All domains D1–D5 (randomisation process, deviations from the intended interventions, missing outcome data, measurement of the outcome, selection of the reported result) and the overall judgement are rated low risk (green).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p182-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 182,
      "pageEnd": 182,
      "target": "../images/img-p182-2.webp",
      "narrative": "draft",
      "content": "Risk of bias summary for a cluster-randomized trial (Figure 3): Cattamanchi 2021, Xpert Ultra versus smear microscopy, outcomes time to diagnosis and time to treatment, weight 10644. For both outcomes, D1a (randomisation process) and D3 (missing outcome data) are low risk, while D1b (timing of identification or recruitment of participants), D2 (deviations from the intended interventions), D4 (measurement of the outcome), D5 (selection of the reported result) and the overall judgement show some concerns (yellow).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-182",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 182",
      "pageStart": 182,
      "pageEnd": 182,
      "target": "../sections/page-182.md",
      "narrative": null,
      "content": "Methodological quality assessment \nFigure 2 summarises the results of the risk of bias assessment for one outcome from an iRCT. \nThere was a low risk of bias across all domains for Esmail 2023 for the time to treatment \noutcome (33).  \n \nFigure 2: Summary of risk of bias assessment for iRCT per outcome \n \nFor Cattamanchi 2021 there were some risk of bias concerns for both outcomes because \nparticipants were screened for recruitment after randomisation (Figure 3) (32) . In addition, \nblinding was not feasible due to the nature of the intervention. No information was provided \nregarding deviatio",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarizes risk of bias, judged low for the South African trial and of some concern for the Ugandan cluster trial owing to recruitment and unspecified analyses.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-vfig-p183",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 183,
      "pageEnd": 183,
      "target": "../figures/vfig-p183.png",
      "narrative": "draft",
      "content": "Page excerpt with text and a forest plot (Figure 4) of the adjusted geometric mean ratio for time to TB diagnosis. One study, Cattamanchi 2021 (N = 10664), has an estimate of 0.49 (95% CI 0.39 to 0.62), lying left of the line of no effect at 1, on the side labelled 'Favours Xpert Ultra'.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-183",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 183",
      "pageStart": 183,
      "pageEnd": 183,
      "target": "../sections/page-183.md",
      "narrative": null,
      "content": "Findings of the synthesis \nPrimary outcomes \nBoth studies did not report our primary outcomes of all-cause mortality and cure.  \n \nSecondary outcomes \nBoth studies did not report LTFU. \n \nTime to TB diagnosis \nCattamanchi 2021 (32) was the only study to report time to TB diagnosis and calculated an \nadjusted geometric mean ratio (0.49, 95% CI: 0.39−0.62) (Figure 4). The adjusted geometric \nmean ratio suggested that the mean time to TB diagnosis was less for those who followed the \nmulticomponent diagnostic strategy (which included Xpert Ultra), compared to routine care \n(smear microscopy).",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that neither trial measured mortality, cure or loss to follow-up, and presents evidence that Xpert Ultra shortened time to diagnosis and to treatment compared with smear microscopy.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-vfig-p184",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 184,
      "pageEnd": 184,
      "target": "../figures/vfig-p184.png",
      "narrative": "draft",
      "content": "Page excerpt with text and two forest plots. Figure 5 shows the adjusted geometric mean ratio for time to TB treatment initiation in Cattamanchi 2021 (N = 10644): 0.35 (95% CI 0.21 to 0.56), favouring Xpert Ultra. Figure 6 shows hazard ratios for time to TB treatment initiation in Esmail 2023 (N = 584): 2.30 (1.00 to 5.10) including culture and 3.00 (1.20 to 7.70) excluding culture, on the side labelled 'Favours Xpert Ultra'.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-184",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 184",
      "pageStart": 184,
      "pageEnd": 184,
      "target": "../sections/page-184.md",
      "narrative": null,
      "content": "1.2−7.7) (33). Therefore, it appears the use of Xpert Ultra improves the time to TB treatment \ninitiation when compared to smear microscopy, regardless of whether treatment was \ninitiated due to a positive TB culture. \n \nFigure 5: Forest plot showing adjusted geometric mean ratio for time to TB treatment \ninitiation \n(N – number of participants included for the outcome) \n \n \nFigure 6: Forest plot showing hazard ratios for time to TB treatment initiation \n(N – number of participants included for the outcome) \n \nCertainty of evidence \nWe rated the certainty of evidence moderate for all outcomes",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Concludes that Xpert Ultra appears to shorten time to treatment initiation compared with smear microscopy, and notes that certainty of evidence was rated moderate for all outcomes.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-185",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 185",
      "pageStart": 185,
      "pageEnd": 185,
      "target": "../sections/page-185.md",
      "narrative": null,
      "content": "Table 4: GRADE table for certainty of the evidence \n \nCertainty assessment \n№ of patients \nEffect \nCertainty \nImportanc\ne \n№ of \nstudie\ns \nStudy \ndesign \nRisk of \nbias \nInconsistenc\ny \nIndirectnes\ns \nImprecisio\nn \nOther \nconsideration\ns \nLow and \nModerate \nComplexit\ny NAATS \nSmear \nmicroscopy/cultu\nre \nRelativ\ne \n(95% \nCI) \nAbsolut\ne \n(95% CI) \nTime to diagnosis \n11 \nrandomise\nd trials \nseriousa \nnot serious \nnot serious \nnot \nseriousb \nnone \nThe adjusted geometric mean ratio for time to diagnosis \n(1 trial, 10,644 participants) was 0.49, (95% CI: 0.39-\n0.62). \n⨁⨁\n⨁◯ \nModerate \nTime to treatme",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The GRADE table giving certainty of evidence for time to diagnosis and treatment with low and moderate complexity NAATs, followed by references to the two included trials.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-186",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 186",
      "pageStart": 186,
      "pageEnd": 186,
      "target": "../sections/page-186.md",
      "narrative": null,
      "content": "Discussion \nSummary of findings \nThis review assessed the impact of LC-aNAATs, LC-mNAATs and MC-aNAATs on patient \noutcomes in people being investigated for TB. We included two RCTs published between 2021 \nand 2023 (one iRCT and one pragmatic cRCT) assessing the impact of Xpert Ultra, a type of \nLC-aNAAT (32, 33). Both studies involved adult participants.  \n \nData were available for two of the three secondary outcomes of interest and none for the \ntwo primary outcomes (all-cause mortality and cure). Time to diagnosis was reduced for \nthose tested using Xpert Ultra compared to routine care (adj",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Discusses the findings of the review, which included two trials of Xpert Ultra in adults reporting secondary outcomes only, along with its strengths and limitation that no trials of other NAAT classes were found.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-187",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 187",
      "pageStart": 187,
      "pageEnd": 187,
      "target": "../sections/page-187.md",
      "narrative": null,
      "content": "Implications for future research \nThere is a need for high-quality trials that assess the impact of LC-aNAATs, LC-mNAATs and \nMC-aNAATs against relevant comparators on important patient outcomes to confidently \ninform policy decisions.  Studies included in this review were conducted within the WHO \nAfrican region. Evidence from other WHO regions with high TB incidence such as South-East \nAsia and Western Pacific regions will help generalise findings on the impact of LC-aNAATs (38). \nResearch on the impact of the three classes of NAATs on all-cause mortality and cure is \nneeded. \n \nConclusion",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Calls for high-quality trials of NAAT impact on patient outcomes, including outside the African region, and concludes that evidence that Xpert Ultra improves diagnosis and treatment times is limited.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-188",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 188",
      "pageStart": 188,
      "pageEnd": 188,
      "target": "../sections/page-188.md",
      "narrative": null,
      "content": "Appendices \nAppendix 1: Detailed search strategies \nOvid MEDLINE(R) ALL <1946 to October 16, 2023,> \n 1          Mycobacterium tuberculosis/ or tuberculosis.ti.  \n2          Tuberculosis, Pulmonary/ or Extensively Drug-Resistant Tuberculosis/ or Tuberculosis, \nMultidrug-Resistant/              \n3          (tuberculosis or TB).tw.  \n4          (extrapulmonary or lymph node* or mening* or pulmonary).ti. or (extrapulmonary or \nlymph node* or mening* or pulmonary).ab.      \n \n5          3 and 4 \n6          exp Tuberculosis, Extrapulmonary/        \n \n7          1 or 2 or 5 or 6  \n8          (Xpert*",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Appendix 1, the start of the detailed MEDLINE search strategy for the patient outcomes review.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-189",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 189",
      "pageStart": 189,
      "pageEnd": 189,
      "target": "../sections/page-189.md",
      "narrative": null,
      "content": "28        23 or 24 or 25 or 26 \n29        7 and 18            \n30        27 and 29          \n31        30 and 28          \n32        Follow Up Studies/ or cohort. mp.          \n \n33        (before and after studies).mp. \n \n34        exp Cohort Studies/     \n35        32 or 33 or 34  \n36        30 and 35 \n37        36 or 31             \nOvid Embase 1947-Present, updated daily \n 1          Mycobacterium tuberculosis/ or tuberculosis.ti.  \n2          Lung Tuberculosis/ or extensively drug resistant tuberculosis/ or drug resistant \ntuberculosis/       \n \n3          (tuberculosis or TB).tw.  \n4",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the search strategies, completing MEDLINE and beginning Embase.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-190",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 190",
      "pageStart": 190,
      "pageEnd": 190,
      "target": "../sections/page-190.md",
      "narrative": null,
      "content": "20        exp Survival/    \n21        Time-to-Treatment/ or exp Diagnostic Error/     \n \n22        18 or 19 or 20 or 21     \n23        8 or 9 or 10 or 11 or 12 or 13 or 14 or 15 or 16 or 17     \n \n24        7 and 22 and 23 \n25        (random* or factorial* or placebo* or assign* or allocat* or crossover*).tw.        \n26        ((blind* or mask*) and (single or double or triple or treble)).tw.   \n27        crossover procedure/  \n28        double blind procedure/ or single blind procedure/        \n \n29        randomization/ or placebo/      \n \n30        parallel design/ or Latin square design/",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the Embase search strategy and gives the Web of Science search.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-191",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 191",
      "pageStart": 191,
      "pageEnd": 191,
      "target": "../sections/page-191.md",
      "narrative": null,
      "content": "#7 Search: #5 OR #6     \n \n#8 Search: #7 AND #4  \nScopus (Elsevier) \n((TITLE-ABS-KEY (tuberculosis OR tb OR mdr-tb OR xdr-tb ) AND TITLE-ABS-KEY ( ( xpert* OR \ngenexpert* OR cepheid ) OR ( truenat* OR molbio OR trurif ) OR tb-lamp* OR loopamp* OR \nfluorotype* OR ( standard AND m10 OR sd AND biosensor OR iron-qpcr OR bioneer ) OR naat* \n) AND ABS ( treatment OR effect* OR outcome* OR mortality OR impact* OR survival OR \nerror* ) ) ) AND ( randomized OR controlled OR cohort OR time AND series ) \nCochrane Central Register of Controlled Trials: Issue 10 of 12, October 2023 \n#1        MeSH descript",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives the Scopus, regional database and trial registry search strategies for the patient outcomes review.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-192",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 192",
      "pageStart": 192,
      "pageEnd": 192,
      "target": "../sections/page-192.md",
      "narrative": null,
      "content": "Appendix 2: Table of characteristics of excluded studies \nStudy ID \nReason for exclusion \n1. Auld 2016 (39) \nDuplicate \n2. Lessells 2017 (40) \nDuplicate \n3. Durovni 2014 (41) \nDuplicate \n4. Theron 2014 (42) \nDuplicate \n5. Churchyard 2015 (43) \nDuplicate \n6. NCT02043080 2014 (44) \nInaccessible full text \n7. NCT03187964 2017 (45) \nInaccessible full text \n8. NCT03168945 2016 (46) \nInaccessible full text \n9. Lessells 2011 (40, 47) \nIneligible index test \n10. Durovni 2014 (41) \nIneligible index test \n11. Jayakumar 2016 (48) \nIneligible index test \n12. Bassett 2019 (49) \nIneligible index test \n13. P",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Appendix 2, the first part of a table of excluded studies, listing reasons such as duplicates, inaccessible full texts and ineligible index tests.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-193",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 193",
      "pageStart": 193,
      "pageEnd": 193,
      "target": "../sections/page-193.md",
      "narrative": null,
      "content": "30. Agizew 2019 (66) \nIneligible population \n31. Blanc 2020 (67) (68) \nIneligible population \n32. Ngwira 2019 (69) \nIneligible population \n33. Santos 2021 (70) \nIneligible population \n34. Auld 2020 (71) \nIneligible population \n35. Mupfumi 2014 (72) \nIneligible population \n36. Gupta-Wright 2018 (73) \nIneligible population \n37. VanRie 2013 (74) \nIneligible study design \n38. Alvarez 2015 (75) \nIneligible study design \n39. Cadavid 2022 (76) \nIneligible study design \n40. Williams 2022 (77) \nIneligible study design \n41. Van Den Handel 2015 (78) \nIneligible study design \n42. Schmidt 2017 (79) \nInelig",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the list of excluded studies, with reasons such as ineligible population, study design or study type.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-194",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 194",
      "pageStart": 194,
      "pageEnd": 194,
      "target": "../sections/page-194.md",
      "narrative": null,
      "content": "61. Lessells 2015 (98) \nIneligible study type \n62. Bonnet 2020 (99) \nIneligible study type \n63. Nathavitharana 2021 (100) \nIneligible study type \n64. Ochodo 2019 (101) \nIneligible study type \n65. Haraka 2021 (21) \nIneligible study type \n66. Marcy 2023 (35) \nIneligible comparator \n67. Mukoka 2023 (36) \nIneligible comparator \n68. Leukes 2024 (37) \nIneligible comparator",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Ends the list of excluded studies, including three trials excluded for having ineligible comparators.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-195",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 195",
      "pageStart": 195,
      "pageEnd": 195,
      "target": "../sections/page-195.md",
      "narrative": null,
      "content": "Appendix 3: Details of ineligible but relevant RCTs \nCharacteristics of excluded but relevant RCTs  \nThese have been summarised in Appendix Table 1 and 2 below. Marcy 2023 (35) included \n2,570 children aged <5 years in a stepped-wedge cRCT comparing the WHO standard of care \nfor severe pneumonia (n=1401) to Xpert Ultra (n=1,169). The trial was conducted in 16 \nhospitals across six countries in Africa. In the 12-week follow-up period, 91 (7.8%) children \ndied in the Xpert Ultra group and 110 (7.9%) died in the standard care group. Forty-one \nparticipants in the Xpert Ultra were lost to follow-u",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Appendix 3 describes three relevant but ineligible trials, including a multi-country stepped-wedge trial of Xpert Ultra in young children with severe pneumonia and a trial in people with HIV in Malawi.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-196",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 196",
      "pageStart": 196,
      "pageEnd": 196,
      "target": "../sections/page-196.md",
      "narrative": null,
      "content": "7 days of their first visit. The study also reported on the median time to treatment initiation.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A fragment ending a description of a trial outcome on starting treatment within seven days and median time to treatment.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-197",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 197",
      "pageStart": 197,
      "pageEnd": 197,
      "target": "../sections/page-197.md",
      "narrative": null,
      "content": "192\nAppendix Table 1: Key characteristics of ineligible but relevant RCTs \nAuthor, year  Study \ndesign \nCountry (WHO \nregion) \nFacility \n(geographical \narea) \nTarget \npopulation \nCondition \nIndex test \n(sample type) \nComparator \ntest (sample \ntype) \nOutcomes \nMarcy 2023 \n(35) \nPragmatic \nstepped \nwedge \ncRCT with \n2 arms \n(multi-\ncentre) \nCambodia \n(Western Pacific \nregion) \nCameroon, Côte \nd'Ivoire, \nMozambique, \nUganda, Zambia \n(African region) \n16 hospitals \n(area not \nreported) \n2,570 newly \nhospitalised \nchildren, 2-59 \nmonths, WHO-\ndefined severe \npneumonia \nincluding PLHIV \nSystematic",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Appendix Table 1, giving characteristics of the three ineligible but relevant trials, such as design, countries, settings, tests and outcomes.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-198",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 198",
      "pageStart": 198,
      "pageEnd": 198,
      "target": "../sections/page-198.md",
      "narrative": null,
      "content": "193\nusing smear \nmicroscopy \niRCT: Individual randomised controlled trial  \ncRCT: Cluster randomised controlled trial \nHIV: Human Immunodeficiency Virus \nPLHIV: people living with HIV \nWHO: World Health Organization \nN: Total sample size \nI: Intervention arm \nC: Comparison arm \n*Outcome not relevant to our review. \nAppendix Table 2: Outcome results for ineligible but relevant RCTs \nAuthor, year  \n(study design) \nType of \nanalysis & \nsample size \nLost-to-follow-up \nPatient outcome results \nIntervention arm (n/N, %) \nComparison arm \nOverall (at \nstudy \nendpoint) \nPrimary \nSecondary \n \nMukoka 202",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Appendix Table 2, outcome results for the ineligible trials, including odds ratios for mortality and hazard ratios for time to treatment in the paediatric pneumonia trial.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-199",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 199",
      "pageStart": 199,
      "pageEnd": 199,
      "target": "../sections/page-199.md",
      "narrative": null,
      "content": "194\nLeukes 2024 \n(37) \nModified \nintention to \ntreat \nN= 3983 \nI= 2003 \nC= 1980 \nNot reported \nNot reported \nNot \nreported \n*Not relevant \nTime to TB treatment \ninitiation:  \nCoxme Model Output for \npTB bacteriologically \nconfirmed population \ncensored at 60 days= Hazard \nratio: 6.0 (95% CI: Not \nreported) \nITT: Intention to treat analysis \n*Non-patient outcomes reported  \ni Results of the sensitivity analyses for the primary end point-deaths in ITT analysis. \n¬Two sputum samples were collected from each participant, of which one was tested using the test strategy for the intervention arm and",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the outcome results table with the Leukes 2024 cluster trial and footnotes on analyses and sputum testing.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-200",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 200",
      "pageStart": 200,
      "pageEnd": 200,
      "target": "../sections/page-200.md",
      "narrative": null,
      "content": "Findings \nAll-cause mortality \nAppendix Figure 1 shows the forest plot for all-cause mortality. Only Marcy 2023 reported all-\ncause mortality. Note that raw numbers were not used to calculate risk ratios because \ndividing numbers by the design effect is not an appropriate adjustment for results of stepped-\nwedge trials (21). Instead, the adjusted odds ratios as calculated by the study authors are \npresented. In the intention-to-treat analysis, there was no statistical evidence of an effect \n(0.99, 95% CI: 0.60−1.63). In the per-protocol analysis, Xpert Ultra reduced the odds of all-\ncause mort",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports findings from the ineligible trials, with only the paediatric pneumonia trial reporting all-cause mortality and a hazard ratio for time to treatment that the text interprets as favouring Xpert Ultra.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-201",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 201",
      "pageStart": 201,
      "pageEnd": 201,
      "target": "../sections/page-201.md",
      "narrative": null,
      "content": "Appendix Figure  2: Forest plot showing hazard ratio for time to TB treatment initiation \n(N – number of participants included for the outcome) \n \n \n \nFavours control \nFavours Xpert Ultra",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A caption for the forest plot of hazard ratios for time to TB treatment initiation.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-202",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 202",
      "pageStart": 202,
      "pageEnd": 202,
      "target": "../sections/page-202.md",
      "narrative": null,
      "content": "References \n1. \nBloom BR, Atun R, Cohen T, Dye C, Fraser H, Gomez GB, et al. Tuberculosis. In: Holmes KK, \nBertozzi S, Bloom BR, Jha P, editors. Major Infectious Diseases. Washington (DC): The International \nBank for Reconstruction and Development / The World Bank © 2017 International Bank for \nReconstruction and Development / The World Bank.; 2017. \n2. \nWorld Health Organisation. Global tuberculosis report 2023 2023 [cited 2024 27 March]. \nAvailable from: https://www.who.int/publications/i/item/9789240083851. \n3. \nWorld Health Organisation. Global Tuberculosis Report 2022. 2022a [Available fr",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The start of the reference list for the patient outcomes review, citing TB burden reports and diagnostic studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-203",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 203",
      "pageStart": 203,
      "pageEnd": 203,
      "target": "../sections/page-203.md",
      "narrative": null,
      "content": "18. \nFan L, Guan B, Cheng M, Liu C, Tian Y, Li R, et al. A Comprehensive Evaluation of a Loop-\nMediated Isothermal Amplification Assay for the Diagnosis of Pulmonary Tuberculosis in Children \nUsing Bronchoalveolar Lavage Fluid. Infect Drug Resist. 2022;15:975-87. \n19. \nIwamoto T, Sonobe T, Hayashi K. Loop-mediated isothermal amplification for direct \ndetection of Mycobacterium tuberculosis complex, M. avium, and M. intracellulare in sputum \nsamples. J Clin Microbiol. 2003;41(6):2616-22. \n20. \nJoon D, Nimesh M, Varma-Basil M, Saluja D. Evaluation of improved IS6110 LAMP assay for \ndiagnosis of",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list, citing studies of LAMP and other molecular assays and of TB detection trials.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-204",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 204",
      "pageStart": 204,
      "pageEnd": 204,
      "target": "../sections/page-204.md",
      "narrative": null,
      "content": "36. \nMukoka M, Twabi HH, Msefula C, Semphere R, Ndhlovu G, Lipenga T, et al. Utility of Xpert \nMTB/RIF Ultra and digital chest radiography for the diagnosis and treatment of TB in people living \nwith HIV: a randomised controlled trial (XACT-TB). Trans R Soc Trop Med Hyg. 2023;117(1):28-37. \n37. \nLeukes VN, Hella J, Sabi I, Cossa M, Khosa C, Erkosar B, et al. Study protocol: a pragmatic, \ncluster-randomized controlled trial to evaluate the effect of implementation of the Truenat \nplatform/MTB assays at primary health care clinics in Mozambique and Tanzania (TB-CAPT CORE). \nBMC Infect Dis. 2024;",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list, citing randomized trials of Xpert testing and case-finding strategies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-205",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 205",
      "pageStart": 205,
      "pageEnd": 205,
      "target": "../sections/page-205.md",
      "narrative": null,
      "content": "52. \nHanrahan CF, Nonyane BAS, Mmolawa L, West NS, Siwelana T, Lebina L, et al. Contact \ntracing versus facility-based screening for active TB case finding in rural South Africa: A pragmatic \ncluster-randomized trial (Kharitode TB). PLos Med. 2019;16(4):14. \n53. \nErshova JV, Volchenkov GV, Somova TR, Kuznetsova TA, Kaunetis NV, Kaminski D, et al. \nImpact of GeneXpert MTB/RIF R on treatment initiation and outcomes of RIF-resistant and RIF-\nsusceptible TB patients in Vladimir TB dispensary, Russia. BMC Infect Dis. 2020;20(1):543. \n54. \nIsrctn. Xpert for Tuberculosis: Evaluating a New Diagnostic",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list, citing trials of case-finding strategies and of Xpert testing in South Africa, Botswana and elsewhere.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-206",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 206",
      "pageStart": 206,
      "pageEnd": 206,
      "target": "../sections/page-206.md",
      "narrative": null,
      "content": "67. \nBlanc FX, Badje AD, Bonnet M, Gabillard D, Messou E, Muzoora C, et al. Systematic or Test-\nGuided Treatment for Tuberculosis in HIV-Infected Adults. N Engl J Med. 2020;382(25):2397-410. \n68. \nNct. Systematic Empirical vs. Test-guided Anti-TB Treatment Impact in Severely \nImmunosuppressed HIV-infected Adults Initiating ART With CD4 Cell Counts &lt;100/mm3. 2014. \n69. \nNgwira LG, Corbett EL, Khundi M, Barnes GL, Nkhoma A, Murowa M, et al. Screening for \nTuberculosis With Xpert MTB/RIF Assay Versus Fluorescent Microscopy Among Adults Newly \nDiagnosed With Human Immunodeficiency Virus in Rura",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list, citing trials of empirical versus test-guided TB treatment in people with HIV and studies of Xpert and late diagnosis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-207",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 207",
      "pageStart": 207,
      "pageEnd": 207,
      "target": "../sections/page-207.md",
      "narrative": null,
      "content": "84. \nMwansa-Kambafwile J, Maitshotlo B, Black A. Microbiologically Confirmed Tuberculosis: \nFactors Associated with Pre-Treatment Loss to Follow-Up, and Time to Treatment Initiation. PLoS \nONE. 2017;12(1):e0168659. \n85. \nChaisson LH, Roemer M, Cantu D, Haller B, Millman AJ, Cattamanchi A, et al. Impact of \nGeneXpert MTB/RIF assay on triage of respiratory isolation rooms for inpatients with presumed \ntuberculosis: a hypothetical trial. Clin Infect Dis. 2014;59(10):1353-60. \n86. \nBoyles TH. Why do clinical trials of Xpert® MTB/RIF fail to show an effect on patient relevant \noutcomes? Int J Tuber",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list, citing studies on Xpert impact, loss to follow-up and Cochrane reviews.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-208",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 208",
      "pageStart": 208,
      "pageEnd": 208,
      "target": "../sections/page-208.md",
      "narrative": null,
      "content": "101. \nOchodo EA, Kalema N, Schumacher S, Steingart K, Young T, Mallett S, et al. Variation in the \nobserved effect of Xpert MTB/RIF testing for tuberculosis on mortality: A systematic review and \nanalysis of trial design considerations. Wellcome Open Res. 2019;4:173.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The final reference in the patient outcomes review, a systematic review of Xpert MTB/RIF effects on mortality.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-209",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 209",
      "pageStart": 209,
      "pageEnd": 209,
      "target": "../sections/page-209.md",
      "narrative": null,
      "content": "Web Annex B.5. LC-mNAATs for detection of pulmonary and \nextrapulmonary tuberculosis in adults and adolescents: a \nsystematic review of diagnostic accuracy \n \nLeeberk Raja Inbaraj1, Jefferson Daniel2, Vignes Anand Srinivasalu1, Mukesh Kumar Sathyanarayanan1, \nAdhin Bhaskar1, Katie Scandrett3, Karen Steingart4 , Yemisi Takwoingi3  \n \n1. ICMR- National Institute for Research in Tuberculosis, Chennai, India \n2. Christian Medical College, Vellore, India \n3. Institute of Applied Health Research, University of Birmingham, United Kingdom \n4. Liverpool School of Tropical Medicine, United Kingdom",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title, authors and affiliations of Web Annex B.5, a systematic review of the diagnostic accuracy of low complexity manual NAATs (LC-mNAATs) for pulmonary and extrapulmonary TB in adults and adolescents.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-210",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 210",
      "pageStart": 210,
      "pageEnd": 210,
      "target": "../sections/page-210.md",
      "narrative": null,
      "content": "Background \nTuberculosis (TB) is one of the deadliest diseases with high mortality and morbidity. Globally, in 2022, \n7.5 million people were newly diagnosed with TB and 1.3 million died due to TB. (WHO Global TB \nReport 2023). One of the major obstacles to the elimination of TB is the delay in accurate diagnosis. \nOver the years, various rapid molecular tests have been recommended by WHO (mWRD) as the initial \ndiagnostic test for TB. Based on the technology, performance and complexity of implementation, the \nWHO categorized these tests into different classes (low, moderate, and high complexit",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives background for the LC-mNAAT review, citing 2022 TB diagnoses and deaths and the problem of diagnostic delay, and describes the TB-LAMP test as the class's index test. Methods follow a generic protocol shared by six diagnostic accuracy reviews.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-211",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 211",
      "pageStart": 211,
      "pageEnd": 211,
      "target": "../sections/page-211.md",
      "narrative": null,
      "content": "Search methods \nElectronic searches \nThe Cochrane Infectious Diseases Group Information Specialist (Vittoria Lutje (VL)) performed \nliterature searches between 2nd and 3rd October 2023, without language restrictions, in the following \ndatabases, using the search terms in Appendix 1: Cochrane Central Register of Controlled Trials \n(CENTRAL), included in the Cochrane Library, issue 10, October 2023; MEDLINE (Ovid); Embase (Ovid); \nScience Citation Index and Biosis previews (ISI Web of Knowledge); WHO Global Index Medicus; and \nSCOPUS (Elsevier). VL searched ClinicalTrials.gov (clinicaltrials.gov",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the October 2023 searches, eligibility criteria and culture reference standard for the LC-mNAAT review.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-212",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 212",
      "pageStart": 212,
      "pageEnd": 212,
      "target": "../sections/page-212.md",
      "narrative": null,
      "content": "of TB while a negative culture indicated no TB. \nComposite reference standard \nWe defined a composite reference standard as a positive culture or a clinical decision to initiate \ntreatment for TB (i.e., clinically diagnosed TB). The composite reference standard may be based on \nthe results of microbiological tests, culture or NAATs other than the index test being evaluated; \nimaging studies; histology; and clinical characteristics, and include at least one component test that is \npositive, according to the definition of the primary study authors. Clinical characteristics may include \ncough lon",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Defines the composite reference standard and outlines analysis methods, then reports that 29 studies were included, 27 on pulmonary TB and three on extrapulmonary TB.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-vfig-p213",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 213,
      "pageEnd": 213,
      "target": "../figures/vfig-p213.png",
      "narrative": "draft",
      "content": "PRISMA 2020 study flow diagram (Figure 1). 2806 database records; 1668 duplicates removed; 1138 screened and 988 excluded; 150 reports sought, 6 not retrieved, 144 assessed and 119 excluded (e.g. in-house kit used 29, case-control design 25, conference abstract 15, not a DTA study 13). From other methods (organisations 1, citation searching 2, WHO public call 4), 7 reports were assessed and 3 excluded. 29 reports and studies were included: PTB 27 and EPTB 3.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-213",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 213",
      "pageStart": 213,
      "pageEnd": 213,
      "target": "../sections/page-213.md",
      "narrative": null,
      "content": "208\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \nFigure 1. Study flow diagram \n*One study contributed data for both pulmonary TB (PTB) and extrapulmonary TB (EPTB). DTA = diagnostic test accuracy \nCharacteristics of included studies \nKey characteristics of the 27 included studies for detection of pulmonary TB and the three studies for detection of extrapulmonary TB are described in Table \nRecords identified from: \nDatabases (n = 2806) \nRegisters (n = 0) \nRecords removed before \nscreening: \nDuplicate records (n = 1668) \nRecords marked as ineligible \nby automation tools (n",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The study flow diagram for the LC-mNAAT review, with an introduction to the characteristics of included studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-214",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 214",
      "pageStart": 214,
      "pageEnd": 214,
      "target": "../sections/page-214.md",
      "narrative": null,
      "content": "209\n1. All the studies were cross-sectional and assessed TB-LAMP using a microbiological reference standard. No study used a composite reference standard. \n \nTable 1. Characteristics of included studies \nStudy \nCountry \nReference \nstandard \nProportion of \nPLHIV  \nClinical \nsetting \nLaboratory \ntype \nHigh TB \nburden \nSpecimen type \nProportion \nsmear positive \nparticipants \nBojang 2016 \nGambia \nMGIT \n2/50 (4%) \nOutpatient \nCentral \nNo \nExpectorated sputum \n104/300 (35%) \nCheng 2020 \nCambodia \nLJ & MGIT \n3/28 (11%) \nOutpatient \nPeripheral \nNo \nExpectorated sputum \n115/499 (23%) \nDanfack 2018 \nCam",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "States that all included studies were cross-sectional evaluations of TB-LAMP against culture, and begins Table 1 of study characteristics by country, HIV prevalence, setting and specimen.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-215",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 215",
      "pageStart": 215,
      "pageEnd": 215,
      "target": "../sections/page-215.md",
      "narrative": null,
      "content": "210\nStudy \nCountry \nReference \nstandard \nProportion of \nPLHIV  \nClinical \nsetting \nLaboratory \ntype \nHigh TB \nburden \nSpecimen type \nProportion \nsmear positive \nparticipants \nOu 2016 \nChina \nMGIT \nNot reported \nOutpatient \nCentral \nYes \nExpectorated sputum \n187/1519 (12%) \nOu 2019 \nChina \nLJ \nNot reported \nOutpatient \nCentral \nYes \nExpectorated sputum \n267/3126 (9%) \nPham 20183 \nPeru, South \nAfrica, Brazil, \nand Vietnam \nLJ/MGIT \n96/964 (10%) \nNot \nreported \nCentral \nYes \nExpectorated sputum \n216/964 (22%) \nPromsena 20221 \nThailand \nMGIT & \nOgawa \nNot reported \nInpatient \nCentral \nYes \nMultipl",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the table of TB-LAMP study characteristics, covering studies from China, multi-country sites and others, with abbreviations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-216",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 216",
      "pageStart": 216,
      "pageEnd": 216,
      "target": "../sections/page-216.md",
      "narrative": null,
      "content": "Detection of pulmonary tuberculosis  \nOf the 27 studies included, 62.9% (17/27) were conducted in high TB burden settings. One of the \nstudies (Getahun 2017) assessed only smear negative patients and so was not included in the main \nanalysis for detection of pulmonary TB. One study (Pham 2018) collected two sputum samples per \nparticipant and each sample was tested with TB-LAMP. The primary analysis reported by the study \nauthors was based on 964 participants and 1928 sputum samples. Raw data were not available from \nthe authors, so we extracted data for 976 participants from the previously pu",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the pulmonary TB studies, most in high-burden settings, and how multi-sample and smear-negative studies were handled, and summarizes risk of bias and applicability judgements.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p217-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 217,
      "pageEnd": 217,
      "target": "../images/img-p217-1.webp",
      "narrative": "draft",
      "content": "Risk of bias and applicability summary grid for 27 studies on pulmonary TB, from Bojang 2016 to Yadav 2021. Rows list the studies; columns show four risk-of-bias domains (patient selection, index test, reference standard, flow and timing) and three applicability domains (patient selection, index test, reference standard), each rated high (red), unclear (yellow) or low (green). Most cells are green; a few are yellow, and red (high) ratings appear for Getahun 2017 (patient selection bias) and for applicability of patient selection in Mitarai 2011, Promsena 2022 and Xin 2023.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-217",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 217",
      "pageStart": 217,
      "pageEnd": 217,
      "target": "../sections/page-217.md",
      "narrative": null,
      "content": "Figure 2. Risk of bias and applicability summary for each study: pulmonary TB",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A caption for the figure of risk of bias and applicability for each pulmonary TB study.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p218-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 218,
      "pageEnd": 218,
      "target": "../images/img-p218-1.webp",
      "narrative": "draft",
      "content": "Stacked horizontal bar charts showing the percentage of studies judged high (red), unclear (yellow) or low (green) for each domain, with study counts on the bars. Risk of bias: patient selection 1 high, 1 unclear, 25 low; index test 27 low; reference standard 6 unclear, 21 low; flow and timing 1 unclear, 26 low. Applicability concerns: patient selection 3 high, 2 unclear, 22 low; index test 3 unclear, 24 low; reference standard 1 unclear, 26 low.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-218",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 218",
      "pageStart": 218,
      "pageEnd": 218,
      "target": "../sections/page-218.md",
      "narrative": null,
      "content": "Figure 3. Risk of bias and applicability summary: pulmonary TB \nReview authors' judgements about each domain presented as percentages across included studies. The \nnumbers shown on each bar indicate the number of studies in each judgement category.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A caption for the summary figure of risk of bias and applicability across pulmonary TB studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-219",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 219",
      "pageStart": 219,
      "pageEnd": 219,
      "target": "../sections/page-219.md",
      "narrative": null,
      "content": "214\nFindings \nTable 2 summarises the results for LC-mNAATs for pulmonary TB. Results are presented below with additional forest plots provided in Appendix 2. \n \nTable 2. Diagnostic accuracy of LC-mNAATs for pulmonary TB in adults and adolescents \nAnalysis \nStudies Total (cases) \nSummary sensitivity (95% CI) Summary specificity (95% CI) PPV (95% CI)* \nNPV (95% CI)* \nDetection of pulmonary TB \nSputum \n26 18297 (4108) 84.1 (78.3 to 88.6) \n96.1 (94.2 to 97.4) \n70.6 (60.0 to 79.1) \n98.2 (97.5 to 98.7) \nBAL \n2 96 (12) \n− \n− \n− \n− \nGastric aspirate \n1 6 (2) \n− \n− \n− \n− \nGastric lavage \n1 24 (4) \n− \n−",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents Table 2 of LC-mNAAT diagnostic accuracy for pulmonary TB, giving pooled sensitivity and specificity for sputum and subgroups.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p220-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 220,
      "pageEnd": 220,
      "target": "../images/img-p220-1.webp",
      "narrative": "draft",
      "content": "Forest plot of sensitivity and specificity of LC-mNAATs for pulmonary TB in respiratory samples across 26 studies, from Nakiyingi 2018 to Yadav 2017. Each row gives TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as blue squares with confidence-interval lines on 0–1 scales. Sensitivity ranges from 0.55 to 1.00 and specificity from 0.70 to 1.00, with most specificities near 1.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-220",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 220",
      "pageStart": 220,
      "pageEnd": 220,
      "target": "../sections/page-220.md",
      "narrative": null,
      "content": "Should LC-mNAATs on respiratory samples be used to diagnose pulmonary TB in \nadults and adolescents with signs and symptoms of pulmonary TB, against a \nmicrobiological reference standard? \nTwenty-six (18297 participants) studies assessed diagnostic accuracy using sputum specimens. The \nsensitivities were between 55% and 100% and the specificities were between 70% and 100% (Figure \n4). The summary sensitivity (95% CI) was 84.1% (78.3 to 88.6) and the summary specificity was \n96.1% (94.2 to 97.4); high certainty (sensitivity and specificity) evidence.  \n \nFor a population of 1000 adults and adol",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that across 26 studies of sputum, TB-LAMP sensitivity ranged widely and gives the summary estimates, then introduces the analysis in people living with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p221-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 221,
      "pageEnd": 221,
      "target": "../images/img-p221-1.webp",
      "narrative": "draft",
      "content": "Risk of bias and applicability summary grid for 8 studies on pulmonary TB detection in people living with HIV. Rows list the studies; columns show four risk-of-bias domains (patient selection, index test, reference standard, flow and timing) and three applicability domains (patient selection, index test, reference standard), each rated high (red), unclear (yellow) or low (green). All ratings are low except unclear ratings for patient selection in N'Guessan 2016 and reference standard in Nakiyingi 2018.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p221-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 221,
      "pageEnd": 221,
      "target": "../images/img-p221-2.webp",
      "narrative": "draft",
      "content": "Forest plot of LC-mNAAT sensitivity and specificity for pulmonary TB in people living with HIV, 8 studies from Nakiyingi 2018 to Danfack 2018. Each row gives TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as blue squares with confidence-interval lines on 0–1 scales. Sensitivity ranges from 0.52 to 1.00; specificity is mostly 0.88–1.00, with one outlier of 0.27 (N'Guessan 2016).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-221",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 221",
      "pageStart": 221,
      "pageEnd": 221,
      "target": "../sections/page-221.md",
      "narrative": null,
      "content": "Figure 5. Risk of bias and applicability for each study: pulmonary TB detection in \nPLHIV \n \nThe sensitivities in the eight studies (2991 participants) included in the meta-analysis ranged \nbetween 52% and 100% and the specificities between 27% and 100% (Figure 6). The summary \nsensitivity (95% CI) was 77.1% (60.8 to 87.9) and the summary specificity was 95.9% (84.9 to 99.0); \nlow certainty (sensitivity) and moderate certainty (specificity) evidence. \n \nFor a population of 1000 adults and adolescents PLHIV where 100 have pulmonary TB on culture, 114 \nwould be LC-mNAAT-positive and 37 (32%) wou",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports a summary sensitivity of about 77% and specificity of about 96% for LC-mNAATs in people living with HIV, and lower sensitivity in those with CD4 counts below 200.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p222-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 222,
      "pageEnd": 222,
      "target": "../images/img-p222-1.webp",
      "narrative": "draft",
      "content": "Cropped forest plot fragment for people living with HIV with CD4 count <200: one study (Spooner 2022) with TP 14, FP 8, FN 9, TN 204, sensitivity 0.61 (0.39–0.80) and specificity 0.96 (0.93–0.98). The heading of the next subgroup (CD4 count ≥200) is cut off at the bottom.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p222-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 222,
      "pageEnd": 222,
      "target": "../images/img-p222-2.webp",
      "narrative": "draft",
      "content": "Risk of bias and applicability summary grid for 3 studies on pulmonary TB detection in HIV-negative adults and adolescents (Danfack 2023, Nliwasa 2016, Odeume 2021). Rows list the studies; columns show four risk-of-bias domains (patient selection, index test, reference standard, flow and timing) and three applicability domains (patient selection, index test, reference standard), each rated high (red), unclear (yellow) or low (green). All cells are rated low (green).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p222-3",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 222,
      "pageEnd": 222,
      "target": "../images/img-p222-3.webp",
      "narrative": "draft",
      "content": "Forest plot of LC-mNAAT sensitivity and specificity for pulmonary TB in HIV-negative people, 3 studies. Each row gives TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as blue squares with confidence-interval lines on 0–1 scales. Sensitivity is 0.53 (Nliwasa 2016), 0.76 (Odeume 2021) and 0.89 (Danfack 2023); specificity is 0.98–1.00 in all three.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-222",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 222",
      "pageStart": 222,
      "pageEnd": 222,
      "target": "../sections/page-222.md",
      "narrative": null,
      "content": "Figure 7. Forest plot of LC-mNAATs for detection of pulmonary TB: PLHIV by CD4 \ncount  \nFN: false negative; FP: false positive; TN: true negative; TP: true positive. \n \nShould LC-mNAATs on respiratory samples be used to diagnose pulmonary TB in \nHIV negative adults and adolescents with signs and symptoms of pulmonary TB, \nagainst a microbiological reference standard? \nFigure 8 summarises the results of the risk of bias and applicability assessment for each of the three \nincluded studies. The risk of bias and applicability concern was low across all the studies.  \nFigure 8. Risk of bias and app",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Part of a web annex systematic review on low-complexity manual nucleic acid amplification tests (LC-mNAATs, TB-LAMP) for TB, giving a forest plot for people with HIV by CD4 count and results for HIV-negative adults and adolescents. Three low-risk studies gave a summary sensitivity of about 77% and specificity of about 99% for pulmonary TB, followed by the start of additional analyses by smear status.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p223-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 223,
      "pageEnd": 223,
      "target": "../images/img-p223-1.webp",
      "narrative": "draft",
      "content": "Forest plot of LC-mNAAT sensitivity and specificity for pulmonary TB in smear-positive people, 22 studies from Danfack 2018 to Yadav 2021. Each row gives TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as blue squares with confidence-interval lines on 0–1 scales. Sensitivity is high (0.64–1.00), while specificity is widely scattered (0.00–1.00) with very wide intervals, and several studies are marked 'Not estimable'.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-223",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 223",
      "pageStart": 223,
      "pageEnd": 223,
      "target": "../sections/page-223.md",
      "narrative": null,
      "content": "against a microbiological reference standard? \nTwenty-one (1908 participants) studies evaluated sputum specimens, but specificity was not \nestimable for five studies. The sensitivities in the remaining 16 studies (1568 participants) included in \nthe meta-analysis were between 64% and 100% and the specificities were between 0% and 100% \n(Figure 10). The summary sensitivity (95% CI) was 95.9% (92.2 to 97.9) and the summary specificity \nwas 43.7% (26.8 to 62.2). \n \nFor a population of 1000 smear-positive adults and adolescents where 100 have pulmonary TB on \nculture, 603 would be LC-mNAAT-positiv",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports LC-mNAAT accuracy for pulmonary TB by sputum smear status: in smear-positive people sensitivity was about 96% but specificity only about 44%. For smear-negative people, results per 1000 tested show fewer false positives but more missed cases.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p224-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 224,
      "pageEnd": 224,
      "target": "../images/img-p224-1.webp",
      "narrative": "draft",
      "content": "Forest plot of LC-mNAAT sensitivity and specificity for pulmonary TB in smear-negative people, 25 studies from Nakiyingi 2018 to Yadav 2021. Each row gives TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as blue squares with confidence-interval lines on 0–1 scales. Sensitivity is lower and variable (0.24–1.00), while specificity is mostly 0.94–1.00, with lower values for Ou 2016 (0.87) and Mitarai 2011 (0.74).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p224-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 224,
      "pageEnd": 224,
      "target": "../images/img-p224-2.webp",
      "narrative": "draft",
      "content": "Risk of bias and applicability summary grid for 3 studies on extrapulmonary TB (Danfack 2024a, Promsena 2022, Singh 2021). Rows list the studies; columns show four risk-of-bias domains (patient selection, index test, reference standard, flow and timing) and three applicability domains (patient selection, index test, reference standard), each rated high (red), unclear (yellow) or low (green). All ratings are low except unclear reference-standard bias for Promsena 2022 and Singh 2021 and unclear reference-standard applicability for Singh 2021.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-224",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 224",
      "pageStart": 224,
      "pageEnd": 224,
      "target": "../sections/page-224.md",
      "narrative": null,
      "content": "Figure 11. Forest plot of LC-mNAATs for detection of pulmonary TB: smear negative, \nrespiratory samples and microbiological reference standard \nStudies are sorted on the plot by sensitivity, specificity and study. FN: false negative; FP: false positive; TN: true \nnegative; TP: true positive. \n \nDetection of extrapulmonary tuberculosis \nMethodological quality  \nOf the three studies included for extrapulmonary TB (Figure 12), we judged two studies to have an \nunclear risk of bias due to a lack of information about blinding to the index test results. One study \nwas judged to have an unclear appli",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives the forest plot for smear-negative samples and begins the results for extrapulmonary TB, noting unclear risk of bias in two of three studies and introducing the question on lymph node TB.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p225-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 225,
      "pageEnd": 225,
      "target": "../images/img-p225-1.webp",
      "narrative": "draft",
      "content": "Cropped forest plot of TB-LAMP for lymph node TB with partly cut-off headings. In lymph node biopsy (3 studies), sensitivity is 0.93 (Danfack 2024a), 1.00 (Promsena 2022) and 1.00 (Singh 2021), with specificity 0.88–1.00; in lymph node aspirate (Danfack 2024a), sensitivity is 0.67 and specificity 0.75. A further heading for LN pus is cut off.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p225-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 225,
      "pageEnd": 225,
      "target": "../images/img-p225-2.webp",
      "narrative": "draft",
      "content": "Forest plot of LC-mNAAT sensitivity and specificity for TB meningitis in cerebrospinal fluid, 2 studies. Each row gives TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as blue squares with confidence-interval lines on 0–1 scales. Danfack 2024a: sensitivity 0.00 (0.00–0.84), specificity 0.97; Singh 2021: sensitivity 1.00 (0.03–1.00), specificity 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-225",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 225",
      "pageStart": 225,
      "pageEnd": 225,
      "target": "../sections/page-225.md",
      "narrative": null,
      "content": "node TB using lymph node tissue from biopsy and a microbiological reference standard (Figure 13). \nThe estimated sensitivities were between 93% and 100% and specificities were between 88% and \n100%. The summary sensitivity (95% CI) was 94.3% (79.8 to 98.6) and the summary specificity was \n90.0% (79.5 to 95.4). For illustration at a 10% prevalence, the estimated positive and negative \npredictive values are 51.2% (30.2 to 70.4) and 99.3% (97.3 to 99.8); low certainty (sensitivity) and \nlow certainty (specificity) evidence. \n \nFor a population of 1000 adults and adolescents where 100 have lymph n",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports LC-mNAAT accuracy for lymph node TB on biopsy tissue, with summary sensitivity about 94% and specificity 90% from low-certainty evidence, and illustrates outcomes per 1000 people tested. It also mentions data from aspirate and pus and introduces two small studies on abdominal TB in ascitic fluid.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p226-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 226,
      "pageEnd": 226,
      "target": "../images/img-p226-1.webp",
      "narrative": "draft",
      "content": "Forest plot of LC-mNAAT sensitivity and specificity for abdominal TB in ascitic fluid, 2 studies. Each row gives TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as blue squares with confidence-interval lines on 0–1 scales. Danfack 2024a: sensitivity 0.17, specificity 0.95; Singh 2021: sensitivity 1.00, specificity 0.94.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p226-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 226,
      "pageEnd": 226,
      "target": "../images/img-p226-2.webp",
      "narrative": "draft",
      "content": "Cropped forest plot of TB-LAMP for pleural TB with partly cut-off labels. In pleural fluid, Danfack 2024a has sensitivity 0.48 and specificity 0.96, and Singh 2021 sensitivity 0.75 and specificity 0.89; in pleural biopsy (Danfack 2024a), sensitivity is 0.00 (0.00–0.10) and specificity 0.93. A heading for pleural pus is cut off at the bottom.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p226-3",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 226,
      "pageEnd": 226,
      "target": "../images/img-p226-3.webp",
      "narrative": "draft",
      "content": "Cropped single-row forest plot with a cut-off header: Danfack 2024a with TP 1, FP 0, FN 0, TN 4, sensitivity 1.00 (0.03–1.00) and specificity 1.00 (0.40–1.00), plotted on 0–1 scales.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-226",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 226",
      "pageStart": 226,
      "pageEnd": 226,
      "target": "../sections/page-226.md",
      "narrative": null,
      "content": "sensitivities were 17% and 100% and estimated specificities were 94% and 95%. \n \n \nFigure 15. Forest plot of LC-mNAATs for detection of abdominal TB in ascitic fluid \nFN = false negative; FP = false positive; TN = true negative; TP = true positive.  \n \nShould LC-mNAATs on pleural fluid be used to diagnose pleural TB in adults and \nadolescents with signs and symptoms of pleural TB, against a microbiological reference \nstandard? \nTwo studies (292 participants; 37 cases) assessed the accuracy of LC-mNAATs for detecting pleural \nTB using pleural fluid and a microbiological reference standard (Figu",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports small-study results for LC-mNAATs in extrapulmonary TB, including abdominal TB in ascitic fluid, pleural TB in pleural fluid and bone or joint TB. It ends introducing a single study on genitourinary TB in urine.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p227-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 227,
      "pageEnd": 227,
      "target": "../images/img-p227-1.webp",
      "narrative": "draft",
      "content": "Forest plot of LC-mNAAT sensitivity and specificity for genitourinary TB in urine, 1 study. Each row gives TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as blue squares with confidence-interval lines on 0–1 scales. Danfack 2024a: TP 1, FP 0, FN 1, TN 30; sensitivity 0.50 (0.01–0.99), specificity 1.00 (0.88–1.00).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-227",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 227",
      "pageStart": 227,
      "pageEnd": 227,
      "target": "../sections/page-227.md",
      "narrative": null,
      "content": "specificity were 50% and 100%.  \nFigure 18. Forest plot of LC-mNAATs for detection of genitourinary TB in urine \nFN = false negative; FP = false positive; TN = true negative; TP = true positive.  \n \nDiscussion \nOur review included 27 studies that assessed pulmonary TB and three studies assessed \nextrapulmonary TB. The estimated summary sensitivity (95% CI) was 84.1% (78.3 to 88.6) and the \nsummary specificity was 96.1% (94.2 to 97.4) for the detection of pulmonary TB (18297 participants; \n26 studies). The summary sensitivity of TB-LAMP for detection of pulmonary TB was lower [84.1% \n(78.3 to 8",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the genitourinary TB result and discusses the review's findings, noting a pooled sensitivity of about 84% and specificity of about 96% for pulmonary TB, lower sensitivity than Xpert MTB/RIF Ultra but similar specificity. It ends with acknowledgements to the study authors who supplied data.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-228",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 228",
      "pageStart": 228,
      "pageEnd": 228,
      "target": "../sections/page-228.md",
      "narrative": null,
      "content": "References \nIncluded studies \nBojang AL, Mendy FS, Tientcheu LD, Otu J, Antonio M, Kampmann B, et al. Comparison of TB-LAMP, \nGeneXpert MTB/RIF and culture for diagnosis of pulmonary tuberculosis in The Gambia. Journal of \nInfection [Internet]. 2016 Mar 1 [cited 2020 Sep 30];72(3):332–7.  \nCheng S, Pheng SH, Heng S, Marks GB, Bañuls AL, Mao TE, et al. Evaluation of Loopamp Assay for the \nDiagnosis of Pulmonary Tuberculosis in Cambodia. BioMed Research International [Internet]. 2020 \nJun 1 [cited 2022 Dec 8];2020:6828043.  \nDonfack VD, Ngando L, Pefura EY, Che D, Ateba G, Bigna JR, et al. Compa",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The reference list of included studies for the adult TB-LAMP review, citing diagnostic accuracy studies from The Gambia, Cambodia, Cameroon, Nigeria and China.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-229",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 229",
      "pageStart": 229,
      "pageEnd": 229,
      "target": "../sections/page-229.md",
      "narrative": null,
      "content": "Pulmonary Tuberculosis at the County-Level Laboratory in China. Cattamanchi A, editor. PLoS ONE. \n2014 May 1;9(5):e94544. \nOu X, Wang S, Dong H, Pang Y, Li Q, Xia H, et al. Multicenter evaluation of a real-time loop-mediated \nisothermal amplification (RealAmp) test for rapid diagnosis of Mycobacterium tuberculosis. Journal \nof Microbiological Methods. 2016 Oct 1;129:39–43. \nOu XC, Li H, Liu DX, Xia H, Ma XG, Wang SH, et al. Comparison of Xpert MTB/RIF, RealAmp, and CPA \nTests in Detecting Mycobacterium tuberculosis. PubMed. 2019 Mar 1;32(3):215–9. \nPham TH, Peter J, Mello FCQ, Parraga T, Lan N",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list for the adult TB-LAMP review, citing studies from China, multicentre and paediatric evaluations, and GRADE guidance on rating test accuracy evidence.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-230",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 230",
      "pageStart": 230,
      "pageEnd": 230,
      "target": "../sections/page-230.md",
      "narrative": null,
      "content": "Journal of Clinical Epidemiology. 2020 Jun;122:142–52. \nWorld Health Organization. The use of loop-mediated isothermal amplification (TB-LAMP) for the \ndiagnosis of pulmonary tuberculosis: policy guidance [Internet]. iris.who.int. Geneva: World Health \nOrganization; 2016 [cited 2024 Mar 13]. Available from: https://iris.who.int/handle/10665/249154 \nWorld Health Organization. Global tuberculosis report 2023 [Internet]. Geneva: World Health \nOrganization; 2023 [cited 2024 Mar 13]. Available from: https://www.who.int/teams/global-\ntuberculosis-programme/tb-reports \nZifodya JS, Kreniske JS, Schill",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The end of the reference list, citing WHO TB-LAMP policy guidance, the Global tuberculosis report 2023 and a Cochrane review of Xpert Ultra.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-231",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 231",
      "pageStart": 231,
      "pageEnd": 231,
      "target": "../sections/page-231.md",
      "narrative": null,
      "content": "Appendix \nAppendix 1. Detailed search strategies \n \nOvid MEDLINE(R) ALL <1946 to September 29, 2023> \n1 \nMycobacterium tuberculosis/ \n \n2 \nTuberculosis, Pulmonary/ \n \n3 \nExtensively Drug-Resistant Tuberculosis/ \n4 \nTuberculosis, Multidrug-Resistant/ \n \n5 \n(tuberculosis or TB).tw.  \n6 \n((extrapulmonary or lymph node* or mening* or pulmonary) and TB).ti. or ((extrapulmonary \nor lymph node* or mening* or pulmonary) and TB).ab.  \n7 \nexp Tuberculosis, Extrapulmonary/ \n \n8 \n5 and 6  \n9 \n7 or 8  \n10 \n1 or 2 or 3 or 4 or 9 \n \n11 \nTB-LAMP*.mp.  \n12 \nLoop-Mediated Isothermal Amplification.mp. \n \n13 \nLAM",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Appendix 1 giving the detailed literature search strategies for the TB-LAMP review in databases such as MEDLINE and Embase.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-232",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 232",
      "pageStart": 232,
      "pageEnd": 232,
      "target": "../sections/page-232.md",
      "narrative": null,
      "content": "#10 Search: #6 AND #9  \nScopus Elsevier \n( ( TITLE-ABS-KEY ( loop-mediated AND isothermal AND amplification ) ) OR ( TITLE-ABS-KEY ( loop \nAND mediated AND isothermal AND amplification ) ) OR ( TITLE-ABS-KEY ( tb-lamp* OR loopamp* ) ) \n) AND ( ( TITLE-ABS-KEY ( ( tuberculosis OR tb ) AND ( extrapulmonary OR lymph AND node* OR \nmening* OR pulmonary ) ) ) OR ( TITLE-ABS-KEY ( ( multidrug AND resistant AND tuberculosis ) OR ( \nextensively AND drug AND resistant AND tuberculosis ) ) ) OR ( TITLE-ABS-KEY ( mdr-tb OR xdr-tb ) ) ) \nProQuest Dissertations & Theses A&I \nTuberculosis and Loop-Mediated I",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the search strategies, listing queries used in Scopus, ProQuest, the Cochrane Central Register, trial registries and Global Index Medicus.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p233-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 233,
      "pageEnd": 233,
      "target": "../images/img-p233-1.webp",
      "narrative": "draft",
      "content": "Cropped forest plot of TB-LAMP for pulmonary TB in other specimen types, with partly cut-off labels. The first panel (3 studies) shows sensitivity 0.64 for Danfack 2024b, 'Not estimable' for Promsena 2022 and 1.00 for Yadav 2021, with specificity 0.94–1.00; the panel for gastric aspirate specimens shows sensitivity 0.50 (Promsena 2022) and 'Not estimable' (Yadav 2021), specificity 1.00. A heading for gastric lavage specimens is cut off.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p233-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 233,
      "pageEnd": 233,
      "target": "../images/img-p233-2.webp",
      "narrative": "draft",
      "content": "Two-panel forest plot of TB-LAMP for pulmonary TB (PTB) by laboratory setting: 'peripheral lab' (7 studies, Nliwasa 2016 to Wang 2019) and 'central lab' (19 studies, Spooner 2022 to Yadav 2017). Each row gives TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as blue squares with confidence-interval lines on 0–1 scales. Peripheral-lab sensitivity ranges 0.65–0.89 with specificity 0.96–1.00; central-lab sensitivity ranges 0.57–1.00 with specificity mostly 0.86–1.00 and a low of 0.70 (Wahid 2020).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-233",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 233",
      "pageStart": 233,
      "pageEnd": 233,
      "target": "../sections/page-233.md",
      "narrative": null,
      "content": "Appendix 2. Additional forest plots  \n \nAppendix Figures 1 to 3 summarise currently available data for other sample types and subgroups.  \n \nAppendix Figure 1. Forest plot of LC-mNAATs for pulmonary TB: BAL, gastric \naspirate and gastric lavage specimens \nStudies are sorted on the plot by study. BAL= bronchoalveolar lavage; FN = false negative; FP = false positive; \nPTB = pulmonary TB; TN = true negative; TP = true positive. \n \n \nAppendix Figure 2. Forest plot of LC-aNAATs for pulmonary TB: lab setting  \nStudies are sorted on the plot by sensitivity, specificity and study. FN = false negative;",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Appendix 2 introducing additional forest plots for other specimen types, such as bronchoalveolar lavage and gastric samples, and for laboratory settings.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p234-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 234,
      "pageEnd": 234,
      "target": "../images/img-p234-1.webp",
      "narrative": "draft",
      "content": "Forest plot of TB-LAMP for pulmonary TB in high TB burden settings, 17 studies from Nakiyingi 2018 to Yadav 2017. Each row gives TP, FP, FN and TN counts with sensitivity and specificity (95% CI), plotted as blue squares with confidence-interval lines on 0–1 scales. Sensitivity ranges from 0.55 to 1.00 and specificity mostly from 0.86 to 1.00, with a low of 0.70 (Wahid 2020).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-234",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 234",
      "pageStart": 234,
      "pageEnd": 234,
      "target": "../sections/page-234.md",
      "narrative": null,
      "content": "Appendix Figure 3. Forest plot of LC-aNAATs for pulmonary TB: high burden setting  \nStudies are sorted on the plot by sensitivity, specificity and study. FN = false negative; FP = false positive; PTB = \npulmonary TB; TN = true negative; TP = true positive.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Caption of an appendix forest plot of TB test accuracy for pulmonary TB in high-burden settings.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-235",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 235",
      "pageStart": 235,
      "pageEnd": 235,
      "target": "../sections/page-235.md",
      "narrative": null,
      "content": "Page 230 of 1168 \n \nAppendix 3. Summary of findings tables  \n \n1. Should LC-mNAAT on respiratory samples be used to diagnose PTB in \nadults and adolescents with signs and symptoms of PTB, MRS? \n \nPatient or population: adults and adolescents with signs and symptoms of PTB, MRS \nPooled sensitivity: 0.84 (95% CI: 0.78 to 0.89) Pooled specificity: 0.96 (95% CI: 0.94 to 0.97) \nTest result \nNumber of results per 1,000 patients tested (95% \nCI) \nNumber of \nparticipants  \n(studies) \nCertainty of \nthe Evidence \n(GRADE) \nPrevalence \n2.5% \nPrevalence 10% \n \nPrevalence 30% \n \nTrue positives \n21 (20 to 22",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summary of findings tables for LC-mNAATs in adults, giving numbers of true and false results per 1000 tested at several prevalence levels with GRADE certainty. Pulmonary TB on respiratory samples is rated high certainty, with the table for people with HIV and explanatory notes on downgrading following.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-236",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 236",
      "pageStart": 236,
      "pageEnd": 236,
      "target": "../sections/page-236.md",
      "narrative": null,
      "content": "Page 231 of 1168 \n \nc. The point estimates of individual studies ranged from 27% to 100% and 95% Cis did not overlap for a few \nstudies. We downgraded one level for inconsistency. \nd. The 95% CI was wide (84.9 to 99.0) for the specificity. The number of participants contributing to the \nestimation of specificity was large (2531 participants). This was a judgement, and we did not downgrade for \nimprecision. \n \n \n \n3. Should LC-mNAAT on respiratory samples be used to diagnose PTB in \nPLHIV with CD4 <200 with signs and symptoms of PTB, MRS? \n \nPatient or population: PLHIV with CD4 <200 with signs",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the summary of findings tables, with notes on inconsistency and imprecision and a table for people with HIV with CD4 below 200, rated very low certainty. It begins the table for lymph node TB.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-237",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 237",
      "pageStart": 237,
      "pageEnd": 237,
      "target": "../sections/page-237.md",
      "narrative": null,
      "content": "Page 232 of 1168 \n \nTest result \nNumber of results per 1,000 patients tested (95% \nCI) \nNumber of \nparticipants  \n(studies) \nCertainty of \nthe Evidence \n(GRADE) \nPrevalence \n2.5% \nPrevalence 10% \n \nPrevalence 20% \n \nTrue negatives \n878 (780 to \n926) \n810 (720 to \n855) \n720 (640 to \n760) \n60 \n(3) \n⨁⨁◯◯ \nLowb \nFalse positives \n97 (49 to 195) \n90 (45 to 180) \n80 (40 to 160) \nCI: confidence interval \nExplanations \na. There were very few cases (35) with lymph node TB contributing to sensitivity estimation. For a prevalence \nvalue of 10%, we thought the 95% CI around true positives and false negativ",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the summary of findings table for lymph node TB, explaining that the evidence was downgraded two levels for very serious imprecision due to few participants.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-238",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 238",
      "pageStart": 238,
      "pageEnd": 238,
      "target": "../sections/page-238.md",
      "narrative": null,
      "content": "Page 233 of 1168 \n \n5. Should LC-mNAAT on CSF be used to diagnose TB meningitis in adults \nand adolescents with signs and symptoms of TB meningitis, MRS? \n \nPatient or population: adults and adolescents with signs and symptoms of TB meningitis, MRS \nRange of sensitivities:0.01 to 0.9 Range of specificities:0.97 to 0.99 \nTest result \nNumber of results per 1,000 patients tested (95% \nCI) \nNumber of \nparticipants  \n(studies) \nCertainty of \nthe Evidence \n(GRADE) \nPrevalence2.5% Prevalence10% Prevalence20% \nTrue positives \n0 to 25 \n1 to 99 \n2 to 198 \n3 \n(2) \n⨁◯◯◯ \nVery lowa,b \nFalse negatives \n0 to",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summary of findings table for LC-mNAATs on cerebrospinal fluid for TB meningitis, rated very low to low certainty because of inconsistent sensitivity and very few cases.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-239",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 239",
      "pageStart": 239,
      "pageEnd": 239,
      "target": "../sections/page-239.md",
      "narrative": null,
      "content": "Page 234 of 1168 \n \nWeb Annex B.6. LC-mNAATs for detection of pulmonary and \nextrapulmonary tuberculosis in children: a systematic review of \ndiagnostic accuracy \n \nLeeberk Raja Inbaraj1*, Mukesh Kumar Sathyanarayanan1, Jefferson Daniel2, Vignes Anand Srinivasalu1 \nAdhin Bhaskar1, Katie Scandrett3, Yemisi Takwoingi3  \n \n1. ICMR- National Institute for Research in Tuberculosis, Chennai, India \n2. Christian Medical College, Vellore, India \n3. Institute of Applied Health Research, University of Birmingham, United Kingdom  \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n*Corresponding author: \nLeeberk Raja Inb",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Title page of Web Annex B.6, a systematic review of LC-mNAATs for detecting pulmonary and extrapulmonary TB in children, listing authors from ICMR, Christian Medical College and the University of Birmingham.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-240",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 240",
      "pageStart": 240,
      "pageEnd": 240,
      "target": "../sections/page-240.md",
      "narrative": null,
      "content": "Page 235 of 1168 \n \nBackground \nThe World Health Organization (WHO) estimated 1.3 million children (aged 0–14 years) were affected \nby tuberculosis (TB) in 2022. Globally, paediatric TB contributed to 16% of deaths among HIV-negative \nchildren in the same year. Multiple factors contribute to diagnostic delays and the consequent rise in \nincidence and mortality among children. The unique pathophysiology and clinical manifestations of \nTB in children make diagnosis particularly difficult. While there are practical difficulties in obtaining \nspecimens for diagnosis, about 15% are positive on acid",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives the background to the paediatric TB-LAMP review, citing WHO estimates of 1.3 million children with TB in 2022 and the difficulty of diagnosis in children. It describes the TB-LAMP test, which takes about an hour in three steps and gives no information on drug resistance.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-241",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 241",
      "pageStart": 241,
      "pageEnd": 241,
      "target": "../sections/page-241.md",
      "narrative": null,
      "content": "Page 236 of 1168 \n \ndrugs. \nObjectives  \n1. To assess the diagnostic accuracy of LC-mNAATs for detection of pulmonary and \nextrapulmonary TB by the site of disease in children with presumptive TB.  \n2. To investigate the potential sources of heterogeneity. For detection of pulmonary TB, we \naimed to investigate the effect of smear microscopy status (smear positive/smear negative), \nHIV status (positive/negative), and the history of TB on test accuracy. \nMethods \nMethods are described in detail in the generic protocol for six diagnostic accuracy reviews that will \nprovide evidence to inform the",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Sets out the review's objectives, to assess TB-LAMP accuracy by disease site in children and explore heterogeneity, and refers to a generic protocol informing the 2024 WHO guideline update. It describes the search methods and eligibility criteria, including children under 10 and respiratory specimens.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-242",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 242",
      "pageStart": 242,
      "pageEnd": 242,
      "target": "../sections/page-242.md",
      "narrative": null,
      "content": "Page 237 of 1168 \n \naspirate, nasopharyngeal aspirate). For extrapulmonary TB, we included non-respiratory specimens \nsuch as cerebrospinal fluid (CSF), pleural fluid, lymph node aspirate or tissue. We considered blood for \ndisseminated TB. We only included studies that provided at least five specimens for a given form of \nextrapulmonary TB. We excluded studies designed to find people with active TB in community \nsettings. If a study included children and adolescents or adults and if disaggregated data were not \navailable in the published paper, we contacted study authors for the data. If the",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the methods, covering specimen types for extrapulmonary TB, exclusion criteria and handling of mixed-age data. It defines the index test, the culture and composite reference standards and the use of GRADE to assess certainty.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-243",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 243",
      "pageStart": 243,
      "pageEnd": 243,
      "target": "../sections/page-243.md",
      "narrative": null,
      "content": "Page 238 of 1168 \n \nResults of the search \nWe identified 2806 research articles for screening from searches of electronic sources and screened \nthe titles and abstracts of 1138 unique articles. Of the 1138 articles, we excluded 988. We performed \nfull-text screening of 151 articles. We included four studies for pulmonary TB and one of them for both \npulmonary and extrapulmonary TB. Figure 1 shows the flow of studies through the screening and \nselection process.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports the search results: 1138 unique records screened, 151 full texts reviewed and four studies included for pulmonary TB, one also covering extrapulmonary TB.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-244",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 244",
      "pageStart": 244,
      "pageEnd": 244,
      "target": "../sections/page-244.md",
      "narrative": null,
      "content": "Page 239 of 1168 \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \nFigure.1 Study flow diagram  \n*One study contributed to both pulmonary TB (PTB) and extrapulmonary TB (EPTB). DTA= diagnostic test accuracy. \nRecords identified from: \nDatabases (n =2806) \nRegisters (n = 0) \nRecords removed before \nscreening: \nDuplicate records removed  \n(n = 1668 ) \nRecords marked as ineligible \nby automation tools (n = 0) \nRecords removed for other \nreasons (n = 0) \nRecords screened (n = 1138) \nRecords excluded (n = 988) \nReports sought for retrieval \n(n = 150) \nReports not retrieved (n =",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A study flow diagram for the paediatric TB-LAMP review, showing records identified, screened and excluded with reasons such as case-control design or in-house kits.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-245",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 245",
      "pageStart": 245,
      "pageEnd": 245,
      "target": "../sections/page-245.md",
      "narrative": null,
      "content": "Page 240 of 1168 \n \nCharacteristics of included studies \nKey characteristics of the four included studies for detection of pulmonary TB are described in Table \n1. One of the studies also assessed extrapulmonary TB (Promsena 2022). All studies were cross-\nsectional in design and assessed TB-LAMP using a microbiological reference standard in central \nlaboratories. No study used a composite reference standard. None of the studies reported non-\ndeterminate test results. \n \nTable 1. Key characteristics of included studies \nStudy \nReference \nstandard \nCountry  \nHIV \nstatus  \nClinical \nsetting \nHigh",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the four included cross-sectional studies from The Gambia, Thailand, India and Cameroon, with a table of their reference standards, settings and specimen types. It notes risk of bias concerns and high applicability concern for one inpatient study.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-vfig-p246",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 246,
      "pageEnd": 246,
      "target": "../figures/vfig-p246.png",
      "narrative": "draft",
      "content": "A legend key for risk-of-bias judgements: a red circle with a minus sign for High, a yellow circle with a question mark for Unclear and a green circle with a plus sign for Low.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-246",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 246",
      "pageStart": 246,
      "pageEnd": 246,
      "target": "../sections/page-246.md",
      "narrative": null,
      "content": "Page 241 of 1168 \n \n \nFigure 2. Risk of bias and applicability summary for each study \n \nFindings \nMain results are presented below with additional forest plots provided in Appendix 2 and summary \nof findings tables in Appendix 3. Figure 3 summarises the sensitivity and specificity of the studies \nincluded across and for each specimen type. Four studies (314 participants) assessed diagnostic \naccuracy using different specimen types. The sensitivities were between 50% and 100% and the \nspecificities were between 67% and 100%. Given the mix of specimen types and paucity of data, we \ndid not perf",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces findings for the paediatric review, with sensitivities from 50% to 100% and specificities from 67% to 100% across four studies; no meta-analysis was done because of mixed specimens and sparse data. A risk of bias figure is partly extracted as garbled text.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-247",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 247",
      "pageStart": 247,
      "pageEnd": 247,
      "target": "../sections/page-247.md",
      "narrative": null,
      "content": "Page 242 of 1168 \n \n \nFigure 3. Forest plot of LC-mNAATs for detection of pulmonary TB by specimen type \nBAL= bronchoalveolar lavage; FN: false negative; FP: false positive; NPA = nasopharyngeal aspirate; PTB = \npulmonary TB; TN: true negative; TP: true positive. \n \nShould LC-mNAATs on respiratory samples be used to diagnose pulmonary TB \nin children with signs and symptoms of pulmonary TB, against a \nmicrobiological reference standard? \nThree studies assessed the accuracy of LC-mNAATs for detecting pulmonary TB using respiratory \nsamples (sputum, BAL and tracheal aspirate) and a microbiologic",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives the forest plot by specimen type and reports that three studies on respiratory samples showed sensitivities of 60–100% and specificities of 95–100%, with very low to low certainty. Results for stool and other specimens follow. Much of the forest plot data was extracted as garbled text with missing letters.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-248",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 248",
      "pageStart": 248,
      "pageEnd": 248,
      "target": "../sections/page-248.md",
      "narrative": null,
      "content": "Page 243 of 1168 \n \nchildren with signs and symptoms of pulmonary TB, against a microbiological \nreference standard? \nThree studies (176 participants including 14 with pulmonary TB) assessed the accuracy of LC-\nmNAATs for detecting pulmonary TB using gastric aspirate and a microbiological reference standard \n(Figure 3). Sensitivity was not estimable for two studies and was 64% in the third study. The \nspecificities were between 93% and 100%. \n \nShould LC-mNAATs on gastric lavage be used to diagnose pulmonary TB \nusing in children with signs and symptoms of pulmonary TB, against a \nmicrobiologi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports paediatric TB-LAMP results for gastric aspirate, tracheal aspirate and nasopharyngeal aspirate, mostly from single small studies. It then introduces extrapulmonary TB results and their quality assessment.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p249-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 249,
      "pageEnd": 249,
      "target": "../images/img-p249-1.webp",
      "narrative": "draft",
      "content": "Risk of bias and applicability summary grid for 1 study on extrapulmonary TB (Promsena 2022). Rows list the studies; columns show four risk-of-bias domains (patient selection, index test, reference standard, flow and timing) and three applicability domains (patient selection, index test, reference standard), each rated high (red), unclear (yellow) or low (green). All domains are rated low except a high (red) applicability concern for patient selection.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-249",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 249",
      "pageStart": 249,
      "pageEnd": 249,
      "target": "../sections/page-249.md",
      "narrative": null,
      "content": "Page 244 of 1168 \n \n \nFigure 5. Risk of bias and applicability summary for detection of extrapulmonary TB \nReview authors' judgements about each domain presented per study.  \n \nFindings  \nShould LC-mNAATs be used to diagnose lymph node TB in lymph node tissue \nin children with signs and symptoms of lymph node TB, against a \nmicrobiological reference standard? \nOne study (nine participants) assessed the accuracy of LC-mNAATs for detecting lymph node TB \nusing lymph node tissue and a microbiological reference standard (Figure 6). Estimated sensitivity \nand specificity were 100%. \n \n \nFigure 6. F",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports a single small study of TB-LAMP on lymph node tissue in children and discusses that the evidence is limited by few, small studies, calling for adequately powered research. Acknowledgements and a partly garbled forest plot follow.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-250",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 250",
      "pageStart": 250,
      "pageEnd": 250,
      "target": "../sections/page-250.md",
      "narrative": null,
      "content": "Page 245 of 1168 \n \nReferences \nIncluded studies  \nBojang AL, Mendy FS, Tientcheu LD, Otu J, Antonio M, Kampmann B, et al. Comparison of TB-LAMP, \nGeneXpert MTB/RIF and culture for diagnosis of pulmonary tuberculosis in The Gambia. Journal of \nInfection. 2016 Mar;72(3):332-7. doi: 10.1016/j.jinf.2015.11.011. Epub 2015 Dec 24. PMID: \n26724771 \nDonfack VD et al. Performance Evaluation of TB-LAMP in children. 2024 (Unpublished) \nPromsena P, Jantarabenjakul W, Suntarattiwong P, Sudjaritruk T, Anugulruengkitt S, Rotcheewaphan \nS, et al. Diagnostic accuracy of loop-mediated isothermal amplification",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The reference list for the paediatric TB-LAMP review, citing included studies from The Gambia, Cameroon, Thailand and India and background literature on childhood TB diagnosis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-251",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 251",
      "pageStart": 251,
      "pageEnd": 251,
      "target": "../sections/page-251.md",
      "narrative": null,
      "content": "Page 246 of 1168 \n \nAppendix \n \nAppendix 1. Detailed search strategies \nDetailed search strategies \nOvid MEDLINE(R) ALL <1946 to September 29, 2023> \n1 \nMycobacterium tuberculosis/ \n \n2 \nTuberculosis, Pulmonary/ \n \n3 \nExtensively Drug-Resistant Tuberculosis/ \n4 \nTuberculosis, Multidrug-Resistant/ \n \n5 \n(tuberculosis or TB).tw.  \n6 \n((extrapulmonary or lymph node* or mening* or pulmonary) and TB).ti. or ((extrapulmonary \nor lymph node* or mening* or pulmonary) and TB).ab.  \n7 \nexp Tuberculosis, Extrapulmonary/ \n \n8 \n5 and 6  \n9 \n7 or 8  \n10 \n1 or 2 or 3 or 4 or 9 \n \n11 \nTB-LAMP*.mp.  \n12 \nLoop-",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Appendix 1 giving detailed search strategies for the paediatric TB-LAMP review in MEDLINE, Embase and Web of Science.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-252",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 252",
      "pageStart": 252,
      "pageEnd": 252,
      "target": "../sections/page-252.md",
      "narrative": null,
      "content": "Page 247 of 1168 \n \n#9 Search: #7 OR #8 \n \n#10 Search: #6 AND #9  \nScopus Elseviers \n( ( TITLE-ABS-KEY ( loop-mediated AND isothermal AND amplification ) ) OR ( TITLE-ABS-KEY ( loop \nAND mediated AND isothermal AND amplification ) ) OR ( TITLE-ABS-KEY ( tb-lamp* OR loopamp* ) ) \n) AND ( ( TITLE-ABS-KEY ( ( tuberculosis OR tb ) AND ( extrapulmonary OR lymph AND node* OR \nmening* OR pulmonary ) ) ) OR ( TITLE-ABS-KEY ( ( multidrug AND resistant AND tuberculosis ) OR ( \nextensively AND drug AND resistant AND tuberculosis ) ) ) OR ( TITLE-ABS-KEY ( mdr-tb OR xdr-tb ) ) ) \nProQuest Dissertations &",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the search strategies, listing queries in Scopus, ProQuest, the Cochrane Central Register, trial registries and Global Index Medicus.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-253",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 253",
      "pageStart": 253,
      "pageEnd": 253,
      "target": "../sections/page-253.md",
      "narrative": null,
      "content": "Page 248 of 1168 \n \nAppendix 2. Additional forest plots  \nAppendix Figures 1 to 3 summarise available data for different subgroups.  \n \nAppendix Figure 1. Forest plot of LC-mNAATs for pulmonary TB by age group for \nrespiratory specimen types \nT   A P for PT ,   1 year, sputum\n tudy\nTP\nFP\nFN\nTN\n ensitivity (95% CI)\n pecificity (95% CI)\n ensitivity (95% CI)\n pecificity (95% CI)\n a a  2021\n0\n0\n0\n2\n ot esti a le\n1 00  0 16  1 00 \n0\n0 2\n0 4\n0 6\n0 8\n1\n0\n0 2\n0 4\n0 6\n0 8\n1\nT   A P for PT , 1 to 4 years, sputum\n tudy\nTP\nFP\nFN\nTN\n ensitivity (95% CI)\n pecificity (95% CI)\n ensitivity (95% CI)\n pecificity",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Appendix forest plots of paediatric TB-LAMP accuracy by age group and respiratory specimen type. The plot data was extracted as garbled text with missing letters and is largely unreadable.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p254-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 254,
      "pageEnd": 254,
      "target": "../images/img-p254-1.webp",
      "narrative": "draft",
      "content": "Six single-study forest plots (Danfack 2024) of TB-LAMP for pulmonary TB in children by HIV status and specimen. For children living with HIV (CLHIV), gastric aspirate, NPA and stool all show sensitivity 'Not estimable' and specificity 0.93–1.00. For HIV-negative children, sensitivity is 0.64 (gastric aspirate), 0.58 (NPA) and 1.00 (stool), with specificity 0.91–0.93.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-254",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 254",
      "pageStart": 254,
      "pageEnd": 254,
      "target": "../sections/page-254.md",
      "narrative": null,
      "content": "Page 249 of 1168 \n \nBAL= bronchoalveolar lavage; FN: false negative; FP: false positive; NPA = nasopharyngeal aspirate; PTB = \npulmonary TB; TN: true negative; TP: true positive. \n \n \n \nAppendix Figure 2. Forest plot of LC-mNAATs for pulmonary TB by age group for \nstool specimen type \nFN: false negative; FP: false positive; PTB = pulmonary TB; TN: true negative; TP: true positive. \n \n \n \nAppendix Figure 3. Forest plot of LC-mNAATs for pulmonary TB by HIV status  \nCLHIV= children living with HIV; FN: false negative; FP: false positive; NPA = nasopharyngeal aspirate; PTB = \npulmonary TB; TN: tru",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Appendix forest plots of paediatric TB-LAMP accuracy by age group for stool and by HIV status. Much of the plot data was extracted as garbled text.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-255",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 255",
      "pageStart": 255,
      "pageEnd": 255,
      "target": "../sections/page-255.md",
      "narrative": null,
      "content": "250 \n \nAppendix 3. Summary of findings tables  \n \n1. Should LC-mNAAT on respiratory samples be used to diagnose PTB in \nchildren with signs and symptoms of PTB, MRS? \nPatient or population: Children with signs and symptoms of PTB, MRS \nRange of sensitivities: 0.60 to 0.99  Range of specificities: 0.95 to 0.99 \nTest result \nNumber of results per 1,000 patients tested (95% CI) \nNumber of \nparticipants  \n(studies) \nCertainty of \nthe Evidence \n(GRADE) \nPrevalence 1% \nPrevalence 5% \nPrevalence 10% \nTrue positives \n6 to 10 \n30 to 50 \n60 to 99 \n8 \n(4) \n⨁◯◯◯ \nVery lowa, b,c \nFalse negatives \n0 to 4 \n0",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summary of findings tables for TB-LAMP in children, giving true and false results per 1000 tested at different prevalence levels, with very low to moderate certainty and notes on downgrading.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-256",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 256",
      "pageStart": 256,
      "pageEnd": 256,
      "target": "../sections/page-256.md",
      "narrative": null,
      "content": "251 \n \ntuberculosis burden country. We downgraded one level for indirectness due to applicability concerns in \nother settings.  \nb. The study had a wide 95% CI (59-100%) for sensitivity. We downgraded one level for serious \ninconsistency. \nc. Too few children with pulmonary tuberculosis. We downgraded one level for serious imprecision.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explanatory notes completing the summary of findings table, citing downgrading for indirectness, inconsistency and imprecision.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-257",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 257",
      "pageStart": 257,
      "pageEnd": 257,
      "target": "../sections/page-257.md",
      "narrative": null,
      "content": "252 \n \nWeb Annex B.7. Parallel use of low-complexity automated nucleic \nacid amplification tests on respiratory samples and stool with or \nwithout lateral flow urine lipoarabinomannan assays to detect \npulmonary tuberculosis disease in children \n \nLaura Olbrich*, Stephanie Bjerrum, Bada Yang, Hayley Poore, Alia Razid, Brittney Sweetser, Johanna \nÅhsberg, Mathias Weis Damkjær, Alexander Kay, Andreas Lundh, Ruvandhi Nathavitharana, Ian Schiller, \nNandini Dendukuri, Maunank Shah, Devan Jaganath    \n• Laura Olbrich, Division of Infectious Diseases and Tropical Medicine, LMU University Hospital Mun",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Title page of Web Annex B.7, a systematic review of parallel use of automated nucleic acid amplification tests on respiratory samples and stool, with or without urine LF-LAM, to detect pulmonary TB in children, listing authors and affiliations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-258",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 258",
      "pageStart": 258,
      "pageEnd": 258,
      "target": "../sections/page-258.md",
      "narrative": null,
      "content": "253 \n \nEXECUTIVE SUMMARY \nBackground \nLow complexity automated nucleic acid amplification tests (LC-aNAATs) are part of the molecular World \nHealth Organization (WHO)-recommended rapid diagnostics widely used to diagnose tuberculosis \ndisease. They are recommended as the first test for symptomatic children presenting with presumptive  \ntuberculosis disease. The lateral flow urine lipoarabinomannan (LF-LAM) antigen test is recommended \nby WHO to assist in diagnosing tuberculosis disease among children with HIV. Previous systematic \nreviews have traditionally assessed diagnostic accuracy of LC-a",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Begins the executive summary of the review of parallel TB testing in children, explaining that automated NAATs and urine LF-LAM have mostly been assessed separately though used together in practice. It sets out the questions and reports that parallel testing of respiratory samples and stool increased sensitivity with a small loss of specificity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-259",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 259",
      "pageStart": 259,
      "pageEnd": 259,
      "target": "../sections/page-259.md",
      "narrative": null,
      "content": "254 \n \nparticipant, tuberculosis prevalence 40.9% [1,464/3,579]), the estimated diagnostic accuracy of the \nparallel use of LC-aNAAT on respiratory samples and stool had a pooled sensitivity of 29.0% (95% CrI \n17.3 to 44.7) and a pooled specificity of 99.1% (95% CrI 96.1 to 99.8). Compared to LC-aNAAT on \nrespiratory samples alone, parallel testing had 4.7 (95% CrI 2.1 to 8.9) percentage points higher \nsensitivity and -0.5 (95% CrI -1.4 to -0.1) percentage points lower specificity. Compared to LC-aNAAT on \nstool alone, parallel testing had 10.5 (95% CrI 6.9 to 15.0) percentage points higher se",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the executive summary with pooled accuracy for parallel respiratory and stool testing against a composite reference standard and results for children with HIV. It concludes that parallel testing improves sensitivity and may be most beneficial in high-prevalence settings, with greater risk of false positives and overtreatment where prevalence is low.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-260",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 260",
      "pageStart": 260,
      "pageEnd": 260,
      "target": "../sections/page-260.md",
      "narrative": null,
      "content": "255 \n \nrecommendation to consider repeat molecular testing in children if the pre-test probability of \npulmonary TB is 5% or more. \nAmong CHIV, the incremental increase in sensitivity of parallel stool and respiratory LC-aNAAT was \nlower than among children without HIV. This likely reflects the lower sensitivity in general of molecular \nTB testing in people living with HIV given their paucibacillary disease. The urine LF-LAM test is a non-\ninvasive, point-of-care test to detect tuberculosis and is currently recommended as an add-on test. We \ndemonstrated increased sensitivity when LF-LAM on ur",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Discusses that the sensitivity gain from parallel testing was smaller in children with HIV and that adding urine LF-LAM increased sensitivity further. It notes LF-LAM may be the only test available in some settings and that building capacity or referral for molecular testing should be considered.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-261",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 261",
      "pageStart": 261,
      "pageEnd": 261,
      "target": "../sections/page-261.md",
      "narrative": null,
      "content": "256 \n \n1. BACKGROUND   \n1.1 Target condition being diagnosed  \nOf the approximately 10 million annual tuberculosis cases worldwide, an estimated 1.2 million occur in \nchildren [World Health Organization 2022a]. Tuberculosis remains among the top ten causes of \nmortality in children, accounting for 14% and 11% of tuberculosis-related deaths in HIV-negative people \nand people  with HIV, respectively, largely due to missed diagnosis [World Health Organization 2022a, \n2018]. Especially children <5 years and immunocompromised children are at greatest risk of tuberculosis \nand tuberculosis -related",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes childhood TB as the review's target condition, with about 1.2 million cases in children annually and high mortality from missed diagnosis. It explains the limits of clinical and radiographic diagnosis and outlines WHO-recommended rapid tests, with a figure summarising tests by specimen type.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p262-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 262,
      "pageEnd": 262,
      "target": "../images/img-p262-1.webp",
      "narrative": "draft",
      "content": "Table listing TB diagnostic tests with acceptable specimen types and whether they detect rifampicin resistance. Xpert MTB/RIF accepts sputum, gastric fluid, nasopharyngeal aspirate, stool, CSF, lymph node aspirate or biopsy, pleural, peritoneal, pericardial and synovial fluid, urine and blood (yes); Xpert Ultra accepts sputum, gastric fluid, nasopharyngeal aspirate, stool, CSF and lymph node aspirate or biopsy (yes); Truenat MTB and MTB Plus accepts sputum (yes); TB-LAMP accepts sputum (no); LF-LAM accepts urine (no). Footnotes say blood is recommended for people living with HIV with disseminated TB and urine for children and adolescents living with HIV.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-262",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 262",
      "pageStart": 262,
      "pageEnd": 262,
      "target": "../sections/page-262.md",
      "narrative": null,
      "content": "257 \n \n \n \nSeveral clinical scores have been proposed to aid in establishing a diagnosis, especially at lower levels of \nhealth care where most children present [World Health Organization 2018]. In 2022, the WHO \nrecommended the use of treatment decision algorithms for the interim, with evidence of its \nperformance to be reviewed in 2024 [World Health Organization 2022b]. To be broadly applicable across \na range of health care levels, one treatment decision algorithms including CXR scoring and one without \nwere proposed. The algorithms are designed to achieve a high sensitivity at cost of spec",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes clinical scores and the treatment decision algorithms recommended by WHO in 2022, which favour sensitivity over specificity, and the need for better tests. It defines the parallel testing strategy and introduces the two index tests.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-263",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 263",
      "pageStart": 263,
      "pageEnd": 263,
      "target": "../sections/page-263.md",
      "narrative": null,
      "content": "258 \n \nurine. LC-aNAAT includes the Xpert MTB/RIF and Xpert MTB/RIF Ultra (Cepheid, Sunnyvale, United \nStates of America [USA]); Truenat MTB and MTB Plus (Molbio Diagnostics, Goa, India); STANDARD M10 \n(SD Biosensor, Republic of Korea); and IRON-qPCR (BIONEER, Republic of Korea). For LF-LAM, Determine \nTB LAM Ag (DetermineTM TB LAM Ag; Abbott, Chicago, IL, USA) is currently the only commercially \navailable LF-LAM test.  \n1.2.1 Index test, LC-aNAATs \nLC-aNAATs are a subgroup of mWRD used for tuberculosis diagnosis and detection of resistance to \nIsoniazid/Rifampicin. A number of commercial LC-a",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the automated NAATs assessed, including Xpert MTB/RIF and Ultra, Truenat, STANDARD M10 and IRON-qPCR, and the only commercial LF-LAM test. It explains methods for processing stool samples, such as the simple one-step and sucrose flotation methods.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-264",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 264",
      "pageStart": 264,
      "pageEnd": 264,
      "target": "../sections/page-264.md",
      "narrative": null,
      "content": "259 \n \nKit (SPK) containing several plastic consumables (filter device, stool transfer supplies) can be used jointly \nwith a specially prepared buffer [Banada 2016; Walters 2018]. Following preparation, the sample is \nfiltered instead of using sedimentation. A recently presented head-to-head comparison demonstrated a \nsimilar sensitivity with estimates ranging from 46.8% for the OSF and 48.3% for SOS to 52.1% for SPK, \nwhile specificity was 97% for all [Nabeta 2021]. \n1.2.2 Index test, LF-LAM \nUrine LF-LAM is a commercially available point-of-care test for tuberculosis disease. Lipoarabinomann",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the stool processing kit method and a head-to-head comparison of stool methods with similar sensitivity, then explains how the urine LF-LAM test detects a mycobacterial cell wall component. It argues there may be added value in combining WHO-recommended rapid tests for children with and without HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-265",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 265",
      "pageStart": 265,
      "pageEnd": 265,
      "target": "../sections/page-265.md",
      "narrative": null,
      "content": "260 \n \nsupported from past studies in children that have shown improved accuracy of parallel LC-aNAAT with \ndifferent respiratory sample types [Song 2021; Zar 2019]. \nThis systematic review will estimate the change in diagnostic accuracy of parallel use of LC-aNAATs \n(from respiratory specimens and stool) and LF-LAM compared to using each of the tests alone to \ndiagnose pulmonary tuberculosis among children with and without HIV. The findings will inform an \nupcoming WHO guideline development.  \n2. REVIEW QUESTION  \nWe addressed the following PICO (population, intervention, comparison, outcome)",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "States the review's aim to estimate the change in accuracy from parallel use of automated NAATs on respiratory samples and stool and urine LF-LAM, to inform WHO guidelines. It lists the PICO questions, including effects on patient-important outcomes such as cure, mortality and time to diagnosis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-266",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 266",
      "pageStart": 266,
      "pageEnd": 266,
      "target": "../sections/page-266.md",
      "narrative": null,
      "content": "261 \n \ntreatment initiation.  \n3. METHODS \nThis review was based on a protocol, which was informed by methodology used in previous reviews and \nother protocols in the field [Inbaraj 2023; Zifodya 2021; Kohli 2021; Shapiro 2021; Bjerrum 2019b; Shah \n2016b].  \n3.1. Criteria for considering studies for this review \n3.1.1. Types of studies \nFor children who were HIV negative or with unknown HIV status, we included comparative diagnostic \naccuracy studies that evaluated the accuracy of the parallel use of LC-aNAATs on a respiratory specimen \nand LC-aNAAT on stool, and the accuracy of one of the tes",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Begins the methods of a systematic review on parallel testing for pulmonary tuberculosis in children, defining eligible comparative diagnostic accuracy studies of low-complexity automated NAATs (LC-aNAATs) on respiratory samples and stool, plus urine LF-LAM for children with HIV. Table 5 sets out the primary comparisons and minimum study design requirements.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-267",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 267",
      "pageStart": 267,
      "pageEnd": 267,
      "target": "../sections/page-267.md",
      "narrative": null,
      "content": "262 \n \ntwo-by-two table data). Additionally, we reviewed fully paired studies that were originally designed to \nassess the accuracy of LF-LAM or LC-aNAAT on stool, with LC-aNAAT and microbiological culture on \nrespiratory specimens for the reference standard. We included these studies if data were made \navailable for reanalysis and estimation of comparative diagnostic accuracy. \nWe included studies with an attempt to obtain fully paired data for the index tests, but where data on \nsome participants were unavailable, e.g., either a respiratory sample or stool sample was not available \nfor LC-aN",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains that fully paired studies were included even when some participants lacked a sample or result, since samples are hard to obtain from children, and describes eligible participants and index tests. It notes the stool processing methods considered and begins listing the automated outputs of Xpert MTB/RIF and Xpert Ultra.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-268",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 268",
      "pageStart": 268,
      "pageEnd": 268,
      "target": "../sections/page-268.md",
      "narrative": null,
      "content": "263 \n \n1. MTB (Mycobacterium tuberculosis complex) DETECTED; RIF (rifampicin) resistance DETECTED \n2. MTB DETECTED; Rif resistance NOT DETECTED \n3. MTB DETECTED; Rif resistance INDETERMINATE \n4. MTB NOT DETECTED \n5. INVALID/ERROR/NO RESULT \nInvalid, error, and no result indicate that the presence or absence of M tuberculosis cannot be \ndetermined [A. W. Kay 2020]. Xpert MTB/ RIF Ultra also incorporates a semi-quantitative classification \nfor results: trace, very low, low, moderate, and high. We considered 'trace' as positive for M tuberculosis \nin children [World Health Organization 2017]. \nTr",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists the result categories of the Xpert assays, Truenat, MTB Plus, STANDARD M10, IRON-qPCR and LF-LAM, noting that Ultra 'trace' results were counted as positive in children. It then describes the microbiological reference standard and how its quality was assessed with QUADAS-2.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-269",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 269",
      "pageStart": 269,
      "pageEnd": 269,
      "target": "../sections/page-269.md",
      "narrative": null,
      "content": "264 \n \nWe included studies using a CRS consisting of a combination of microbiological tests and clinical criteria, \nincluding M. tuberculosis culture, NAATs on any sample type (pulmonary and extrapulmonary), or \nclinical criteria specified by the primary study authors. We included NAATs on non-respiratory sample \ntypes as children who have pulmonary tuberculosis often also have extrapulmonary TB. Clinical criteria \nto classify tuberculosis was defined by each study, and may be based on imaging studies, histology, \nclinical characteristics, and treatment response. Clinical characteristics may i",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Defines the composite reference standard, combining microbiological tests on any sample type with study-defined clinical criteria, and how TB presence or absence was classified. It then outlines the electronic literature search, limited to studies published after 2010.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-270",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 270",
      "pageStart": 270,
      "pageEnd": 270,
      "target": "../sections/page-270.md",
      "narrative": null,
      "content": "265 \n \nWe examined reference lists of included studies, relevant reviews, studies, commentaries, and editorials \nand searched the WHO websites related to HIV and tuberculosis. We contacted experts in the field and \nleading research groups working on new diagnostics such as SMART4TB, STOP TB and FIND (Foundation \nfor Innovative New Diagnostics). In addition, we reviewed studies submitted after a WHO call for \nliterature.  \n3.3 Data collection and analysis \n3.3.1 Selection of studies \nWe uploaded titles and abstracts retrieved from the electronic literature search to EndNote to remove \nduplicate",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes other search sources, such as expert groups and a WHO call for data, and how two reviewers independently screened records in Covidence and extracted data, including individual-participant data shared by study authors.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-271",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 271",
      "pageStart": 271,
      "pageEnd": 271,
      "target": "../sections/page-271.md",
      "narrative": null,
      "content": "266 \n \ninvolved in any of the included studies did not participate in data extraction of their own study. We \nextracted the data and recorded it using Microsoft® Excel and entered it into Review Manager Web \ndirectly [Collaboration 2020].  \n3.3.3 Assessment of methodological quality \nWe assessed the methodological quality of included studies using QUADAS-2 [Whiting 2011] and \nQUADAS-C [Yang 2021a] tools (Appendices 2 and 3). QUADAS-2 is a tool for assessing risk of bias in and \napplicability of single test accuracy estimates; QUADAS-C is an extension to QUADAS-2 for assessing risk \nof bias in",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains that data were managed in Excel and Review Manager and that study quality was assessed independently with the QUADAS-2 and QUADAS-C tools. It begins describing the statistical analysis, in which any positive result counted as positive for the parallel strategy.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-272",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 272",
      "pageStart": 272,
      "pageEnd": 272,
      "target": "../sections/page-272.md",
      "narrative": null,
      "content": "267 \n \ntesting pathway were regarded as ‘not positive’. This included (a) participants with an inadequate or \nunavailable specimen for testing and (b) participants with an invalid or technical failure (i.e. non-\ndeterminate test results). For the primary analysis, we analysed participants with a ‘not positive’ as test \nnegative. We performed a sensitivity analysis in which those participants with at least one missing or \nnon-determinate index test result were excluded (see below). \n3.3.5 Approach to unavailable or non-determinate results  \nTo avoid selection bias, we analysed the total number",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains that participants with missing or invalid index test results were treated as test-negative in the primary analysis, with a sensitivity analysis excluding them. It also defines investigations of heterogeneity, including subgroups by danger signs of serious illness in children with and without HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-007",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 7: Subgroup comparisons of interest for assessing heterogeneity",
      "pageStart": 272,
      "pageEnd": 272,
      "target": "../tables/table-7.csv",
      "narrative": "draft",
      "content": "Subgroup comparisons planned to investigate heterogeneity in the review of tests for TB in children. Each row is a variable (such as age, sample type, clinical setting, CD4 count or percentage, advanced HIV disease and serious illness) with its definition, the population it applies to and the comparison made.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-273",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 273",
      "pageStart": 273,
      "pageEnd": 273,
      "target": "../sections/page-273.md",
      "narrative": null,
      "content": "268 \n \n3.3.7 Sensitivity analysis \nWe performed the following sensitivity analyses by limiting inclusion in the meta-analysis based on the \nfollowing criteria: \n• \nOnly participants screening positive for tuberculosis using CXR  \n• \nExclusion of studies at high or unclear risk of bias (QUADAS-C) and high concerns regarding \napplicability (QUADAS-2) in at least one domain  \n• \nAlterative prior distributions which are less informative: increasing the variance of the normal \ndistributions over the summary log odds of the sensitivity or specificity to 100 and using a uniform \ndistribution ranging",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists the sensitivity analyses performed and explains why publication bias was not formally assessed and how certainty of evidence was graded with GRADE. It begins the results, reporting that 155 unique records were screened and 42 full texts assessed.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-274",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 274",
      "pageStart": 274,
      "pageEnd": 274,
      "target": "../sections/page-274.md",
      "narrative": null,
      "content": "269 \n \nsix studies. We included 36 unique studies, all of which had data on a parallel strategy of LC-aNAAT on \nrespiratory specimens and LC-aNAAT on stool, and additionally LF-LAM in CHIV; however, none of the \nincluded studies reported on parallel accuracy. We did not identify studies on impact/patient-important \noutcomes. Data extraction directly from the publication was not feasible for any of the included studies. \nOn request, authors provided data re-analyzed for the purpose of this review for 14 studies, 12 of which \nmade data available for PICO 1 and six for PICO 2 (i.e., 4 studies add",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that 36 studies were included but none reported parallel-testing accuracy or patient-important outcomes, so authors re-analysed data for 14 studies across the two review questions.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-vfig-p275",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 275,
      "pageEnd": 275,
      "target": "../figures/vfig-p275.png",
      "narrative": "draft",
      "content": "A PRISMA flow diagram of studies in the review: 170 studies from databases/registers (MEDLINE 130, ClinicalTrials.gov 39, WHO public call for data 1) and 0 from other sources; 15 duplicates removed; 155 screened and 113 excluded; 42 sought and assessed for eligibility, 6 excluded and data not retrievable for 22; 14 studies included (PICO 1: 8, PICO 2: 2, PICO 1 and 2: 4). In this render, some boxes overlap and surrounding page text is drawn over the lower part.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-275",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 275",
      "pageStart": 275,
      "pageEnd": 275,
      "target": "../sections/page-275.md",
      "narrative": null,
      "content": "270 \n \n  \n  \nFigure 2: Flow of studies in the review. \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n4.2  Methodological quality of included studies  \n4.2.1. QUADAS- 2 – PICO 1 \nFigure 3a show risk of bias and applicability concerns for each of the 12 included studies for children \nwithout HIV or with HIV status unknown, while Figure 3b shows the same for the seven studies included \nfor CHIV.  \n \nFigure 3: Risk of bias and applicability concerns related to single test accuracy estimates (QUADAS-2), PICO 1. Review authors' \njudgements about each domain for each inclu",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Shows the study flow diagram (Figure 2) and introduces the QUADAS-2 risk-of-bias and applicability figures for studies of children without HIV and children with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p276-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 276,
      "pageEnd": 276,
      "target": "../images/img-p276-1.webp",
      "narrative": "draft",
      "content": "Two-panel risk of bias and applicability grid. Panel a lists 12 studies (Ainan 2021 to Wobudeya 2024) and panel b lists 7 studies (Chipinduro 2017 to Orikiriza 2021); columns cover risk-of-bias domains including patient selection, index test, reference standards (MRS/CRS or microbiological/composite) and flow and timing, plus three applicability domains. Cells are coloured low (green), high (red), unclear (yellow) or not applicable (grey); the composite reference standard column is red for most studies.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-276",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 276",
      "pageStart": 276,
      "pageEnd": 276,
      "target": "../sections/page-276.md",
      "narrative": null,
      "content": "271 \n \n \n \nFor children without HIV or HIV unknown, in the patient selection domain, we judged 11 studies (91.6%) \nto have low risk of bias. One study (8.3%; [Donkeng 2022]) had unclear risk of bias. Regarding \napplicability, eight studies (66.7%) had low concern because the studies included the appropriate \nparticipants and settings  [Ainan 2021; Babo 2023; Hanrahan 2018; Hasan 2017; Jaganath 2024; Olbrich \n2024; M. P. Nicol 2013; Wobudeya 2024]. We judged three studies (25.0%) to have high concern, as \nthese studies exclusively recruited participants in inpatient settings [Chabala 2024; Kabi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports QUADAS-2 judgements for PICO 1, finding most studies at low risk of bias in patient selection, with applicability concerns for those recruiting only inpatients. It also covers the index test, reference standard and flow and timing domains.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-277",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 277",
      "pageStart": 277,
      "pageEnd": 277,
      "target": "../sections/page-277.md",
      "narrative": null,
      "content": "272 \n \nparticipants were included in the analysis. Four studies had unclear risk of bias as we could not \ndetermine if all index tests were collected at the same time as the reference standard (33.3%; [Ainan \n2021; Chabala 2024; Orikiriza 2021; Wobudeya 2024]). \n4.2.2. QUADAS-C – PICO 1 \nFigure 4a show risk of bias in the comparative accuracy estimates for each of the 12 studies comparing \nthe parallel use of LC-aNAAT on respiratory samples plus LC-aNAAT on stool to each of those alone for \nchildren without HIV and Figure 4b for CHIV.  \n \n \n \n \n \n \n \n \n \n \nFigure 4: Risk of bias in comparative",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the flow and timing judgements for PICO 1 and introduces Figure 4, the QUADAS-C assessment of bias in the comparative accuracy estimates.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p278-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 278,
      "pageEnd": 278,
      "target": "../images/img-p278-1.webp",
      "narrative": "draft",
      "content": "Two-panel risk-of-bias grid with five domains: patient selection, index test, micro reference standard, comp reference standard, and flow and timing. Panel a covers 12 studies (Ainan 2021 to Wobudeya 2024) and panel b 7 studies (Chipinduro 2017 to Orikiriza 2021); cells are low (green, +), high (red, −), unclear (yellow, ?) or not applicable (grey, NA). Most cells are low, but the composite reference standard is rated high in nearly every study, with unclear ratings in Hanrahan 2018 (a) and Jaganath 2024 (b).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-278",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 278",
      "pageStart": 278,
      "pageEnd": 278,
      "target": "../sections/page-278.md",
      "narrative": null,
      "content": "273 \n \n \n \nAll studies for the purpose of this review used a paired diagnostic accuracy design.  \nIn the patient selection domain, we judged eleven studies (91.7%) for studies in children without HIV \nand five (71.4%) for CHIV to have low risk of bias [Chipinduro 2017; Jaganath 2024; LaCourse 2018; \nOlbrich 2024; Orikiriza 2021]. One study (8.3%,[Donkeng 2022]) for children without HIV and two (28.6% \n[Donkeng 2022; Hanrahan 2018]) for CHIV had unclear risk of bias, as the risk of patient selection bias \nfor both index tests was judged unclear per QUADAS-2 due to the patient recruitment not be",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports QUADAS-C judgements for PICO 1, finding most studies at low risk of bias for patient selection and index tests. Studies using a composite reference standard that included the index NAATs were judged at high risk of bias in that domain.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-vfig-p279",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 279,
      "pageEnd": 279,
      "target": "../figures/vfig-p279.png",
      "narrative": "draft",
      "content": "A QUADAS-2 table of risk of bias (patient selection, index test (parallel), index test (LAM), micro reference standard, comp reference standard, flow and timing) and applicability concerns for six studies: Chipinduro 2017, Hanrahan 2018, Jaganath 2024, LaCourse 2018, Olbrich 2024 and Orikiriza 2021. Most cells are green (low); the composite reference standard is rated high (red) in five studies.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-279",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 279",
      "pageStart": 279,
      "pageEnd": 279,
      "target": "../sections/page-279.md",
      "narrative": null,
      "content": "274 \n \nwere judged to have a high risk of bias. \nIn the flow and timing domain, we judged nine studies (75.0 %, [Ainan 2021; Babo 2023; Hanrahan 2018; \nHasan 2017; Jaganath 2024; Kabir 2021; M. P. Nicol 2013; Olbrich 2024; Orikiriza 2021]) for children \nwithout HIV and five (71.4%, [Chipinduro 2017; Jaganath 2024; LaCourse 2018; Olbrich 2024; Orikiriza \n2021]) for CHIV to have low risk of bias for the comparison, as all participants were included in the \nanalyses. Three (25.0%, [Chabala 2024; Donkeng 2022; Wobudeya 2024]) studies for children without \nHIV and two (28.7%, [Donkeng 2022; Hanraha",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the QUADAS-C flow and timing results for PICO 1 and introduces Figure 5, the QUADAS-2 assessment for the six PICO 2 studies, including the figure's risk-of-bias grid.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-280",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 280",
      "pageStart": 280,
      "pageEnd": 280,
      "target": "../sections/page-280.md",
      "narrative": null,
      "content": "275 \n \nOlbrich 2024; Orikiriza 2021].  \nRegarding applicability of the reference standard, we thought that all studies had low concerns as all \nperformed a test to identify M tuberculosis species (speciation) for the MRS, and appropriate criteria for \nCRS.  \nIn the flow and timing domain, we judged five studies (83.3%) to be at low risk of bias noting that all \nparticipants were included in the analysis. One study had unclear risk of bias as we could not tell (16.7%; \n[Orikiriza 2021]). \n4.2.4. QUADAS-C – PICO 2 \nFigure 6 shows risk of bias in comparative accuracy estimates for six studies com",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes low applicability concerns for reference standards in PICO 2 studies and introduces Figure 6, the QUADAS-C assessment comparing parallel NAAT plus LF-LAM testing with each test alone.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-281",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 281",
      "pageStart": 281,
      "pageEnd": 281,
      "target": "../sections/page-281.md",
      "narrative": null,
      "content": "276 \n \n \n \nAll studies for the purpose of this review used a paired diagnostic accuracy design.  \nIn the index test domain, we judged five studies (83.3%) to have low risk of bias as the risk of bias for \neach index test was judged low in QUADAS-2, while we could not tell for one study (16.7%; [Chipinduro \n2017]), and thus judged unclear risk of bias.  \nIn the reference standard domain, we judged all studies based on a MRS to have low risk of bias as the \nstudies avoided incorporating LC-aNAAT as part of the MRS. For CRS, we judged five studies (83.3%) to \nhave high risk of bias because they i",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports QUADAS-C judgements for PICO 2, with high risk of bias for composite reference standards that included NAATs, and begins the description of key study characteristics, followed by the risk-of-bias grid.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-282",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 282",
      "pageStart": 282,
      "pageEnd": 282,
      "target": "../sections/page-282.md",
      "narrative": null,
      "content": "277 \n \n(eight, 66.7%) recruited from 10 countries within Sub-Saharan Africa. \nAll 14 studies included symptomatic, but five studies also enrolled participants based on CXR \nabnormalities, being seriously ill or with advanced HIV disease.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "States that most studies recruited in sub-Saharan Africa and that all enrolled symptomatic children, some also by chest X-ray, serious illness or advanced HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-283",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 283",
      "pageStart": 283,
      "pageEnd": 283,
      "target": "../sections/page-283.md",
      "narrative": null,
      "content": "278 \n \n4.4. PICO 1: What is the incremental diagnostic accuracy of parallel use of LC-\naNAATs on a respiratory sample and stool for diagnosis of pulmonary \ntuberculosis disease in children that are HIV negative or unknown HIV status \nwith signs and symptoms or screened positive for pulmonary tuberculosis, \ncompared to use of the LC-aNAAT on one sample type (either respiratory or \nstool)?   \n4.4.1 Comparative diagnostic accuracy against a microbiological reference \nstandard  \nSummary of Findings Tables for the incremental diagnostic accuracy of the parallel use of LC-aNAATs \ncombined versus eac",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents PICO 1 results against a microbiological reference standard, describing the eight studies comparing parallel respiratory and stool NAAT testing with either sample alone. It reports study numbers and TB prevalence and includes forest plots of sensitivity and specificity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p284-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 284,
      "pageEnd": 284,
      "target": "../images/img-p284-1.webp",
      "narrative": "draft",
      "content": "Forest plot, panel a, of 8 studies (Chishala 2024 to Orikiriza 2021) giving TP, FP, FN, TN, sensitivity and specificity (95% CI), with black squares and CI lines on 0–1 scales and a summary diamond. The summary estimate is sensitivity 0.80 (0.68–0.90) and specificity 0.93 (0.87–0.97).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p284-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 284,
      "pageEnd": 284,
      "target": "../images/img-p284-2.webp",
      "narrative": "draft",
      "content": "Forest plot, panel b, of 7 studies (Chipinduro 2017 to Orikiriza 2021) giving TP, FP, FN, TN, sensitivity and specificity (95% CI), with black squares and CI lines on 0–1 scales and a summary diamond. Hanrahan 2018 sensitivity is shown as NaN; the summary estimate is sensitivity 0.70 (0.51–0.85) and specificity 0.95 (0.92–0.98).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-284",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 284",
      "pageStart": 284,
      "pageEnd": 284,
      "target": "../sections/page-284.md",
      "narrative": null,
      "content": "279 \n \n \n \n \n \nDiagnostic accuracy of the component tests: LC-aNAAT on respiratory sample or \nLC-aNAAT on stool based on the microbiological reference standard  \nFor children without HIV or HIV status unknown, the sensitivity of LC-aNAAT on respiratory samples \nalone ranged from 47.8% [Olbrich 2024] to 100% [Hanrahan 2018] and specificity from 84.5% [M. P.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Begins reporting the range of sensitivity and specificity of NAAT on respiratory samples alone in children without HIV or with unknown status.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-285",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 285",
      "pageStart": 285,
      "pageEnd": 285,
      "target": "../sections/page-285.md",
      "narrative": null,
      "content": "280 \n \nNicol 2013] to 100% [Donkeng 2022] between studies. The pooled sensitivity and specificity for LC-\naNAAT on respiratory samples alone were 72.6% (95% CrI 59.0 to 84.6) and 95.0% (95% CrI 90.6 to 97.7). \nThe sensitivity of LC-aNAAT on stool alone ranged from 0.0% [Hanrahan 2018] to 85.7% [Hasan 2017] \nand specificity from 89.2% [M. P. Nicol 2013] to 100% [Hanrahan 2018; Hasan 2017; Donkeng 2022] \nbetween studies. The pooled sensitivity and specificity for LC-aNAAT on stool alone were 56.3% (95%CrI \n42.9 to 70.0) and 97.5% (95% CrI 94.8 to 99.0), (eight studies; 2,145 participants, 173 [8",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports pooled accuracy of NAAT on respiratory samples alone and on stool alone, against a microbiological reference standard, for children without HIV and with HIV. It then gives the pooled gain in sensitivity and small loss of specificity from parallel testing compared with respiratory testing alone.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p286-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 286,
      "pageEnd": 286,
      "target": "../images/img-p286-1.webp",
      "narrative": "draft",
      "content": "Forest plots of study-specific differences in sensitivity and specificity (parallel LC-aNAAT testing versus LC-aNAAT on respiratory samples alone) for 8 studies, Chishala 2024 to Orikiriza 2021, with total N and risk-of-bias ratings (P, I, R, FT) shown as coloured dots on the right. Differences in sensitivity range from 0 to 0.143 and in specificity from −0.08 to 0; the summary difference is 0.072 (95% CrI −0.0971 to 0.241) for sensitivity and −0.0162 (−0.0831 to 0.0396) for specificity.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-286",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 286",
      "pageStart": 286,
      "pageEnd": 286,
      "target": "../sections/page-286.md",
      "narrative": null,
      "content": "281 \n \nFigure 8: Forest plots of study-specific and summary estimates of difference in accuracy between the parallel use of LC-aNAATs combined versus LC-aNAAT on respiratory samples \nalone for tuberculosis in children that are HIV negative or unknown HIV status with presumed pulmonary tuberculosis by MRS. The black squares represent the point estimate of \none study, the black horizontal line its 95% confidence interval (CI). The QUADAS-C risk-of-bias judgements by study are included on the right. Note that study-specific 95% CIs \nsometimes contain impossible values (difference in sensitivity <",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Caption for Figure 8, forest plots of the difference in accuracy between parallel NAAT testing and respiratory-sample testing alone in children without HIV, with notes on the confidence intervals.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p287-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 287,
      "pageEnd": 287,
      "target": "../images/img-p287-1.webp",
      "narrative": "draft",
      "content": "Forest plot of pooled sensitivity and specificity (95% CrI) for LC-aNAAT on stool, LC-aNAAT on respiratory samples and parallel testing (N 2145 with 173 TB cases by MRS; about 3579 with 1513 cases by CRS), shown separately by microbiological (MRS) and composite (CRS) reference standard, with differences (Δ parallel vs. resp, Δ parallel vs. stool). By MRS, sensitivity is 56.3% (stool), 72.6% (resp) and 79.9% (parallel), with specificity 93.4–97.5%; by CRS, sensitivity is 16.4%, 21.7% and 29.0%, with specificity 99.1–99.7%.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p287-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 287,
      "pageEnd": 287,
      "target": "../images/img-p287-2.webp",
      "narrative": "draft",
      "content": "Forest plot of pooled sensitivity and specificity (95% CrI) for LC-aNAAT on stool, LC-aNAAT on respiratory samples and parallel testing (N 697 with 44 TB cases by MRS; 718 with 287 cases by CRS), shown separately by microbiological (MRS) and composite (CRS) reference standard, with differences (Δ parallel vs. resp, Δ parallel vs. stool). By MRS, sensitivity is 60.8% (stool), 66.4% (resp) and 70.2% (parallel), with specificity 95.4–97.2%; by CRS, sensitivity is 12.6%, 14.6% and 16.1%, with specificity 95.5–98.3%.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-287",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 287",
      "pageStart": 287,
      "pageEnd": 287,
      "target": "../sections/page-287.md",
      "narrative": null,
      "content": "282 \n \nFigure 9: Forest plot of pooled sensitivity and specificity for all studie in children without HIV or HIV unknown, by each index test. \nThe squares represent the pooled sensitivity and specificity, the black horizontal line its 95% credible interval (CrI). The pooled \ndifference in sensitivity and specificity between parallel testing and LC-aNAAT alone is indicated by a line connecting two \nsquares. This pooled difference may not correspond to the difference between the pooled single test accuracy estimates (see \n3.3.4: Statistical analysis and data synthesis). \n \n \nFigure 10: Forest pl",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Captions for Figures 9 and 10, forest plots of pooled sensitivity and specificity by index test for children without HIV and children with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-288",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 288",
      "pageStart": 288,
      "pageEnd": 288,
      "target": "../sections/page-288.md",
      "narrative": null,
      "content": "283 \n \nIncremental diagnostic accuracy of parallel use of LC-aNAATs combined versus \nLC-aNAAT on stool alone based on the microbiological reference standard \nIn children without HIV or HIV status unknown, incremental difference in sensitivity for parallel use of \nLC-aNAATs versus LC-aNAAT on stool alone ranged from 6.7% [Chabala 2024] to 31.7% [Jaganath 2024] \nbetween studies. The difference in specificity ranged from 0% [Donkeng 2022; Orikiriza 2021] to -13.0% \n[M. P. Nicol 2013]. See Figure 11.  The pooled difference in sensitivity was 23.5% (95% CrI 5.8 to 40.3; PI \n8.0 to 46.9) and −4.1% (",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that, against a microbiological reference standard, parallel testing raised pooled sensitivity by about 23.5% over stool testing alone in children without HIV, with a small specificity loss, and by about 8.5% in children with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p289-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 289,
      "pageEnd": 289,
      "target": "../images/img-p289-1.webp",
      "narrative": "draft",
      "content": "Forest plots of study-specific differences in sensitivity and specificity (parallel LC-aNAAT testing versus LC-aNAAT on stool alone) for 8 studies, Chishala 2024 to Orikiriza 2021, with total N and risk-of-bias ratings (P, I, R, FT) shown as coloured dots on the right. Sensitivity gains range from 0.067 to 1 (Hanrahan 2018) and specificity differences from −0.13 to 0; the summary difference is 0.2349 (95% CrI 0.0559 to 0.403) for sensitivity and −0.0405 (−0.105 to 0.004) for specificity.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-289",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 289",
      "pageStart": 289,
      "pageEnd": 289,
      "target": "../sections/page-289.md",
      "narrative": null,
      "content": "284 \n \nFigure 11: Forest plots of study-specific and summary estimates of difference in accuracy between the parallel use of LC-aNAATs versus LC-aNAAT on stool alone for tuberculosis \nin children that are HIV negative or unknown HIV status with presumed pulmonary tuberculosis by MRS. The black squares represent the point estimate of one study, the black \nhorizontal line its 95% confidence interval (CI). The QUADAS-C risk-of-bias judgements by study are included on the right. Note that study-specific 95% CIs sometimes contain \nimpossible values (difference in sensitivity < 0 and difference in s",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Caption for Figure 11, forest plots of the difference in accuracy between parallel NAAT testing and stool testing alone in children without HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-290",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 290",
      "pageStart": 290,
      "pageEnd": 290,
      "target": "../sections/page-290.md",
      "narrative": null,
      "content": "285 \n \n4.4.2 Comparative diagnostic accuracy against a composite reference standard \nSummary of Findings Tables for the incremental diagnostic accuracy of the parallel use of LC-aNAATs \ncombined versus each LC-aNAAT alone using the CRS can be found in Appendix 12. \nDescription of included studies \nWe identified 12 studies that compared the accuracy of a parallel strategy of LC-aNAATs combined \nversus each LC-aNAAT alone in children without HIV or HIV status unknown against a CRS, involving \n3,579 participants, 1,464 (40.9%) of which had tuberculosis disease. In CHIV, seven studies assessed the",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Begins the PICO 1 analysis against a composite reference standard, describing 12 studies in children without HIV and seven in children with HIV, with TB prevalence ranging widely. It notes that composite reference definitions varied and includes forest plots of accuracy.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p291-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 291,
      "pageEnd": 291,
      "target": "../images/img-p291-1.webp",
      "narrative": "draft",
      "content": "Forest plot, panel a, of 11 studies (Ainan 2021 to Wobudeya 2024) giving TP, FP, FN, TN, sensitivity and specificity (95% CI), with black squares and CI lines on 0–1 scales and a summary diamond. Sensitivity is generally low (0.05–0.88) and specificity high (0.77–1.00); the summary estimate is sensitivity 0.29 (0.17–0.45) and specificity 0.99 (0.96–1.00).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p291-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 291,
      "pageEnd": 291,
      "target": "../images/img-p291-2.webp",
      "narrative": "draft",
      "content": "Forest plot, panel b, of 7 studies (Chipinduro 2017 to Orikiriza 2021) giving TP, FP, FN, TN, sensitivity and specificity (95% CI), with black squares and CI lines on 0–1 scales and a summary diamond. Sensitivity ranges from 0.00 to 1.00 and specificity from 0.79 to 1.00; the summary estimate is sensitivity 0.16 (0.08–0.31) and specificity 0.95 (0.90–0.98).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-291",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 291",
      "pageStart": 291,
      "pageEnd": 291,
      "target": "../sections/page-291.md",
      "narrative": null,
      "content": "286 \n \n \n \n \n \nDiagnostic accuracy of the component tests: LC-aNAAT on respiratory sample or \nLC-aNAAT on stool based on the composite reference standard  \nIn children without HIV or HIV status unknown, the sensitivity of LC-aNAAT on respiratory samples alone \nranged from 44.6% ([Hanrahan 2018]) to 81.2% [Donkeng 2022] between studies, and specificity was at \n100% for all studies except one study where specificity was 78.4% [Jaganath 2024]. The pooled",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Begins reporting the sensitivity and specificity ranges of respiratory-sample NAAT testing against the composite reference standard in children without HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-292",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 292",
      "pageStart": 292,
      "pageEnd": 292,
      "target": "../sections/page-292.md",
      "narrative": null,
      "content": "287 \n \nsensitivity was 21.7% (95% CrI 12.8 to 34.4) and the specificity 99.8% (95% CrI 97.9 to 100), (12 studies; \n3,579 participants, 1,464 [40.9%] with tuberculosis). The sensitivity of LC-aNAAT on stool alone ranged \nfrom 0.0% [Hanrahan 2018] to 68.8% [Donkeng 2022] and specificity from 93.5% [Olbrich 2024] to 100% \n[Ainan 2021; Chabala 2024; Donkeng 2022; Hanrahan 2018; Kabir 2021; Orikiriza 2021; Wobudeya 2024] \nbetween studies. The pooled sensitivity was 16.4% (95% CrI 8.4 to 29.4) and the specificity 99.3% (95% \nCrI 97.9 to 99.8), (12 studies; 3,579 participants, 1,464 [40.9%] with tube",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports pooled sensitivity and specificity of respiratory and stool NAAT testing against the composite reference standard, which gives much lower sensitivity estimates than the microbiological standard. It then gives the pooled differences for parallel testing versus respiratory testing alone in children with and without HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p293-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 293,
      "pageEnd": 293,
      "target": "../images/img-p293-1.webp",
      "narrative": "draft",
      "content": "Forest plots of study-specific differences in sensitivity and specificity (parallel LC-aNAAT testing versus LC-aNAAT on respiratory samples alone) for 11 studies, Ainan 2021 to Wobudeya 2024, with total N and risk-of-bias ratings (P, I, R, FT) shown as coloured dots on the right. Sensitivity differences range from 0 to 0.36 (Kabir 2021) and specificity differences from −0.06 to 0; the summary difference is 0.072 (95% CrI −0.0967 to 0.253) for sensitivity and −0.0057 (−0.0356 to 0.0121) for specificity. The reference-standard (R) column is red for most studies.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-293",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 293",
      "pageStart": 293,
      "pageEnd": 293,
      "target": "../sections/page-293.md",
      "narrative": null,
      "content": "288 \n \nFigure 13: Forest plots of study-specific and summary estimates of difference in accuracy between the parallel use of LC-aNAATs combined versus LC-aNAAT on respiratory samples \nalone for tuberculosis in children that are (a) HIV negative or unknown HIV status and (b) CHIV with presumed pulmonary tuberculosis by a CRS. The black squares represent the \npoint estimate of one study, the black horizontal line its 95% confidence interval (CI). The QUADAS-C risk-of-bias judgements by study are included on the right. Note that study-\nspecific 95% CIs sometimes contain impossible values (differe",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Caption for Figure 13, forest plots of the difference in accuracy between parallel NAAT testing and respiratory testing alone against a composite reference standard.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-294",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 294",
      "pageStart": 294,
      "pageEnd": 294,
      "target": "../sections/page-294.md",
      "narrative": null,
      "content": "289 \n \nIncremental accuracy of the parallel use of LC-aNAATs combined versus LC-\naNAAT on stool alone based on the composite reference standard \nFor children without HIV or HIV status unknown, within studies, incremental difference in sensitivity for \nparallel use of LC-aNAATs combined versus LC-aNAAT on stool alone ranged from 4.1% [Orikiriza 2021] \nto 31.8% [Chabala 2024]. The difference in specificity for parallel use of LC-aNAATs versus LC-aNAAT on \nstool alone was 0.0% for all studies [Ainan 2021; Babo 2023; Chabala 2024; Donkeng 2022; Hanrahan \n2018; Kabir 2021; M. P. Nicol 2013; Olbrich",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that, against a composite reference standard, parallel testing increased pooled sensitivity by about 10.5% over stool testing alone in children without HIV with almost no loss of specificity, and gives the corresponding results for children with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p295-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 295,
      "pageEnd": 295,
      "target": "../images/img-p295-1.webp",
      "narrative": "draft",
      "content": "Forest plots of study-specific differences in sensitivity and specificity (parallel LC-aNAAT testing versus LC-aNAAT on stool alone) for 11 studies, Ainan 2021 to Wobudeya 2024, with total N and risk-of-bias ratings (P, I, R, FT) shown as coloured dots on the right. Sensitivity differences range from 0.041 to 0.318 and specificity differences are mostly 0 (−0.19 for Jaganath); the summary difference is 0.1237 (95% CrI −0.046 to 0.300) for sensitivity and −0.0016 (−0.032 to 0.014) for specificity.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-295",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 295",
      "pageStart": 295,
      "pageEnd": 295,
      "target": "../sections/page-295.md",
      "narrative": null,
      "content": "290 \n \nFigure 14: Forest plots of study-specific and summary estimates of difference in accuracy between the parallel use of LC-aNAATs combined versus LC-aNAAT on stool alone for \ntuberculosis in children that are HIV negative or unknown HIV status with presumed pulmonary tuberculosis by a CRS. The black squares represent the point estimate of one study, \nthe black horizontal line its 95% confidence interval (CI). The QUADAS-C risk-of-bias judgements by study are included on the right. Note that study-specific 95% CIs sometimes \ncontain impossible values (difference in sensitivity < 0 and diff",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Caption for Figure 14, forest plots of the difference in accuracy between parallel NAAT testing and stool testing alone against a composite reference standard.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-296",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 296",
      "pageStart": 296,
      "pageEnd": 296,
      "target": "../sections/page-296.md",
      "narrative": null,
      "content": "291 \n \n4.4.3 Subgroup analyses using a microbiological reference standard \nThe diagnostic accuracy of the parallel use of LC-aNAAT combined, LC-aNAAT on respiratory samples, \nand LC-aNAAT on stool in children without HIV or HIV status unknown is shown for the following \nsubgroups against a MRS in Appendix 16: inpatient setting; symptomatic participants; age 0-4 versus 5-9 \nyears; sample type NPA versus other respiratory samples; severe acute malnutrition present.   \n4.4.4 Non‐determinate results \nSix studies reported invalid results for either LC-aNAAT on respiratory samples or stool in childr",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists subgroup analyses for PICO 1, such as inpatients, age group, sample type and severe acute malnutrition, and reports the low rates of invalid NAAT results. It also gives sensitivity-analysis results that exclude participants with missing or invalid results.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-297",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 297",
      "pageStart": 297,
      "pageEnd": 297,
      "target": "../sections/page-297.md",
      "narrative": null,
      "content": "292 \n \n4.5. PICO 2: What is the incremental diagnostic accuracy of parallel use of LC-\naNAATs on a respiratory sample and stool and LF-LAM on urine for diagnosis of \npulmonary tuberculosis disease in children with HIV with signs and symptoms \nor screened positive for pulmonary tuberculosis, compared to any of the tests \n(either LC-aNAATs combined or LF-LAM) alone? \n4.5.1 Comparative diagnostic accuracy against a microbiological reference \nstandard  \nDescription of included studies \nWe identified six studies that compared a parallel strategy of LC-aNAAT on respiratory specimen plus LC-\naNAAT on",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces PICO 2 on parallel respiratory and stool NAAT plus urine LF-LAM testing in children with HIV, describing six studies with 653 participants. Its forest-plot caption refers to adults and adolescents, which seems inconsistent with the child population.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p298-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 298,
      "pageEnd": 298,
      "target": "../images/img-p298-1.webp",
      "narrative": "draft",
      "content": "Forest plot of 6 studies (Chipinduro 2017 to Orikiriza 2021) giving TP, FP, FN, TN, sensitivity and specificity (95% CI), with black squares and CI lines on 0–1 scales and a summary diamond. Hanrahan 2018 sensitivity is NaN; other sensitivities range 0.64–0.89 and specificities 0.63–0.92; the summary estimate is sensitivity 0.78 (0.60–0.90) and specificity 0.84 (0.74–0.90).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-298",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 298",
      "pageStart": 298,
      "pageEnd": 298,
      "target": "../sections/page-298.md",
      "narrative": null,
      "content": "293 \n \n \n \nDiagnostic accuracy of the component tests: LC-aNAATs combined or urine LF-\nLAM based on the microbiological reference standard  \nThe sensitivity of LC-aNAATs combined alone ranged from 25.0% [Orikiriza 2021] to 85.7% [Jaganath \n2024] and the specificity from 90.0% ([Olbrich 2024]) to 100% [Hanrahan 2018; Orikiriza 2021] between \nstudies. The pooled sensitivity and specificity for LC-aNAATs combined alone were 69.3% (95% CrI 49.9 \nto 84.0) and 95.4% (95% CrI 91.1 to 98.0), (six studies, 653 participants, 43 [6.6%] of which had \ntuberculosis). \nThe sensitivity of LF-LAM alone ranged",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports pooled accuracy for combined NAATs alone and for LF-LAM alone in children with HIV, against a microbiological reference standard, and the pooled gain in sensitivity and loss of specificity from adding LF-LAM.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p299-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 299,
      "pageEnd": 299,
      "target": "../images/img-p299-1.webp",
      "narrative": "draft",
      "content": "Forest plots of study-specific differences in sensitivity and specificity (parallel LC-aNAAT test plus LF-LAM versus parallel LC-aNAAT test) for 6 studies, Chipinduro 2017 to Orikiriza 2021, with total N and risk-of-bias ratings (P, I, R, FT) shown as coloured dots on the right. Sensitivity differences range from 0 to 0.5 (Hanrahan 2018 NaN) and specificity differences from −0.28 to −0.02; the summary difference is 0.0816 (95% CrI −0.1482 to 0.3123) for sensitivity and −0.1142 (−0.2172 to −0.0368) for specificity.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-299",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 299",
      "pageStart": 299,
      "pageEnd": 299,
      "target": "../sections/page-299.md",
      "narrative": null,
      "content": "294 \n \nFigure 16:  Forest plots of study-specific and summary estimates of difference in accuracy between the parallel use of LC-aNAATs combined and LF-LAM versus LC-aNAATs \ncombined alone for tuberculosis in children with HIV with presumed pulmonary tuberculosis by MRS. The black squares represent the point estimate of one study, the black \nhorizontal line its 95% confidence interval (CI). The QUADAS-C risk-of-bias judgements by study are included on the right. Note that study-specific 95% CIs sometimes contain \nimpossible values (difference in sensitivity < 0 and difference in specificity >",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Caption for Figure 16, forest plots of the difference in accuracy between combined NAATs plus LF-LAM and combined NAATs alone in children with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p300-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 300,
      "pageEnd": 300,
      "target": "../images/img-p300-1.webp",
      "narrative": "draft",
      "content": "Forest plot of pooled sensitivity and specificity (95% CrI) for LF-LAM, combined LC-aNAATs and parallel testing, by microbiological (MRS; N 653, 43 TB cases) and composite (CRS; N 674, 286 cases) reference standard, with differences for parallel testing. By MRS, sensitivity is 27.4% (LF-LAM), 69.3% (combined LC-aNAATs) and 77.6% (parallel), with specificity 88.2%, 95.4% and 83.9%; by CRS, sensitivity is 17.7%, 14.8% and 30%, with specificity 88.7%, 95.5% and 83.3%.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-300",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 300",
      "pageStart": 300,
      "pageEnd": 300,
      "target": "../sections/page-300.md",
      "narrative": null,
      "content": "295 \n \nFigure 17: Forest plot of pooled sensitivity and specificity for all studies, by each index test. The squares represent the pooled \nsensitivity and specificity, the black horizontal line its 95% credible interval (CrI). The pooled difference in sensitivity and \nspecificity between parallel testing and LC-aNAAT alone is indicated by a line connecting two squares. This pooled difference may \nnot correspond to the difference between the pooled single test accuracy estimates (see 3.3.4: Statistical analysis and data \nsynthesis). \n \n \nIncremental diagnostic accuracy LC-aNAATs combined and LF",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Caption for Figure 17, then reports that parallel testing raised sensitivity by about 50% over LF-LAM alone, and gives component test accuracy against a composite reference standard.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p301-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 301,
      "pageEnd": 301,
      "target": "../images/img-p301-1.webp",
      "narrative": "draft",
      "content": "Forest plot of 6 studies (Chipinduro 2017 to Orikiriza 2021) giving TP, FP, FN, TN, sensitivity and specificity (95% CI), with black squares and CI lines on 0–1 scales and a summary diamond. Sensitivity ranges from 0.06 (Hanrahan 2018) to 0.88 (Chipinduro 2017) and specificity from 0.25 to 0.93; the summary estimate is sensitivity 0.30 (0.13–0.55) and specificity 0.83 (0.70–0.90).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-301",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 301",
      "pageStart": 301,
      "pageEnd": 301,
      "target": "../sections/page-301.md",
      "narrative": null,
      "content": "296 \n \nFigure 18: Forest plots the parallel use of LC-aNAAT on respiratory sample and LC-aNAAT on stool. Sensitivity and specificity for \ntuberculosis in children without HIV or HIV status unknown using a MRS. The squares represent the sensitivity and specificity of \none study, the black line its confidence interval (CI). TP = true positive; FP = false positive; FN = false negative; TN = true negative \n \n \n \nDiagnostic accuracy of the component tests: LC-aNAATs combined or LF-LAM \nbased on the composite reference standard  \nThe sensitivity of LC-aNAATs combined alone ranged from 0.0% [Hanrahan",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Caption for Figure 18 and pooled accuracy of combined NAATs and LF-LAM alone against a composite reference standard, with the incremental accuracy of adding LF-LAM.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-302",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 302",
      "pageStart": 302,
      "pageEnd": 302,
      "target": "../sections/page-302.md",
      "narrative": null,
      "content": "297 \n \nparticipants, MRS (PICO 1) \nSee Appendix 23:  Summary of findings of incremental accuracy for parallel use of LC-aNAATs on \nrespiratory samples and stool and urinary LF-LAM versus LC-aNAATs combined alone for all \nparticipants, CRS (PICO 1)",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A short fragment pointing to the summary-of-findings appendices for PICO 2.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p303-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 303,
      "pageEnd": 303,
      "target": "../images/img-p303-1.webp",
      "narrative": "draft",
      "content": "Forest plots of study-specific differences in sensitivity and specificity (parallel LC-aNAAT test plus LF-LAM versus parallel LC-aNAAT test) for 6 studies, Chipinduro 2017 to Orikiriza 2021, with total N and risk-of-bias ratings (P, I, R, FT) shown as coloured dots on the right. Sensitivity differences range from 0 to 0.333 and specificity differences from −0.75 (Hanrahan 2018) to −0.05; the summary difference is 0.149 (95% CrI −0.0594 to 0.411) for sensitivity and −0.120 (−0.257 to −0.0258) for specificity.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-303",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 303",
      "pageStart": 303,
      "pageEnd": 303,
      "target": "../sections/page-303.md",
      "narrative": null,
      "content": "298 \n \nFigure 19: Forest plots of study-specific and summary estimates of difference in accuracy between the parallel use of LC-aNAATs combined and LF-LAM for tuberculosis in children \nwith HIV with presumed pulmonary tuberculosis by a CRS. The black squares represent the point estimate of one study, the black horizontal line its 95% confidence interval (CI). \nThe QUADAS-C risk-of-bias judgements by study are included on the right. Note that study-specific 95% CIs sometimes contain impossible values (difference in sensitivity < 0 and \ndifference in specificity > 0). This is because the method",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Caption for Figure 19, forest plots of the accuracy difference for combined NAATs plus LF-LAM against a composite reference standard in children with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-304",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 304",
      "pageStart": 304,
      "pageEnd": 304,
      "target": "../sections/page-304.md",
      "narrative": null,
      "content": "Incremental diagnostic accuracy of LC-aNAATs combined and LF-LAM versus \nLF-LAM alone based on the composite reference standard \nThe pooled difference in accuracy of the parallel use of LC-aNAATs combined and LF-LAM versus LF-\nLAM alone was 10.1% (95% CrI 5.1 to 17.2; PI 20.3 to 79.1) for sensitivity and -3.5% (95% CrI -21.6 to \n-4.9; PI -16.2 to -3.4) for specificity.  \n \n4.5.3 Subgroup analyses using a microbiological reference standard \nThe comparative diagnostic accuracy for LC-aNAATs combined and the \nparallel use of the two tests is shown for the following subgroups against a \nMRS. We al",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports the incremental accuracy of parallel testing over LF-LAM alone against a composite reference standard, and lists the PICO 2 subgroup analyses, including advanced HIV disease and low CD4 count.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-305",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 305",
      "pageStart": 305,
      "pageEnd": 305,
      "target": "../sections/page-305.md",
      "narrative": null,
      "content": "Table 4: Summary accuracy for subgroups of interest. Summary accuracy by test with the difference between the parallel \nuse (LC-aNAAT and LF-LAM) versus LC-aNAAT alone for detection of tuberculosis against a MRS \n \nStudies   \n(participants)\nSummary sensitivity \n95% CrI \nSummary specificity \n95% CrI \nPrimary analysis (MRS), all participants  \nLC-aNAATs combined \n6 (653) \n69.3% \n49.9, 84.0 \n95.4% \n91.1, 98.0 \nLF-LAM \n27.4% \n13.5, 47.5 \n88.2% \n78.2, 93.8 \nParallel  (LC-aNAATs combined + LF-LAM)  \n77.8 \n59.9, 89.8 \n83.9% \n73.9, 90.4 \nΔ between parallel and LC-aNAATs combined \n6.9 % \n1.5, 20.1 \n‐3.",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents Table 4, summarising sensitivity and specificity for combined NAATs, LF-LAM and parallel testing in all participants and in subgroups such as advanced HIV disease. It also reports sensitivity analyses that change the estimated gains and losses.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-306",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 306",
      "pageStart": 306,
      "pageEnd": 306,
      "target": "../sections/page-306.md",
      "narrative": null,
      "content": "(95% CrI −6.9 to −1.3).   \nWhen excluding participants with at least one unavailable or non-determinate index test result (6 \nstudies, 548 participants), pooled difference in sensitivity comparing the accuracy of parallel testing \nto LC-aNAAT combined was 8.4% (95% CrI 1.9 to 24.4) while pooled difference in specificity was -\n11.6% (95% CrI -21.2 to -5.9), compared to the primary analysis with a pooled difference in \nsensitivity of 6.9% (95% CrI 1.5 to 20.1) and specificity of −3.4% (95% CrI −6.9 to −1.3). See Appendix \n26.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that excluding participants with missing or invalid results gave a slightly larger sensitivity gain and a larger specificity loss for parallel testing than the primary analysis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-307",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 307",
      "pageStart": 307,
      "pageEnd": 307,
      "target": "../sections/page-307.md",
      "narrative": null,
      "content": "5. DISCUSSION  \nThis systematic review summarizes the evidence on the parallel use of LC-aNAAT on respiratory \nsamples and LC-aNAAT on stool in children without HIV with presumptive pulmonary tuberculosis \ndisease, compared to each test alone, based on 12 fully paired comparative diagnostic accuracy \nstudies. Furthermore, it summarises the evidence on the parallel use of LC-aNAAT on respiratory \nsamples plus LC-aNAAT on stool and LF-LAM on urine in children with HIV with presumptive \npulmonary tuberculosis disease, compared to each test alone, based on six fully paired comparative \ndiagnostic",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Opens the discussion by summarising the evidence base of 12 and six fully paired studies for the two review questions. It then begins setting out the main PICO 1 results as expected outcomes per 1000 children.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-308",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 308",
      "pageStart": 308,
      "pageEnd": 308,
      "target": "../sections/page-308.md",
      "narrative": null,
      "content": "• \nPooled sensitivity: 70.2% (51.1 to 84.7) | Pooled specificity: 95.4% (91.7 to 97.8) \n➔ For a population of 1,000 children where 400 have tuberculosis on culture, 308 would test \npositive using the parallel testing strategy; of these, 281 would have pulmonary tuberculosis \n(true positives) and 11 would not have pulmonary tuberculosis (false positives); 692 would test \nnegative by the parallel testing strategy; of these 573 would not have pulmonary tuberculosis \n(true negatives) and 119 would have pulmonary tuberculosis (false-negatives). \nCompared to LC-aNAAT on respiratory samples alone, di",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Translates pooled accuracy of parallel respiratory and stool NAAT testing into numbers of true and false positives and negatives per 1000 children. It also shows how many extra children would be correctly or wrongly identified compared with each test alone.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-309",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 309",
      "pageStart": 309,
      "pageEnd": 309,
      "target": "../sections/page-309.md",
      "narrative": null,
      "content": "test negative; of these, 3 would not have pulmonary tuberculosis (true negatives) and 14 \nwould have pulmonary tuberculosis (false-negatives). \nCompared to LC-aNAAT on stool alone, difference in accuracy is: \n• \nPooled difference in sensitivity: 10.5% (6.9 to 15.0)  \n• \nPooled difference in specificity: -0.1% (-0.7 to 0) \n➔ For a population of 1,000 children where 100 have tuberculosis on culture,  \ncompared to testing with LC-aNAAT on stool alone, 11 children would additionally test \npositive with the parallel strategy; of these, 10 would have pulmonary tuberculosis (true \npositives) and 1 wo",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the per-1000-children estimates for parallel testing compared with stool testing alone and for children with HIV, then opens the discussion of PICO 1.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-310",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 310",
      "pageStart": 310,
      "pageEnd": 310,
      "target": "../sections/page-310.md",
      "narrative": null,
      "content": "testing had a sensitivity of 79.9% and specificity of 93.4% for detection of pulmonary tuberculosis \nbased on the MRS. Sensitivity estimates ranged from 60-100% in the includes studies and may \nreflect that three of the eight included studies had less than 10 children each who were culture-\npositive, mostly due to low tuberculosis prevalence. However, the estimated difference in sensitivity \nbetween parallel and individual testing was fairly similar across studies. Compared to LC-aNAAT on \nrespiratory testing, parallel testing increased sensitivity by an absolute value of 7.1% at a loss of \nsp",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Discusses that parallel testing in children without HIV raised sensitivity modestly over respiratory testing and more over stool testing, with small specificity losses, suggesting the two sample types complement each other. It notes that the sensitivity analyses gave similar overall estimates.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-311",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 311",
      "pageStart": 311,
      "pageEnd": 311,
      "target": "../sections/page-311.md",
      "narrative": null,
      "content": "5.3 Summary of main results of PICO 2 \n___________________________________________________________________________ \nSee Appendix 22 and 23. Summary table, comparative accuracy for parallel use of LC-aNAATs \ncombined and LF-LAM versus LC-aNAATs combined alone using a MRS, presented for all \nparticipants, by subgroups, and sensitivity analyses. \n___________________________________________________________________________ \nBased on six studies included in this meta-analysis for a microbiological reference standard (653 \nparticipant, tuberculosis prevalence 6.6% [43/653]), the estimated diagnostic",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarises the PICO 2 results, giving pooled accuracy of parallel NAAT and LF-LAM testing in children with HIV and translating it into per-1000 estimates. These show that adding LF-LAM detects more cases but produces many more false positives.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-312",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 312",
      "pageStart": 312,
      "pageEnd": 312,
      "target": "../sections/page-312.md",
      "narrative": null,
      "content": "46 would not have pulmonary tuberculosis (true negatives) and 61 would have pulmonary \ntuberculosis (false-negatives). \n5.4 Parallel use of LC-aNAAT on respiratory samples plus stool and LF-LAM \nfor diagnosis of tuberculosis compared to each assay alone (PICO 2) \nFor the parallel use of LC-aNAATS combined and LF-LAM in CHIV, the overall diagnostic accuracy of \nparallel testing was estimated at a sensitivity of 77.6% and a specificity of 83.9% with MRS. \nSensitivity estimates ranged from 64% to 89% between studies, and similar to the result in PICO 1, \nmay reflect that five of the six included",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports diagnostic accuracy results from a systematic review on tuberculosis testing in children with HIV, finishing the single-test example and then presenting parallel use of rapid molecular tests on respiratory and stool samples together with urine LF-LAM. Parallel testing showed a sensitivity of about 78% and specificity of 84%, with wide variation between studies, and sensitivity rose compared with molecular tests alone. Subgroup comparisons by age and CD4 count are also discussed, with overlapping credible intervals.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-313",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 313",
      "pageStart": 313,
      "pageEnd": 313,
      "target": "../sections/page-313.md",
      "narrative": null,
      "content": "older than five years with a CD4-count lower than 200 per microliter. \n5.5 Strengths and weaknesses of the review  \nThis review represents the first estimations of incremental accuracy using parallel- compared to \nsingle-test strategies in children, based on comprehensive searching, strict inclusion, and \nstandardized data extraction. For both review questions, sample sizes were large and comparative \naccuracy study analyses were based on a number of studies. We communicated extensively with the \nauthors of the primary studies and were able to retrieve comprehensive data.  \nHowever, the result",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Discusses the strengths of the review, such as comprehensive searching and data from study authors, and its weaknesses, including unpublished re-analysed data and that fewer than half of eligible studies contributed. It compares the findings with earlier reviews in adults and with Cochrane estimates for single tests.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-314",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 314",
      "pageStart": 314,
      "pageEnd": 314,
      "target": "../sections/page-314.md",
      "narrative": null,
      "content": "Kay 2022]. For LF-LAM, estimated sensitivity in this review was lower compared to a previous \nsystematic review and meta-analysis (27.4% versus 46.59%) but with a higher specificity (88.2% \nversus 76.45%); however, methodologies including reference standard definitions may have varied \n[Seid 2022].",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Compares the review's LF-LAM accuracy estimates with an earlier meta-analysis, noting lower sensitivity but higher specificity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-315",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 315",
      "pageStart": 315,
      "pageEnd": 315,
      "target": "../sections/page-315.md",
      "narrative": null,
      "content": "6. Author’s conclusions \n6.1 Implications for practice  \nAlmost half of the estimated tuberculosis cases in children and two-thirds among children less than \n5 years old are not reported to public health programs worldwide [World Health Organization 2023]. \nUnderdiagnosis is the main contributor to this case detection gap, and leads to delays in treatment \ninitiation [World Health Organization 2023]. As most children who die from tuberculosis never \nreceived treatment, there is a global need to increase detection of tuberculosis disease in children. \nConsequently, approaches and tools that inc",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Sets out the authors' conclusions on testing children for tuberculosis, stressing the large gap in case detection and the priority given to sensitivity over specificity. It discusses the benefits of microbiological confirmation and of parallel testing on several sample types, and identifies research needs including extrapulmonary disease and drug-resistance detection.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-316",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 316",
      "pageStart": 316,
      "pageEnd": 316,
      "target": "../sections/page-316.md",
      "narrative": null,
      "content": "limited to use of Xpert MTB/RIF and Xpert Ultra, and further research is needed on other LC-aNAAT \nplatforms for children. There is an urgent need for implementation research to understand the \nbarriers and facilitators to guide the best strategies for tuberculosis testing in children, and there is a \nlack of data on the impact of parallel testing on patient outcomes for children with presumptive \ntuberculosis.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Concludes the research implications, calling for studies on other molecular platforms, implementation research and data on patient outcomes of parallel testing in children.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-317",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 317",
      "pageStart": 317,
      "pageEnd": 317,
      "target": "../sections/page-317.md",
      "narrative": null,
      "content": "Acknowledgements  \nWe are grateful to Vittoria Lutje, the CIDG Information Specialist, for help with the search strategy. \nWe thank all authors of primary studies for answering our questions and providing data.  \nContributions of authors  \nAll authors contributed significantly to this review. LO, SB, BY, JÅ, AK, AL, MS, and DJ wrote the \nprotocol. LO and DJ performed the majority of the screening and data extraction with assistance \nfrom BS, HP, and AR; AK gave input in case of discordances during screening and aligning this review \nwith other systematic reviews. The method was developed mainl",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Acknowledges help with the search strategy, describes each author's contribution to the review and lists declarations of interest.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-318",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 318",
      "pageStart": 318,
      "pageEnd": 318,
      "target": "../sections/page-318.md",
      "narrative": null,
      "content": "References \nAbbott. \"Determine TB LAM Package Insert.\" \nAinan, S., F. F. Furia, F. Mhimbira, N. P. Mnyambwa, N. Mgina, et al. 2021. \"Xpert R MTB/RIF assay \ntesting on stool for the diagnosis of paediatric pulmonary TB in Tanzania.\" Public Health in \nAction 11 (2): 75-79. \nBabo, Y., B. Seremolo, M. Bogale, A. Bedru, Y. Wabe, et al. 2023. \"Comparison of Xpert MTB/RIF \nUltra Results of Stool and Sputum in Children with Presumptive Tuberculosis in Southern \nEthiopia.\" Tropical Medicine & Infectious Disease 8 (7): 30. \nBanada, P. P., U. Naidoo, S. Deshpande, F. Karim, J. L. Flynn, et al. 2016. \"A N",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The start of the reference list for the review of parallel TB testing in children, citing studies of stool and urine testing and methodological sources.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-319",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 319",
      "pageStart": 319,
      "pageEnd": 319,
      "target": "../sections/page-319.md",
      "narrative": null,
      "content": "de Haas, P., B. Yenew, E. Mengesha, A. Slyzkyi, Z. Gashu, et al. 2021. \"The Simple One-Step (SOS) \nStool Processing Method for Use with the Xpert MTB/RIF Assay for a Child-Friendly Diagnosis \nof Tuberculosis Closer to the Point of Care.\" J Clin Microbiol 59 (8): e0040621. \nDiNardo, A. R., A. Detjen, P. Ustero, K. Ngo, J. Bacha, et al. 2016. \"Culture is an imperfect and \nheterogeneous reference standard in pediatric tuberculosis.\" Tuberculosis (Edinb) 101s: \nS105-s108. \nDonkeng, V. 2022. \"Performance Evaluation of TB-LAMP in children. Centre Pasteur du Cameroun: \nYaoundé, Cameroun, 2022.\". \nDu",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continuation of the reference list, citing studies on stool processing methods, clinical scoring and Xpert assays for tuberculosis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-320",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 320",
      "pageStart": 320,
      "pageEnd": 320,
      "target": "../sections/page-320.md",
      "narrative": null,
      "content": "Kunkel, Amber, Pia Abel Zur Wiesch, Ruvandhi R Nathavitharana, Florian M Marx, Helen E Jenkins, et \nal. 2016. \"Smear positivity in paediatric and adult tuberculosis: systematic review and meta-\nanalysis.\" BMC infectious diseases 16 (1): 1-9. \nLaCourse, S. M., P. B. Pavlinac, L. M. Cranmer, I. N. Njuguna, C. Mugo, et al. 2018. \"Stool Xpert \nMTB/RIF and urine lipoarabinomannan for the diagnosis of tuberculosis in hospitalized HIV-\ninfected children.\" AIDS 32 (1): 69-78. \nLawn, S. D., A. D. Kerkhoff, R. Burton, C. Schutz, A. Boulle, et al. 2017. \"Diagnostic accuracy, \nincremental yield and progno",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continuation of the reference list, citing studies on smear positivity and urine LAM testing in children and adults with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-321",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 321",
      "pageStart": 321,
      "pageEnd": 321,
      "target": "../sections/page-321.md",
      "narrative": null,
      "content": "Olbrich, L. 2024. \"Rapid and Accurate Diagnosis of Pediatric Tuberculosis Disease (RaPaed-TB). \nNCT03734172.\" \nOrikiriza, P., J. Smith, B. Ssekyanzi, D. Nyehangane, I. Mugisha Taremwa, et al. 2021. \"Tuberculosis \ndiagnostic accuracy of stool Xpert MTB/RIF and urine AlereLAM in vulnerable children.\" \nEuropean Respiratory Journal 59 (1): 01. \nPenn-Nicholson, A., S. N. Gomathi, C. Ugarte-Gil, A. Meaza, E. Lavu, et al. 2021. \"A prospective \nmulticentre diagnostic accuracy study for the Truenat tuberculosis assays.\" Eur Respir J 58 \n(5). \nPeter, Jonny G, Lynn S Zijenah, Duncan Chanda, Petra Clowes,",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continuation of the reference list, citing paediatric TB diagnostic trials, Truenat assays and Cochrane methodological guidance.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-322",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 322",
      "pageStart": 322,
      "pageEnd": 322,
      "target": "../sections/page-322.md",
      "narrative": null,
      "content": "Thit, Swe Swe, Ne Myo Aung, Zaw Win Htet, Mark A. Boyd, Htin Aung Saw, et al. 2017. \"The clinical \nutility of the urine-based lateral flow lipoarabinomannan assay in HIV-infected adults in \nMyanmar: an observational study.\" BMC Medicine 15 (1): 145. \nThomas, T. A. 2019. \"Tuberculosis in Children.\" Thorac Surg Clin 29 (1): 109-121. \nCovidence systematic review software., , Melbourne, Australia. Available at www.covidence.org. \nWalters, E., L. Scott, P. Nabeta, A. M. Demers, G. Reubenson, et al. 2018. \"Molecular Detection of \nMycobacterium tuberculosis from Stools in Young Children by Use of a N",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continuation of the reference list, citing studies on LF-LAM, stool molecular detection and tools for assessing diagnostic accuracy studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-323",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 323",
      "pageStart": 323,
      "pageEnd": 323,
      "target": "../sections/page-323.md",
      "narrative": null,
      "content": "Zar, Heather J, Lesley Workman, Washiefa Isaacs, Jacinta Munro, Faye Black, et al. 2012. \"Rapid \nmolecular diagnosis of pulmonary tuberculosis in children using nasopharyngeal specimens.\" \nClinical infectious diseases 55 (8): 1088-1095. \nZar, Heather J, Lesley J Workman, Margaretha Prins, Linda J Bateman, Slindile P Mbhele, et al. 2019. \n\"Tuberculosis diagnosis in children using Xpert Ultra on different respiratory specimens.\" \nAmerican journal of respiratory and critical care medicine 200 (12): 1531-1538. \nZifodya, J. S., J. S. Kreniske, I. Schiller, M. Kohli, N. Dendukuri, et al. 2021. \"Xper",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The final entries of the reference list, citing studies on molecular diagnosis of tuberculosis in children and Xpert Ultra in adults.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-324",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 324",
      "pageStart": 324,
      "pageEnd": 324,
      "target": "../sections/page-324.md",
      "narrative": null,
      "content": "Web Annex B.8. Parallel use of low-complexity automated \nnucleic acid amplification tests and lateral flow urine \nlipoarabinomannan assay to detect tuberculosis disease in adults \nand adolescents with HIV: a systematic review  \n \nStephanie Bjerrum*, Bada Yang, Johanna Åhsberg, Laura Olbrich, Mathias Weis Damkjær, Ruvandhi \nNathavitharana, Tobias Broger, Ioana Diana Olaru, Brittney Sweetser, Hayley Poore, Alia Razid, \nAlexander Kay, Claudia M. Denkinger, Ian Schiller, Nandini Dendukuri, Devan Jaganath, Andreas \nLundh, Maunank Shah  \n \n• \nStephanie Bjerrum, Department of Clinical Research, Resea",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title page and author list of Web Annex B.8, a systematic review of parallel use of rapid molecular tests and urine LF-LAM to detect tuberculosis in adults and adolescents with HIV, with author affiliations and corresponding author.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-325",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 325",
      "pageStart": 325,
      "pageEnd": 325,
      "target": "../sections/page-325.md",
      "narrative": null,
      "content": "EXCECUTIVE SUMMARY  \nBackground  \nLow complexity automated nucleic acid amplification tests (LC-aNAATs) are part of the molecular \nWorld Health Organization (WHO)-recommended rapid diagnostics widely used to diagnose \ntuberculosis disease. The lateral flow urine lipoarabinomannan (LF-LAM) test is recommended by \nWHO to assist in diagnosing tuberculosis disease among people with HIV. Previous systematic \nreviews have traditionally assessed diagnostic accuracy of LC-aNAATs and LF-LAM tests each in \nisolation for detection of tuberculosis, but in clinical practice the tests may be used in paralle",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The executive summary's background, objective and methods for the review of parallel molecular and LF-LAM testing in people with HIV. It reports results on patient outcomes, including an uncertain effect of parallel testing on mortality in inpatient settings.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-326",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 326",
      "pageStart": 326,
      "pageEnd": 326,
      "target": "../sections/page-326.md",
      "narrative": null,
      "content": "reporting higher overall mortality in the parallel testing group; this trial differed from the other two \nin having lower baseline antiretroviral therapy coverage, higher rates of comorbidities, and was \nconducted during the COVID-19 pandemic which may have affected overall care. The parallel testing \nstrategy likely increases the proportion with a microbiologically test positive tuberculosis diagnosis \n(risk ratio 3.06 (95% CI 1.82 to 5.16) and probably increases the proportion initiating tuberculosis \ntreatment (risk ratio 1.47 (95% CI 1.25 to 1.73), compared to routine tuberculosis diagnost",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the executive summary, noting that parallel testing likely increases microbiologically confirmed diagnoses and treatment starts and concluding that the two tests may be complementary. The page then begins the table of contents of the review.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-327",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 327",
      "pageStart": 327,
      "pageEnd": 327,
      "target": "../sections/page-327.md",
      "narrative": null,
      "content": "3.2.2 Searching other resources. ................................................................................................ 332 \n3.3 Data collection and analysis ....................................................................................................... 332 \n3.3.1 Selection of studies ............................................................................................................ 332 \n3.3.2 Data extraction and management ..................................................................................... 333 \n3.3.3 Assessment of methodological quality ...........",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Part of the table of contents of the review, listing methods, results, discussion and conclusion sections.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-328",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 328",
      "pageStart": 328,
      "pageEnd": 328,
      "target": "../sections/page-328.md",
      "narrative": null,
      "content": "Acknowledgements \n367 \nContributions of authors \n367 \nDeclarations of interest 367 \nReferences \n367",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The final lines of the table of contents, listing acknowledgements, author contributions, declarations of interest and references.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-329",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 329",
      "pageStart": 329,
      "pageEnd": 329,
      "target": "../sections/page-329.md",
      "narrative": null,
      "content": "1. BACKGROUND   \n1.1 Target condition being diagnosed  \nTuberculosis remains a challenge to diagnose particularly in people with HIV. Diagnosis is often \nmissed or delayed resulting in worse prognosis and patient outcomes. Globally, tuberculosis resulted \nin 1.1 million deaths among people without HIV and 167,000 deaths among people with HIV in 2022, \nand is the leading cause of death among people with HIV [Organization 2023].  \nTuberculosis predominantly affects the lungs as pulmonary tuberculosis, but can affect any part of \nthe body, such as the lymph nodes, pleura, urogenital system, brain",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes tuberculosis as the leading cause of death among people with HIV and explains the higher rates of extrapulmonary disease and diagnostic difficulty in this group. It introduces urine LF-LAM, noting that its sensitivity increases with greater immunosuppression.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-330",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 330",
      "pageStart": 330,
      "pageEnd": 330,
      "target": "../sections/page-330.md",
      "narrative": null,
      "content": "Previous reviews have traditionally assessed diagnostic accuracy of LC-aNAATs and LF-LAM each in \nisolation for detection of tuberculosis including Xpert MTB/RIF and Xpert Ultra assays (Cepheid, \nSunnyvale, USA) for pulmonary tuberculosis [Zifodya 2021; Shapiro 2021] and extrapulmonary \ntuberculosis [Kohli 2021a] and urine LF-LAM for people with HIV [Bjerrum 2019c]. Some studies have \nfound that urine LF-LAM and respiratory LC-aNAAT may be complementary for the diagnosis of \ntuberculosis disease in people with HIV [Shah 2014; Barr 2020; Lawn 2017b; Broger 2023a]. In the \n2016 Cochrane systemat",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarizes previous reviews that assessed molecular tests and LF-LAM separately, including evidence that combining them raised sensitivity, and gives earlier pooled LF-LAM accuracy estimates. It also begins to describe WHO's guidance on using LF-LAM by patient setting.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-331",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 331",
      "pageStart": 331,
      "pageEnd": 331,
      "target": "../sections/page-331.md",
      "narrative": null,
      "content": "i.e. inpatient and outpatient settings [World Health Organization. 2021a]. In inpatient settings, LF-\nLAM is recommended in people with HIV if: with signs and symptoms of tuberculosis; advanced HIV \ndisease; seriously ill; or with a CD4 cell count of less than or equal to 200 cells/mm3, irrespective of \nsigns and symptoms of tuberculosis. In outpatient settings, LF-LAM is suggested in people with HIV \nif: with signs and symptoms of tuberculosis; seriously ill; or with a CD4 cell count of less than or \nequal to 100 cells/mm3, irrespective of signs and symptoms of tuberculosis. \n1.2.2 Index test",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes WHO criteria for LF-LAM use in inpatient and outpatient settings and introduces the molecular index tests, Xpert MTB/RIF, Xpert Ultra and Truenat, and how they work.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-332",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 332",
      "pageStart": 332,
      "pageEnd": 332,
      "target": "../sections/page-332.md",
      "narrative": null,
      "content": "1.3 Role of the tests \nThe WHO operational handbook on tuberculosis states that “Urine LF-LAM should be used as an \nadd-on to clinical judgement in combination with other tests. It should not be used as a replacement \nor triage test” [World Health Organization. 2021c]. For LC-aNAAT, the suggested role is an initial test \nreplacing smear microscopy for the diagnosis of tuberculosis [World Health Organization. 2021a]. LF-\nLAM and LC-aNAAT may be performed in parallel (i.e. LF-LAM in combination with LC-aNAAT) where \na positive result on either test implies a positive result for the combination o",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains the intended role of LF-LAM as an add-on test and molecular tests as initial tests, and the rationale for estimating the change in accuracy when both are used in parallel to inform WHO guidelines.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-333",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 333",
      "pageStart": 333,
      "pageEnd": 333,
      "target": "../sections/page-333.md",
      "narrative": null,
      "content": "2. Review questions  \nWe addressed the following PICO (population, intervention, comparison, outcome) questions. See \nAppendix 1: PICO questions. \n1. What is the incremental diagnostic accuracy of parallel use of respiratory LC-aNAATs and urine \nLF-LAM for diagnosis of tuberculosis disease in adults and adolescents with HIV who present \nwith presumptive tuberculosis, compared to any of the tests alone?  \nIn addition, we will estimate incremental accuracy in the following subgroups: \n• in patients with signs or symptoms of tuberculosis versus asymptomatic  \n• in inpatient settings versus outpat",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists the review questions on the incremental accuracy of parallel testing in adults and adolescents with HIV, the subgroups of interest and the patient-important outcomes such as mortality and time to treatment.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-334",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 334",
      "pageStart": 334,
      "pageEnd": 334,
      "target": "../sections/page-334.md",
      "narrative": null,
      "content": "3. METHODS \nThe methodology of this review is similar to those used in previous reviews and protocols in the field \n[Shah 2016a; Bjerrum 2019c; Shapiro 2021; Kohli 2021a; Zifodya 2021; Inbaraj 2023a; Inbaraj \n2023b]. \n3.1 Criteria for considering studies for this review \n3.1.1 Types of studies \nWe included comparative diagnostic accuracy studies that evaluated the accuracy of a parallel \ncombination of respiratory LC-aNAATs and urine LF-LAM (henceforth referred to as ‘parallel testing’) \nand the diagnostic accuracy of one of the tests used alone, in the same study group (Table 8), \nagainst a m",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Begins the methods, setting out the types of comparative diagnostic accuracy studies eligible and excluded and the inclusion of adults and adolescents with HIV and presumptive tuberculosis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-008",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 8. Eligible comparisons",
      "pageStart": 334,
      "pageEnd": 334,
      "target": "../tables/table-8.csv",
      "narrative": "draft",
      "content": "Comparisons eligible for the review of parallel testing in adults and adolescents with HIV. For each comparison, the table gives index test A (respiratory LC-aNAAT plus urinary LF-LAM), the comparator index test B, and the test data required when individual participant data are or are not available.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-335",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 335",
      "pageStart": 335,
      "pageEnd": 335,
      "target": "../sections/page-335.md",
      "narrative": null,
      "content": "symptoms compatible with tuberculosis (e.g., a positive WHO four-symptom screen). We also \nincluded participants with a positive screening test for tuberculosis (e.g., abnormal chest X-ray \nfindings, elevated C-reactive protein or positive screening mWRD). We included participants from \nstudies that evaluated the index tests in people with HIV irrespective of tuberculosis symptoms, for \nexample, studies of people with HIV being assessed for initiation of antiretroviral treatment or \npeople with HIV admitted to medical wards. Although participants in these studies are not assessed \nfor signs an",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Further defines eligible participants, including those with positive screening tests or unselected people with HIV in high-prevalence settings, and describes the index tests and the target condition of any form of tuberculosis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-336",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 336",
      "pageStart": 336,
      "pageEnd": 336,
      "target": "../sections/page-336.md",
      "narrative": null,
      "content": "tuberculosis were not regarded as separate target conditions. \n3.1.5 Reference standards \nWe required studies to ascertain the presence or absence of tuberculosis disease using at least one \nof the following two reference standards:  \n3.1.5.1 Microbiological Reference Standard \nFor a microbiological reference standard  \n• Presence of tuberculosis disease was defined as a positive M. tuberculosis culture on \nany sample type or a positive NAAT result for detection of tuberculosis on any \nsample other than LC-aNAAT on respiratory samples. \n• Absence of tuberculosis disease was defined as negative",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Defines the microbiological and composite reference standards used to determine whether participants had tuberculosis disease.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-337",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 337",
      "pageStart": 337,
      "pageEnd": 337,
      "target": "../sections/page-337.md",
      "narrative": null,
      "content": "Some composite reference standards may include one or more index tests (LF-LAM, respiratory LC-\naNAATs, or both) as part of the reference standard; for example, if the decision to start tuberculosis \ntreatment directly depended on the results of respiratory LC-aNAAT. This will likely lead to \noverestimation of diagnostic specificity. If the reference standard incorporated any of the index \ntests, we considered the risk of bias to be high in the reference standard domain.    \nBecause of the uncertainties that surround a clinical diagnosis of tuberculosis, a composite reference \nstandard based o",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Discusses how composite reference standards that include index tests may bias accuracy estimates, then describes the search methods and the process for screening titles, abstracts and full texts.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-338",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 338",
      "pageStart": 338,
      "pageEnd": 338,
      "target": "../sections/page-338.md",
      "narrative": null,
      "content": "articles. We resolved disagreements through discussion and, if necessary, involved a third review \nauthor as arbiter. We listed reasons for exclusion. Review authors who co-authored any of the \nincluded studies did not participate in assessment or data extraction of their own study [Cochrane \nLibrary Content.]. We illustrated the study selection process in a PRISMA diagram [Page 2021].  \n3.3.2 Data extraction and management \nUsing a pre‐developed data extraction form, two review authors worked independently to extract \ndata on the following characteristics. See Appendix 3: Data extraction form",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes study selection, independent data extraction on study, participant, test and reference standard characteristics, quality assessment with QUADAS tools and the Bayesian statistical model for comparing accuracy.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-339",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 339",
      "pageStart": 339,
      "pageEnd": 339,
      "target": "../sections/page-339.md",
      "narrative": null,
      "content": "accomplished by assuming that difference in TP (TP of parallel testing minus TP of component test) \nand difference in TN (TN of parallel testing minus TN of component test) both follow a binomial \ndistribution.  This approach naturally accounts for the dependence between the two tests being \ncompared (i.e. the parallel testing strategy and the component test) and ensures that difference in \nsensitivity cannot be lower than zero and difference in specificity cannot be greater than zero, as \nwould be expected when using two tests in parallel. As the binomial distribution does not allow \nnegative",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains the statistical approach for modelling differences in sensitivity and specificity between parallel and single testing, the prior distributions used and the handling of non-determinate results. It begins listing subgroup analyses for heterogeneity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-340",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 340",
      "pageStart": 340,
      "pageEnd": 340,
      "target": "../sections/page-340.md",
      "narrative": null,
      "content": "• CD4 cell count (CD4 ≤ 200 versus CD4 > 200)  \n• Clinical setting (inpatient versus outpatient) \n• Type of LC-aNAAT. However, there was data only for Xpert MTB/RIF and Xpert \nMTB/RIF Ultra. This analysis was restricted to within-study comparisons \n3.3.7 Sensitivity analysis \nWe performed the following sensitivity analyses by limiting inclusion in the meta-analysis based on \nthe following criteria: \n• Only screening positive for tuberculosis (e.g. chest X-ray abnormal, CRP elevated or \nmWRD screening test positive) \n• Only seriously ill  \n• Only advanced HIV  \n• Only studies using Xpert Ultra",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists the remaining subgroup analyses and the sensitivity analyses, such as restricting to seriously ill patients, advanced HIV or Xpert Ultra. It then explains how data on health impact outcomes were identified from trials and related reviews.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-341",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 341",
      "pageStart": 341,
      "pageEnd": 341,
      "target": "../sections/page-341.md",
      "narrative": null,
      "content": "on Impact of LF-LAM on patient outcomes [Nathavitharana 2021]; all articles with data on impact \nwere further reviewed for inclusion by review author RRN.  \nWe included randomized controlled trials that assessed the effect of test strategies on the following \npatient important outcomes: cure, mortality, loss-to-follow-up, time to diagnosis, and time to \ntreatment initiation. Trials that included participants as described under section 3.1.2 and allocated \nparticipants to one of the test strategies described in Table 8, i.e. a test strategy of LC-aNAAT and LF-\nLAM used in parallel versus a test",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the methods for the part of the systematic review on how testing strategies affect patient outcomes, including only randomized trials that compared parallel respiratory LC-aNAAT and urine LF-LAM testing with either test alone. It lists outcomes such as cure, mortality and time to diagnosis or treatment, and notes that pooled risk ratios were estimated by random-effects meta-analysis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-342",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 342",
      "pageStart": 342,
      "pageEnd": 342,
      "target": "../sections/page-342.md",
      "narrative": null,
      "content": "4. RESULTS \n4.1 Results of the search \nWe identified and screened a total of 457 records for inclusion in this systematic review. Of these, \nwe assessed 130 full-text reports against our inclusion criteria. We excluded 99 records and provided \nreasons.  \nFigure 1 shows the flow of studies in the review. We recorded the excluded studies (n=99) and the \nreasons for their exclusion. We included 31 unique studies: all studies provided data comparing a \nparallel strategy of respiratory LC-aNAAT and urine LF-LAM to any of the tests alone for detection of \ntuberculosis. Of the 31 studies, we could ex",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports the search results of the systematic review: 457 records screened, 130 full texts assessed and 31 studies included, all comparing parallel respiratory LC-aNAAT and urine LF-LAM with either test alone. It explains that most data came from authors or from re-analysis of shared individual participant data.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p343-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 343,
      "pageEnd": 343,
      "target": "../images/img-p343-1.webp",
      "narrative": "draft",
      "content": "Flow diagram of studies in the review across identification, screening and inclusion stages. 449 studies from databases/registers (MEDLINE 385, ClinicalTrials.gov 64) and 8 references from other sources (WHO public call for data 1, citation searching 7); 15 duplicates removed. 442 studies screened (312 excluded), 130 sought for retrieval (0 not retrieved), 130 assessed for eligibility (99 excluded, with reasons listed, e.g. duplicate/subgroup analysis 23, no or insufficient accuracy data 19), and 31 studies included in the review.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-343",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 343",
      "pageStart": 343,
      "pageEnd": 343,
      "target": "../sections/page-343.md",
      "narrative": null,
      "content": "Figure 8. Flow of studies in the review.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A caption for the figure showing the flow of studies through the review.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p344-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 344,
      "pageEnd": 344,
      "target": "../images/img-p344-1.webp",
      "narrative": "draft",
      "content": "Risk of bias and applicability grid for 31 studies (Åhsberg 2023 to Yoon 2019). Columns cover risk of bias (patient selection, index test, micro reference standard, comp reference standard, flow and timing) and applicability (patient selection, index test, reference standard), coloured low (green), high (red), unclear (yellow) or not applicable (grey). Index test and applicability are mostly low, while high ratings are frequent for patient selection, the reference standards and flow and timing.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-344",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 344",
      "pageStart": 344,
      "pageEnd": 344,
      "target": "../sections/page-344.md",
      "narrative": null,
      "content": "4.2 Methodological quality of included studies  \n4.2.1 QUADAS-2  \nFigure 9 shows risk of bias and applicability concerns for each of the 31 studies included. \nIn the patient selection domain, we considered 11 studies (35%) to be at high risk of bias for the \nfollowing reasons: nine studies excluded participants who could not produce sputum samples \n(and/or urine)[Benjamin 2019; Bjerrum 2015; Broger-Cohort4 2020; Calligaro 2017; Kebede 2021; J. \nFigure 9: Risk of bias in and applicability concerns related to single test accuracy estimates (QUADAS-2).  \nReview authors' judgements about each doma",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Begins the assessment of study quality using QUADAS-2, noting that 11 of 31 studies were at high risk of bias in patient selection, mostly because they excluded people unable to give sputum or urine samples.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-345",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 345",
      "pageStart": 345,
      "pageEnd": 345,
      "target": "../sections/page-345.md",
      "narrative": null,
      "content": "Peter 2015a; J.G. Peter 2016; Shah 2014; Stead 2023]; one study enrolled participants who had \nstarted tuberculosis treatment before index testing [de Vasconcellos 2021]; and one study enrolled a \nsubset of participants with a known positive sputum Xpert ultra result for convenience that could \nhave inflated sensitivity [Shapiro 2022]. Regarding applicability, 26 studies (84%) had low concern \nbecause the studies included the appropriate participants and settings. We judged three studies \n(10%) to have high concern: one study because participants were evaluated exclusively as part of a \ncommun",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the QUADAS-2 assessment, describing other sources of bias such as enrolling people already on tuberculosis treatment or selecting known Xpert-positive cases. It reports that most studies raised low applicability concerns, explains the few with high concern, and notes that nine studies were at high risk of bias for flow and timing.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-346",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 346",
      "pageStart": 346,
      "pageEnd": 346,
      "target": "../sections/page-346.md",
      "narrative": null,
      "content": "Wright 2018; Huerga 2017a; Huerga 2021; Liu 2022; Spooner 2022; Stead 2023] or because not all \nparticipants were included in the two-by-two tables [de Vasconcellos 2021; Huerga 2019; Liu 2022]. \nThree studies had unclear risk of bias [Huerga 2023a; Huerga 2020; Van Hoving 2019]. We judged \nthe remaining 19 studies (61%) to be at low risk of bias noting that all participants were included in \nthe analysis.  \n4.2.2 QUADAS-C \nFigure 10 shows risk of bias for 31 studies comparing a parallel use of LC-aNAAT on respiratory \nsamples and LF-LAM on urine to LC-aNAAT on respiratory samples alone. All s",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the QUADAS-2 assessment and begins the QUADAS-C comparative risk-of-bias assessment of the 31 paired studies comparing parallel testing with LC-aNAAT alone. It notes issues such as non-determinate or invalid results, reported as generally low.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p347-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 347,
      "pageEnd": 347,
      "target": "../images/img-p347-1.webp",
      "narrative": "draft",
      "content": "Risk-of-bias grid for comparative accuracy in 31 studies (Åhsberg 2023 to Yoon 2019), with five domains: patient selection, index test, micro reference standard, comp reference standard, and flow and timing. Cells are low (green), high (red), unclear (yellow) or not applicable (grey); the composite reference standard is high or not applicable in almost every study, and high ratings are also common for patient selection and flow and timing.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-347",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 347",
      "pageStart": 347,
      "pageEnd": 347,
      "target": "../sections/page-347.md",
      "narrative": null,
      "content": "4.3 Findings \nFigure 10: Risk of bias in comparative accuracy estimates (QUADAS C) \nReview authors' judgements about each domain for each included study.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces the findings section with the caption of the figure on comparative risk of bias (QUADAS-C) for each study.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-348",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 348",
      "pageStart": 348,
      "pageEnd": 348,
      "target": "../sections/page-348.md",
      "narrative": null,
      "content": "The 31 studies were conducted in 19 different countries, four (21%) in low-income countries, 10 \n(53%) in lower middle-income countries and five (26%) in upper middle-income countries. Fourteen \ncountries (74%) were high burden countries for tuberculosis/HIV [World Health Organization. \n2021b]. A majority of studies, 27 (87%), enrolled participants from 12 countries within Sub-Saharan \nAfrica.  \nWe noted substantial differences between the studies regarding the following characteristics: target \ncondition (e.g. pulmonary tuberculosis, extra pulmonary tuberculosis, mycobacteraemia, or a \ncombin",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the 31 included studies, carried out in 19 countries, mostly in sub-Saharan Africa and in high TB/HIV burden settings. It notes major differences between studies in target condition, CD4 counts, tuberculosis prevalence and setting, from inpatient wards to outpatient clinics and community screening.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p349-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 349,
      "pageEnd": 349,
      "target": "../images/img-p349-2.webp",
      "narrative": "draft",
      "content": "Forest plot labelled 'Parallel' for 30 studies (Ahsberg 2023 to Yoon 2019), giving TP, FP, FN, TN and sensitivity and specificity (95% CI), plotted as blue squares with CI lines on 0–1 scales. Sensitivity ranges from 0.33 (Olbrich 2024) to 1.00 and specificity from 0.57 (Huerga 2019) to 1.00, with most sensitivities between about 0.6 and 0.9.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-349",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 349",
      "pageStart": 349,
      "pageEnd": 349,
      "target": "../sections/page-349.md",
      "narrative": null,
      "content": "screening setting, and one in primary health care facilities. For the remaining two studies, the setting \nwas unclear. The median CD4 cell count ranged from 42 to 317 cells per µL. See Appendix 9: \nCharacteristics of individual studies.  \nDiagnostic accuracy of parallel use of respiratory LC-aNAAT and urine LF-\nLAM based on the microbiological reference standard  \nThe sensitivity of the parallel use of LC-aNAAT on respiratory samples and LF-LAM on urine ranged \nfrom 33% to 100% and specificity from 57% to 100%. See Figure 11. The pooled sensitivity of parallel \ntesting was 77.5% (95% CrI 73.4",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives the pooled accuracy of parallel respiratory LC-aNAAT and urine LF-LAM testing against a microbiological reference standard: sensitivity about 78% and specificity about 89% across 27 studies, with a related forest plot.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-350",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 350",
      "pageStart": 350,
      "pageEnd": 350,
      "target": "../sections/page-350.md",
      "narrative": null,
      "content": "Diagnostic accuracy of the component tests, respiratory LC-aNAAT and urine \nLF-LAM, based on the microbiological reference standard  \nThe sensitivity of respiratory LC-aNAAT alone ranged from 24% to 100%, and specificity from 91% to \n100%. The pooled sensitivity LC-aNAAT alone was 68.0% (95% CrI 60.8 to 74.9) and specificity 96.7% \n(95% CrI 95.7 to 97.6), (27 studies; 12,651 participants, 2,368 (18.7%) with tuberculosis). \nThe sensitivity of urine LF-LAM alone ranged from 0% to 75%, and specificity ranged from 59% to \n100%. The pooled sensitivity of LF-LAM alone was 39.1% (95% CrI 32.6 to 45.9",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports pooled accuracy of the component tests on their own, with LC-aNAAT alone at about 68% sensitivity and 97% specificity and LF-LAM alone at about 39% sensitivity and 92% specificity. It also gives the pooled gain in sensitivity and loss in specificity from parallel testing compared with LC-aNAAT alone.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p351-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 351,
      "pageEnd": 351,
      "target": "../images/img-p351-1.webp",
      "narrative": "draft",
      "content": "Paired forest plot (Figure 12) of 27 studies, from Ahsberg 2023 to Yoon 2019, showing the difference in sensitivity and in specificity between parallel testing (LC-aNAAT plus LF-LAM) and LC-aNAAT alone, with total N and coloured risk-of-bias ratings (P, I, R, FT) for each study. The summary difference is 0.0946 (95% CrI 0.0147 to 0.176) for sensitivity, favouring the parallel test, and -0.0731 (95% CrI -0.1102 to -0.0426) for specificity, favouring the LC-aNAAT test.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-351",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 351",
      "pageStart": 351,
      "pageEnd": 351,
      "target": "../sections/page-351.md",
      "narrative": null,
      "content": "346 \nFigure 12. Forest plot of study-specific estimates of difference in sensitivity and specificity between the parallel use of LC-aNAAT and LF-LAM versus LC-aNAAT alone for tuberculosis in adults and adolescents \nwith HIV, based on the microbiological reference standard. The black squares represent the point estimate of one study, the black horizontal line its 95% confidence interval (CI). The QUADAS-C risk-of-bias \njudgements by study are included on the right. Note that study-specific 95% CIs sometimes contain impossible values (difference in sensitivity < 0 and difference in specificity >",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Caption for a forest plot of study-level differences in sensitivity and specificity between parallel testing and LC-aNAAT alone, noting why some confidence intervals include impossible values.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p352-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 352,
      "pageEnd": 352,
      "target": "../images/img-p352-1.webp",
      "narrative": "draft",
      "content": "Forest plot (Figure 13) of pooled sensitivity and specificity for LF-LAM, LC-aNAAT and their parallel use, against a microbiological reference standard (MRS, 12 651 participants, 18.7% with TB) and a composite reference standard (CRS, 11 109 participants, 33.5% with TB). Against MRS, sensitivity is 39.1% for LF-LAM, 68.0% for LC-aNAAT and 77.5% for parallel use (difference 6.7%), with specificity 91.9%, 96.7% and 89.4% (difference -6.8%). Against CRS, sensitivity is 38.6%, 46.8% and 67.6% (difference 16.0%) and specificity 96.3%, 99.9% and 96.2% (difference -3.5%).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-352",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 352",
      "pageStart": 352,
      "pageEnd": 352,
      "target": "../sections/page-352.md",
      "narrative": null,
      "content": "347 \nFigure 13. Forest plot of pooled difference in sensitivity and specificity (all studies combined) by index test: LF-LAM, LC-aNAAT, and their parallel use.  \nThe squares represent the pooled sensitivity and specificity, the black horizontal line its 95% credible interval (CrI). The pooled difference in sensitivity and specificity between parallel testing and LC-aNAAT alone \nis indicated by a line connecting two squares. This pooled difference may not correspond to the difference between the pooled single test accuracy estimates (see 3.3.4: Statistical analysis and data synthesis). \nMRS = m",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Caption for a forest plot of pooled sensitivity and specificity for LF-LAM, LC-aNAAT and parallel testing, with the pooled differences between them.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-353",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 353",
      "pageStart": 353,
      "pageEnd": 353,
      "target": "../sections/page-353.md",
      "narrative": null,
      "content": "Incremental accuracy of parallel testing versus LF-LAM alone based on the \nmicrobiological reference standard  \nThe study-specific estimates of difference in accuracy between the parallel use of respiratory LC-aNAAT \nand urine LF-LAM versus LF-LAM alone for tuberculosis in adults and adolescents with HIV, based on the \nmicrobiological reference standard, are shown in Appendix 15. The pooled difference in sensitivity and \npooled difference in specificity between parallel testing versus LF-LAM alone was 35.7% (95% CrI 29.3 to \n42.7) and -2.0% (95% CrI -2.7 to -1.5).  \n4.4.2 Comparative diagnosti",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that parallel testing raised sensitivity by about 36 percentage points over LF-LAM alone with a small loss of specificity, using the microbiological reference standard. It then presents the comparative accuracy based on a composite reference standard from 23 studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-354",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 354",
      "pageStart": 354,
      "pageEnd": 354,
      "target": "../sections/page-354.md",
      "narrative": null,
      "content": "4.4.3 Subgroup analyses based on the microbiological reference standard \nThe comparative diagnostic accuracy for the parallel use of respiratory LC-aNAAT and LF-LAM versus any \nof the tests alone is shown for the following subgroups: symptomatic versus asymptomatic, CD4 cell \ncount ≤ 200 versus > 200, and inpatients versus outpatients. We also show results limited to the \nfollowing subgroups: screening positive for tuberculosis, with advanced HIV, and seriously ill. See \nAppendix 11: Summary table, comparative accuracy for parallel use (LC-aNAAT and LF-LAM) versus LC-\naNAAT alone by subgroup.",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces subgroup analyses of parallel testing versus either test alone, by symptoms, CD4 count, inpatient or outpatient status, positive TB screening, advanced HIV and serious illness. It gives results for symptomatic and asymptomatic people and begins the results for people with CD4 counts of 200 or less.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-009",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 9. Summary accuracy by presence of symptoms. Pooled test accuracy along with summary difference between the parallel use (LC-aNAAT and LF-LAM) versus LC-aNAAT alone for detection of tuberculosis using a microbiological reference standard.",
      "pageStart": 354,
      "pageEnd": 354,
      "target": "../tables/table-9.csv",
      "narrative": "draft",
      "content": "Pooled accuracy of LF-LAM, LC-aNAAT and parallel testing (LC-aNAAT plus LF-LAM) for detecting tuberculosis against a microbiological reference standard, in people with HIV with and without symptoms. Rows give each test with the number of studies (participants) and pooled sensitivity and specificity (95% CrI), followed by the summary difference between parallel testing and LC-aNAAT.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-355",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 355",
      "pageStart": 355,
      "pageEnd": 355,
      "target": "../sections/page-355.md",
      "narrative": null,
      "content": "between parallel testing and LC-aNAAT was 9.2% (95% CrI 4.0 to 17.4; 95% PI 0.6 to 57.4) and -9.4% \n(95% CrI -14.7 to -5.6; 95% PI -40.0 to -1.6).  \nTwo studies reported data for participants with CD4 > 200, involving 148 participants, 11 (7%) with \ntuberculosis.  \nTable 10: Summary accuracy by CD4.  \nPooled test accuracy along with difference between the parallel use (LC-aNAAT and LF-LAM) versus LC-aNAAT alone for \ndetection of tuberculosis using a microbiological reference standard. \nParticipants \nStudies (participants) \nPooled sensitivity (95% CrI) \nPooled specificity (95% CrI) \nCD4 ≤ 200",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the subgroup results, giving the sensitivity gain and specificity loss of parallel testing in people with CD4 counts of 200 or less, and noting few data above 200. Tables summarize pooled accuracy by CD4 count and by inpatient or outpatient setting.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-010",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 10: Summary accuracy by CD4. Pooled test accuracy along with difference between the parallel use (LC-aNAAT and LF-LAM) versus LC-aNAAT alone for detection of tuberculosis using a microbiological reference standard.",
      "pageStart": 355,
      "pageEnd": 355,
      "target": "../tables/table-10.csv",
      "narrative": "draft",
      "content": "Pooled accuracy of LF-LAM, LC-aNAAT and parallel testing (LC-aNAAT plus LF-LAM) for detecting tuberculosis against a microbiological reference standard, in people with HIV with CD4 counts of 200 or below and above 200. Rows give each test with the number of studies (participants) and pooled sensitivity and specificity (95% CrI), followed by the summary difference between parallel testing and LC-aNAAT.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/table-011",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 11: Summary accuracy by clinical setting. Pooled test accuracy along with difference between the parallel use (LC-aNAAT and LF-LAM) versus LC-aNAAT alone for detection of tuberculosis using a microbiological reference standard.",
      "pageStart": 355,
      "pageEnd": 355,
      "target": "../tables/table-11.csv",
      "narrative": "draft",
      "content": "Pooled accuracy of LF-LAM, LC-aNAAT and parallel testing (LC-aNAAT plus LF-LAM) for detecting tuberculosis against a microbiological reference standard, in inpatients and outpatients with HIV. Rows give each test with the number of studies (participants) and pooled sensitivity and specificity (95% CrI), followed by the summary difference between parallel testing and LC-aNAAT.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-356",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 356",
      "pageStart": 356,
      "pageEnd": 356,
      "target": "../sections/page-356.md",
      "narrative": null,
      "content": "There was evidence that the difference in sensitivity among inpatients exceeds that of outpatients \n(probability that between-subgroup difference > 0 was 98%). Regarding difference in specificity, there \nwas also evidence of between-subgroup difference (probability that between-subgroup difference < 0 \nwas 97%).  \nComparative accuracy in participants screening positive for tuberculosis  \nSix studies reported data for participants screening positive for tuberculosis involving 1,901 participants, \n346 with tuberculosis (18%). Participants were reported screening positive by chest Xray by [Åhsber",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports evidence that the benefit in sensitivity and the loss in specificity from parallel testing are larger for inpatients than outpatients. It then gives results for people screening positive for tuberculosis by chest X-ray or CRP and for seriously ill participants.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-012",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 12. Summary accuracy for screening positive. Pooled test accuracy along with difference between the parallel use (LC-aNAAT and LF-LAM) versus LC-aNAAT alone for detection of tuberculosis using a microbiological reference standard.",
      "pageStart": 356,
      "pageEnd": 356,
      "target": "../tables/table-12.csv",
      "narrative": "draft",
      "content": "Pooled accuracy of LF-LAM, LC-aNAAT and parallel testing (LC-aNAAT plus LF-LAM) for detecting tuberculosis against a microbiological reference standard, in people with HIV who screened positive for tuberculosis. Rows give each test with the number of studies (participants) and pooled sensitivity and specificity (95% CrI), followed by the summary difference between parallel testing and LC-aNAAT.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/table-013",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 13: Summary accuracy for advanced HIV. Summary accuracy by test with the difference between the parallel use (LC-aNAAT and LF-LAM) versus LC-aNAAT alone for detection of tuberculosis against a microbiological reference standard",
      "pageStart": 356,
      "pageEnd": 356,
      "target": "../tables/table-13.csv",
      "narrative": "draft",
      "content": "Pooled accuracy of LF-LAM, LC-aNAAT and parallel testing (LC-aNAAT plus LF-LAM) for detecting tuberculosis against a microbiological reference standard, in people with advanced HIV disease. Rows give each test with the number of studies (participants) and pooled sensitivity and specificity (95% CrI), followed by the summary difference between parallel testing and LC-aNAAT.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/table-014",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 14: Summary accuracy for seriously ill. Summary accuracy by test with the difference between the parallel use (LC-aNAAT and LF-LAM) versus LC-aNAAT alone for detection of tuberculosis against a microbiological reference standard",
      "pageStart": 356,
      "pageEnd": 357,
      "target": "../tables/table-14.csv",
      "narrative": "draft",
      "content": "Pooled accuracy of LF-LAM, LC-aNAAT and parallel testing (LC-aNAAT plus LF-LAM) for detecting tuberculosis against a microbiological reference standard, in seriously ill people with HIV. Rows give each test with the number of studies (participants) and pooled sensitivity and specificity (95% CrI), followed by the summary difference between parallel testing and LC-aNAAT.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p357-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 357,
      "pageEnd": 357,
      "target": "../images/img-p357-1.webp",
      "narrative": "draft",
      "content": "Table rendered as an image (Table 16) comparing, within the studies Gils 2023 and Reeve 2023, the absolute difference in sensitivity and specificity (point estimate and 95% CI) of parallel testing versus Xpert Ultra and versus Xpert MTB/RIF. P values for interaction are 0.944 and 0.719 for Gils 2023 and 0.919 and 0.411 for Reeve 2023, each noted as no evidence of a difference between Xpert Ultra and Xpert MTB/RIF. Both column groups are headed 'Absolute difference in sensitivity'.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-357",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 357",
      "pageStart": 357,
      "pageEnd": 357,
      "target": "../sections/page-357.md",
      "narrative": null,
      "content": "Summary accuracy by test with the difference between the parallel use (LC-aNAAT and LF-LAM) versus LC-aNAAT alone for \ndetection of tuberculosis against a microbiological reference standard \nParticipants \nStudies (participants) \nPooled sensitivity (95% CrI) \nPooled specificity (95% CrI) \nSeriously ill \n \n \n \n    LF-LAM \n6 (577) \n47.8% (30.6 to 67.0) \n82.4% (71.4 to 90.6) \n    LC-aNAAT \n6 (577) \n67.9% (47.8 to 85.0) \n94.0% (89.4 to 97.1) \n    Parallel \n6 (577) \n78.7% (62.8 to 90.7) \n79.7% (68.7 to 88.5) \nSummary difference between parallel and LC-aNAAT \n13.3% (4.4 to 25.4) \n-14.4% (-24.2 to -7.",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Tabulates accuracy among seriously ill participants and reports additional analyses, including studies that used Xpert Ultra and the proportion of unavailable or non-determinate test results.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-015",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 15. Summary accuracy for Xpert Ultra only. Summary accuracy by test with the difference between the parallel use (Xpert Ultra and LF-LAM) versus Xpert Ultra alone for detection of tuberculosis against a microbiological reference standard",
      "pageStart": 357,
      "pageEnd": 357,
      "target": "../tables/table-15.csv",
      "narrative": "draft",
      "content": "Pooled accuracy of LF-LAM, LC-aNAAT and parallel testing (LC-aNAAT plus LF-LAM) for detecting tuberculosis against a microbiological reference standard, restricted to studies using Xpert Ultra as the LC-aNAAT. Rows give each test with the number of studies (participants) and pooled sensitivity and specificity (95% CrI), followed by the summary difference between parallel testing and LC-aNAAT.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/table-016",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 16. Within study comparison of Xpert vs Ultra",
      "pageStart": 357,
      "pageEnd": 357,
      "target": "../tables/table-16.csv",
      "narrative": "draft",
      "content": "Within-study comparisons from two studies (Gils 2023 and Reeve 2023) of the absolute difference in accuracy between parallel testing with LF-LAM and the LC-aNAAT alone, separately for Xpert Ultra and Xpert MTB/RIF. Columns give the point estimate and lower and upper 95% CI for each of two difference measures, with a P value for interaction between the two Xpert assays.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-358",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 358",
      "pageStart": 358,
      "pageEnd": 358,
      "target": "../sections/page-358.md",
      "narrative": null,
      "content": "Peter 2015a; Van Hoving 2019] \n4.6 Sensitivity analyses \nWhen excluding studies at high of bias (QUADAS-C) and with high or unclear concerns regarding \napplicability (QUADAS-2), the pooled difference in sensitivity increased to 10.0% (95% CrI 4.9 to 17.6) \nand pooled difference in specificity to -10.8 (95% CrI -20.6 to -5.4) (8 studies, 4,304 participants). See \nTable 17.  \nWhen excluding participants with at least one unavailable or non-determinate index test result (1.966 \nparticipants), pooled difference in sensitivity decreased to 4.4% (95% CrI 2.8 to 6.5) while pooled \ndifference in speci",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports sensitivity analyses: excluding high-risk studies raised the sensitivity gain of parallel testing to about 10 points, while excluding participants with missing or non-determinate results lowered it to about 4 points. A table summarizes the results.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-017",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 17. Summary accuracy for sensitivity analysis. Summary accuracy by test with the difference between the parallel use (Xpert Ultra and LF-LAM) versus Xpert Ultra alone for detection of tuberculosis against a microbiological reference standard",
      "pageStart": 358,
      "pageEnd": 358,
      "target": "../tables/table-17.csv",
      "narrative": "draft",
      "content": "Pooled accuracy of LF-LAM, LC-aNAAT and parallel testing (LC-aNAAT plus LF-LAM) for detecting tuberculosis against a microbiological reference standard, in sensitivity analyses (excluding studies at high risk of bias, excluding participants with an unavailable or non-determinate index test result, and using alternative priors). Rows give each test with the number of studies (participants) and pooled sensitivity and specificity (95% CrI), followed by the summary difference between parallel testing and LC-aNAAT.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-359",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 359",
      "pageStart": 359,
      "pageEnd": 359,
      "target": "../sections/page-359.md",
      "narrative": null,
      "content": "4.7 PICO 2 on impact \nPICO 2: What is the effect of parallel use of respiratory LC-aNAATs and urine LF-\nLAM on patient important outcomes in adults and adolescents with HIV who \npresent with presumptive tuberculosis, compared to respiratory LC-aNAAT alone? \nWe identified three trials (n=4602 participants) with data on the impact of parallel use of respiratory LC-\naNAATs and urine LF-LAM for clinical decision-making on mortality, time to tuberculosis diagnosis and \ntime to tuberculosis treatment [Åhsberg 2023; Gupta-Wright 2018; J.G. Peter 2016](Table 18). Two were \nmulti-national randomized tr",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Turns to the second question, on the effect of parallel testing on patient-important outcomes, identifying three trials with 4602 participants that reported mortality and time to diagnosis and treatment. It describes the trial settings and the risk-of-bias assessment, noting possible bias from limited enrolment hours.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p360-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 360,
      "pageEnd": 360,
      "target": "../images/img-p360-1.webp",
      "narrative": "draft",
      "content": "Fragment of the risk-of-bias summary (Figure 14): two rows of seven circles, mostly green with a plus sign (low risk) and some yellow with a question mark (unclear risk). Study and domain labels are not included in this image.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p360-3",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 360,
      "pageEnd": 360,
      "target": "../images/img-p360-3.webp",
      "narrative": "draft",
      "content": "Column labels of the risk-of-bias summary (Figure 14), set vertically: random sequence generation (selection bias), allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), selective reporting (reporting bias), and other bias.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-360",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 360",
      "pageStart": 360,
      "pageEnd": 360,
      "target": "../sections/page-360.md",
      "narrative": null,
      "content": "4.7.1 Impact of respiratory LC-aNAATs and urine LF-LAM on mortality in adult \ninpatients with HIV \nThree trials provided data on the impact of parallel testing with respiratory LC-aNAAT and urine LF-LAM \non mortality, compared to respiratory LC-aNAAT alone, with differences in trial specific intervention \ncharacteristics (Table 11). The pooled risk ratio for the impact of parallel testing of LF-LAM and \nrespiratory LC-aNAAT on mortality was 0.93 (95% CI 0.74 to 1.17) using a random-effects meta-analysis \n(I2 = 72%) (Figure 15).  \nUsing beside urine LF-LAM and study directed treatment recommend",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that across three trials in adult inpatients with HIV, parallel testing had a pooled mortality risk ratio of 0.93 with wide confidence intervals and high heterogeneity. It contrasts one trial showing a 4% absolute mortality reduction with another reporting higher mortality in the intervention arm.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p361-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 361,
      "pageEnd": 361,
      "target": "../images/img-p361-1.webp",
      "narrative": "draft",
      "content": "Forest plot (Figure 15) of risk ratios (M-H, random, 95% CI) for mortality comparing parallel testing with LC-aNAAT in three trials: Ahsberg 2023 (1.28 [0.95, 1.74]), Gupta-Wright 2018 (0.86 [0.74, 1.01]) and Peter 2016 (0.79 [0.65, 0.97]). The pooled risk ratio is 0.93 [0.74, 1.17] (2258 vs 2344 participants), with heterogeneity I² = 71% and test for overall effect P = 0.53; the axis runs from 0.01 to 100, labelled 'Favours [Parallel]' and 'Favours [LC-aNAAT]'.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-vfig-p361",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 361,
      "pageEnd": 361,
      "target": "../figures/vfig-p361.png",
      "narrative": "draft",
      "content": "A table comparing mortality in three trials of interventions including LC-aNAATs and urine LF-LAM in adult HIV-positive inpatients, with columns for population, illness severity, design, time of assessment, mortality analysis, mortality in intervention and control, and other findings. Peter 2016 (1609 adults, RCT, 8 weeks): 17.4% vs 21.9%, ARR 4%; Gupta-Wright 2018 (2574 adults, RCT, 56 days): 18% vs 21%, aRD −2.8% (P = 0.074); Åhsberg 2023 (422 adults, cluster randomized trial, 8 weeks): 32.4% vs 25.2%, log-rank P = .088.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-361",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 361",
      "pageStart": 361,
      "pageEnd": 361,
      "target": "../sections/page-361.md",
      "narrative": null,
      "content": "Table 18.  Comparison of mortality in randomized trials evaluating interventions that included respiratory LC-aNAATs and \nurine LF-LAM in adult inpatients with HIV. \n \nStudy \nPopulation \nIllness severity \nmetrics \nDesign \nTime of \nmortality \nassessment \nMortality analysis Mortality in \nintervention \nMortality in \ncontrol \nOther key \nfindings  \nPeter 2016 1609 HIV-\npositive \nadults, \ninpatients \n(with TB \nsymptoms) \nMedian CD4 84 \ncells per L, \nKarnofksy score \n50, BMI 18.8, \nmedian \nhaemoglobin 9.2 \ng/dL \nRandomized \ncontrolled \ntrial \n8 weeks \nARR 4% (1 to 7%) 17.4% \n(139/798) \n21.9% \n(178/8",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Tabulates the three trials' populations, designs, timing and mortality results, and begins describing how mortality differed by CD4 count.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-018",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 18. Comparison of mortality in randomized trials evaluating interventions that included respiratory LC-aNAATs and urine LF-LAM in adult inpatients with HIV.",
      "pageStart": 361,
      "pageEnd": 361,
      "target": "../tables/table-18.csv",
      "narrative": "draft",
      "content": "Mortality results of three randomized trials in adult inpatients with HIV that evaluated interventions including respiratory LC-aNAATs and urine LF-LAM. For each trial the table gives the population, illness severity metrics, design, time of mortality assessment, mortality analysis, mortality in the intervention and control groups, and other key findings.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-362",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 362",
      "pageStart": 362,
      "pageEnd": 362,
      "target": "../sections/page-362.md",
      "narrative": null,
      "content": "2018 reported that in their trial of adult inpatients with HIV, use of urine LF-LAM in conjunction with \nother tests including respiratory LC-aNAAT (Xpert MTB/RIF) and urine LC-aNAAT (Xpert MTB/RIF) was \nassociated with an adjusted risk difference of -0.1 (-3.3 to 3.1, P = 0.96) in patients with a CD4 count ≥ \n100 cells per µL, and with an adjusted risk difference of -7.1 (-13.7 to -0.4, P = 0.036) in patients with a \nCD4 count < 100 cells per µL. Åhsberg 2023 found mortality of 27% in the intervention group with \nparallel testing compared to 17% with LC-aNAAT alone (p=0.88) for those with cou",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes mortality by CD4 count, with one trial showing a significant mortality reduction from adding LF-LAM only in people with CD4 counts under 100. It then begins reporting the effect of parallel testing on the proportion diagnosed with tuberculosis, with forest plots.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p363-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 363,
      "pageEnd": 363,
      "target": "../images/img-p363-1.webp",
      "narrative": "draft",
      "content": "Forest plot (Figure 16) of risk ratios (M-H, random, 95% CI) comparing parallel testing with LC-aNAAT in two trials of the proportion of adult inpatients with HIV diagnosed with tuberculosis: Ahsberg 2023 (1.82 [1.31, 2.54]) and Gupta-Wright 2018 (1.47 [1.24, 1.73]). The pooled risk ratio is 1.56 [1.29, 1.88] (1460 vs 1533 participants), with I² = 23% and test for overall effect P < 0.00001; the axis runs from 0.01 to 100, labelled 'Favours [LC-aNAAT]' and 'Favours [Parallel]'.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p363-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 363,
      "pageEnd": 363,
      "target": "../images/img-p363-2.webp",
      "narrative": "draft",
      "content": "Forest plot (Figure 17) of risk ratios (M-H, random, 95% CI) comparing parallel testing with LC-aNAAT in two trials of the proportion of adult inpatients with HIV with a confirmed tuberculosis diagnosis: Ahsberg 2023 (4.27 [2.46, 7.40]) and Gupta-Wright 2018 (2.47 [1.94, 3.14]). The pooled risk ratio is 3.06 [1.82, 5.16] (1460 vs 1533 participants), with I² = 68% and test for overall effect P < 0.0001; the axis runs from 0.01 to 100, labelled 'Favours [LC-aNAAT]' and 'Favours [Parallel]'.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-363",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 363",
      "pageStart": 363,
      "pageEnd": 363,
      "target": "../sections/page-363.md",
      "narrative": null,
      "content": "Gupta-Wright 2018 reported that the proportion of participants with a microbiological test positive \ntuberculosis diagnosis (including LF-LAM result) was higher in the intervention group that included use \nof respiratory and urine LC-aNAAT (Xpert MTB/RIF) and LF-LAM compared to the standard-of-care group \nthat included LC-aNAAT but not LF-LAM (16% [210/1287] versus 7% [85/1287]; adjusted risk difference \n9.9% [95% CI 7.5 to12.4]; p<0.0001). Åhsberg 2023 reported that more patients in the intervention \ngroup that included use of LC-aNAAT (Xpert MTB/RIF) and LF-LAM (26% [45/173]) had a microbiol",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that trials found more inpatients received a microbiologically confirmed tuberculosis diagnosis when LF-LAM was added to LC-aNAAT testing. In one trial LF-LAM was the only positive test in 41% of diagnosed cases.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p364-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 364,
      "pageEnd": 364,
      "target": "../images/img-p364-1.webp",
      "narrative": "draft",
      "content": "Forest plot (Figure 18) of risk ratios (M-H, random, 95% CI) comparing parallel testing with LC-aNAAT in two trials of the proportion of adult inpatients with HIV treated for tuberculosis: Ahsberg 2023 (1.47 [0.95, 2.27]) and Gupta-Wright 2018 (1.47 [1.24, 1.75]). The pooled risk ratio is 1.47 [1.25, 1.73] (1460 vs 1533 participants), with I² = 0% and test for overall effect P < 0.00001; the axis runs from 0.01 to 100, labelled 'Favours [LC-aNAAT]' and 'Favours [Parallel]'.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-364",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 364",
      "pageStart": 364,
      "pageEnd": 364,
      "target": "../sections/page-364.md",
      "narrative": null,
      "content": "sputum Xpert MTB/RIF alone.  \n \n4.7.4 Impact of respiratory LC-aNAATs and urine LF-LAM on time to treatment \nand proportion treated for tuberculosis in adult inpatients with HIV \nTwo trials evaluated time to diagnosis, reporting median and interquartile ranges, which precluded \nmeta-analysis of mean differences. Using bedside urine LF-LAM and study directed treatment \nrecommendation, Peter 2016 reported that the time from enrolment to treatment was lower for the \nintervention group compared to the standard-of-care groups: median 0 days (IQR 0 to 2) versus 1 day \n(IQR 0 to 3) (p <0.0001). \nUsin",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports trial findings on time to tuberculosis treatment and the proportion treated, with bedside LF-LAM shortening time from enrolment to treatment in one trial. Another trial noted that fewer than half of those tested with LF-LAM started treatment.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-365",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 365",
      "pageStart": 365,
      "pageEnd": 365,
      "target": "../sections/page-365.md",
      "narrative": null,
      "content": "5. DISCUSSION  \nThis systematic review summarizes the evidence from 31 fully paired comparative accuracy studies on \nthe parallel use of respiratory LC-aNAAT and urine LF-LAM for detection of tuberculosis disease in adults \nand adolescents with HIV presenting with signs and symptoms of tuberculosis or screen positive, \ncompared to each test alone. \nWe identified three randomized controlled trials that evaluated the impact of a parallel testing strategy \nusing respiratory LC-aNAAT and urine LF-LAM versus a strategy with respiratory LC-aNAAT alone on \nmortality and other patient-important outcom",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Opens the discussion by summarizing the evidence from 31 paired accuracy studies and three randomized trials on parallel respiratory LC-aNAAT and urine LF-LAM testing in adults and adolescents with HIV. It restates key pooled accuracy results, including for seriously ill participants.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-366",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 366",
      "pageStart": 366,
      "pageEnd": 366,
      "target": "../sections/page-366.md",
      "narrative": null,
      "content": "o Specificity difference -14.4% (95% CrI -24.2 to -7.4) \n \nFor the subgroup of inpatients, the diagnostic accuracy of the parallel use of respiratory LC-aNAAT and \nurine LF-LAM using a microbiological reference standard is:   \n• \nSensitivity 77.7% (95% CrI 68.5 to 85.9) and specificity 87.2% (95% CrI 81.3 to 91.6) \nCompared to respiratory LC-aNAAT alone, difference in accuracy is: \no Sensitivity difference 11.7% (95% CrI 4.6 to 23.0) \no Specificity difference -9.5% (95% CrI -14.5 to -5.9) \n \nFor the subgroup of outpatients, the diagnostic accuracy of the parallel use of respiratory LC-aNAAT an",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the summary of results with accuracy estimates for inpatients and outpatients and notes that the estimated sensitivity gain of parallel testing was larger with a composite reference standard.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-367",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 367",
      "pageStart": 367,
      "pageEnd": 367,
      "target": "../sections/page-367.md",
      "narrative": null,
      "content": "microbiological reference standard (16.0 % vs. 6.7%), despite an overall estimated sensitivity of parallel \ntesting (67.6%) that was lower than the sensitivity estimate with an microbiological reference standard \n(77.5%).  \nStudy quality may influence estimates of diagnostic accuracy. Comparing parallel testing with respiratory \nLC-aNAAT alone, the increase in sensitivity was highest in studies at low risk of bias in the Patient \nSelection domain (QUADAS-C), namely those that included participants regardless of their inability to \nproduce sputum samples [Gupta-Wright 2018; Lawn 2017b; Stead 20",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Discusses how reference standards and study quality influenced estimates, with the largest sensitivity gains in studies that enrolled people regardless of their ability to produce sputum. It also notes that Xpert Ultra has largely replaced Xpert MTB/RIF and is expected to be more sensitive.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-368",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 368",
      "pageStart": 368,
      "pageEnd": 368,
      "target": "../sections/page-368.md",
      "narrative": null,
      "content": "in sensitivity.  Overall, nine studies in this review evaluated Xpert Ultra and 21 evaluated Xpert MTB/RIF. \nIn sensitivity analysis, there were no differences in the estimated sensitivity and specificity of parallel \ntesting using Xpert Ultra or Xpert MTB/RIF with urinary LF-LAM, compared to either test alone.   \n5.3 Parallel testing - differences to previous Cochrane systematic review  \nSystematic reviews and meta-analyses typically assess tuberculosis diagnostic accuracy by individual \ntests, or may compare diagnostic accuracy review of two or more test typically when a new version of a \nte",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes that sensitivity analyses found no difference between Xpert Ultra and Xpert MTB/RIF in parallel testing, and explains how this review differs from earlier Cochrane reviews by assessing tests in combination. It then turns to the trials on patient outcomes and how they differed in implementing LF-LAM.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-369",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 369",
      "pageStart": 369,
      "pageEnd": 369,
      "target": "../sections/page-369.md",
      "narrative": null,
      "content": "parallel testing with sputum LC-aNAAT and urine LF-LAM has an uncertain effect on mortality (risk ratio \n0.93; 95% CI 0.74 to 1.17) at eight weeks in people with HIV compared to routine tuberculosis diagnostic \ntesting without LF-LAM; however, there was heterogeneity between individual trial interventions and \neffect estimates and the pooled results were statistically imprecise. Parallel testing with sputum LC-\naNAAT and urine LF-LAM in inpatient settings among people with HIV likely increases the proportion of \npeople diagnosed, including increases in the proportion with a microbiologically t",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarizes that in inpatients with HIV parallel testing has an uncertain effect on mortality but likely increases the share diagnosed and probably the share treated for tuberculosis. It also discusses limitations of the evidence and its relevance for decision makers.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-370",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 370",
      "pageStart": 370,
      "pageEnd": 370,
      "target": "../sections/page-370.md",
      "narrative": null,
      "content": "6. Author’s conclusions \n6.1 Implications for practice  \nLC-aNAAT on respiratory sample is often the initial test to detect tuberculosis disease. Urine LF-LAM is \nan add on test that is in combination with clinical judgement and other tests (parallel use) for the \ndiagnosis of HIV-associated tuberculosis. Respiratory LC-aNAAT and urine LF-LAM may be \ncomplementary when performed as part of a parallel testing strategy to diagnose all forms of HIV-\nassociated tuberculosis. Reasons for this observed increase in diagnostic sensitivity with addition of \nurine LF-LAM could be its ease of use, abilit",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Sets out the authors' conclusions for practice, describing respiratory LC-aNAAT as the usual first test and urine LF-LAM as a complementary add-on that can detect more HIV-associated tuberculosis, including in people unable to give sputum. It calls for implementation research on barriers and enablers that affect clinical decisions.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-371",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 371",
      "pageStart": 371,
      "pageEnd": 371,
      "target": "../sections/page-371.md",
      "narrative": null,
      "content": "Other notable research gaps include the lack of data on the effect of parallel use diagnostic strategies on \npatient-important outcomes in children and in outpatient settings.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes further research gaps, including the lack of outcome data for parallel testing in children and in outpatient settings.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-372",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 372",
      "pageStart": 372,
      "pageEnd": 372,
      "target": "../sections/page-372.md",
      "narrative": null,
      "content": "Acknowledgements  \nWe are grateful to Vittoria Lutje, the CIDG Information Specialist, for help with the search strategy. We \nthank all authors of primary studies for answering our questions and providing data.  \nContributions of authors  \nAll authors contributed significantly to this review. SB, BY, and JÅ wrote the review sections. RN wrote \nthe sections related to impact review. SB and JÅ performed the majority of the screening and data \nextraction with assistance from BS, HP, AR and LO. TBR with assistance from IL and CMD re-analysed \npreviously shared individual participant data for the p",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives the review's acknowledgements and authors' contributions, followed by the start of its reference list.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-373",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 373",
      "pageStart": 373,
      "pageEnd": 373,
      "target": "../sections/page-373.md",
      "narrative": null,
      "content": "Beynon, R., M. M. Leeflang, S. McDonald, A. Eisinga, R. L. Mitchell, et al. 2013. \"Search strategies to \nidentify diagnostic accuracy studies in MEDLINE and EMBASE.\" Cochrane Database Syst Rev \n2013 (9): MR000022. \nBIONEER. 2022. \"IRON-qPCR.\" https://eng.bioneer.com/20-a-2600.html. \nBjerrum, S., T. Broger, R. Szekely, S. Mitarai, J. A. Opintan, et al. 2020. \"Diagnostic Accuracy of a Novel \nand Rapid Lipoarabinomannan Test for Diagnosing Tuberculosis Among People With Human \nImmunodeficiency Virus.\" Open Forum Infect Dis 7 (1): ofz530. \nBjerrum, S., E. Kenu, M. Lartey, M. J. Newman, K. K. Addo,",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Part of the review's reference list, citing studies on diagnostic search strategies and the accuracy of urine LAM tests for tuberculosis in people with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-374",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 374",
      "pageStart": 374,
      "pageEnd": 374,
      "target": "../sections/page-374.md",
      "narrative": null,
      "content": "Calligaro, G. L., L. S. Zijenah, J. G. Peter, G. Theron, V. Buser, et al. 2017. \"Effect of new tuberculosis \ndiagnostic technologies on community-based intensified case finding: a multicentre randomised \ncontrolled trial.\" The Lancet Infectious Diseases 17 (4): 441-450. \nCepheid. 2023a. \"Xpert MTB/RIF \". Accessed November 12 https://www.cepheid.com/en-US/tests/tb-\nemerging-infectious-diseases/xpert-mtb-rif.html. \nCepheid. 2023b. \"Xpert MTB/RIF Ultra.\" Accessed November 12. https://www.cepheid.com/en-\nGB/tests/tb-emerging-infectious-diseases/xpert-mtb-rif-ultra.html. \nChu, H., and S. R. Cole. 2",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Part of the review's reference list, citing trials and accuracy studies of tuberculosis diagnostics, Xpert product pages and statistical methods papers.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-375",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 375",
      "pageStart": 375,
      "pageEnd": 375,
      "target": "../sections/page-375.md",
      "narrative": null,
      "content": "Horne, D. J., M. Kohli, J. S. Zifodya, I. Schiller, N. Dendukuri, et al. 2019. \"Xpert MTB/RIF and Xpert \nMTB/RIF Ultra for pulmonary tuberculosis and rifampicin resistance in adults.\" Cochrane \nDatabase Syst Rev 6 (6): CD009593. \nHuerga, H., M. Bastard, A. V. Lubega, M. Akinyi, N. T. Antabak, et al. 2023a. \"Novel FujiLAM assay to \ndetect tuberculosis in HIV-positive ambulatory patients in four African countries: a diagnostic \naccuracy study.\" The Lancet Global Health 11 (1): e126-e135. \nHuerga, H., M. Bastard, A. V. Lubega, M. Akinyi, N. T. Antabak, et al. 2023b. \"Novel FujiLAM assay to \ndetec",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Part of the review's reference list, citing Cochrane reviews of Xpert assays and studies of the FujiLAM assay in people with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-376",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 376",
      "pageStart": 376,
      "pageEnd": 376,
      "target": "../sections/page-376.md",
      "narrative": null,
      "content": "Kohli, M., I. Schiller, N. Dendukuri, M. Yao, K. Dheda, et al. 2021b. \"Xpert MTB/RIF Ultra and Xpert \nMTB/RIF assays for extrapulmonary tuberculosis and rifampicin resistance in adults.\" Cochrane \nDatabase Syst Rev 1 (1): CD012768. \nLawn, S. D., A. D. Kerkhoff, R. Burton, C. Schutz, A. Boulle, et al. 2017a. \"Diagnostic accuracy, \nincremental yield and prognostic value of Determine TB-LAM for routine diagnostic testing for \ntuberculosis in HIV-infected patients requiring acute hospital admission in South Africa: a \nprospective cohort.\" BMC Med 15 (1): 67. \nLawn, S. D., A. D. Kerkhoff, R. Burton",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Part of a reference list citing mainly diagnostic accuracy studies and Cochrane reviews of Xpert assays and urine LAM tests for TB in people with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-377",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 377",
      "pageStart": 377,
      "pageEnd": 377,
      "target": "../sections/page-377.md",
      "narrative": null,
      "content": "antiretroviral therapy initiators: a prospective diagnostic accuracy study.\" MedRxiv : the Preprint \nServer for Health Sciences 05: 05. \nReitsma, J. B., A. S. Glas, A. W. Rutjes, R. J. Scholten, P. M. Bossuyt, and A. H. Zwinderman. 2005. \n\"Bivariate analysis of sensitivity and specificity produces informative summary measures in \ndiagnostic reviews.\" J Clin Epidemiol 58 (10): 982-90. \nSD BIOSENSOR. 2021. \"STANDARD M M10 MDR-TB.\" \nhttps://www.sdbiosensor.com/product/product_view?product_no=23007. \nShah, M., C. Hanrahan, Z. Y. Wang, N. Dendukuri, S. D. Lawn, et al. 2016a. \"Lateral flow urine \nli",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues a reference list of TB diagnostic studies, meta-analysis methods, manufacturer information on STANDARD M10 and Cochrane handbook chapters.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-378",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 378",
      "pageStart": 378,
      "pageEnd": 378,
      "target": "../sections/page-378.md",
      "narrative": null,
      "content": "Thit, S. S., N. M. Aung, Z. W. Htet, M. A. Boyd, H. A. Saw, et al. 2017. \"The clinical utility of the urine-\nbased lateral flow lipoarabinomannan assay in HIV-infected adults in Myanmar: an observational \nstudy.\" BMC Medicine 15 (1): 145. \nVan Hoving, D. J., S. Lahri, H. J. Lategan, M. P. Nicol, G. Maartens, and G. Meintjes. 2019. \"Brief Report: \nReal-World Performance and Interobserver Agreement of Urine Lipoarabinomannan in \nDiagnosing HIV-Associated Tuberculosis in an Emergency Center.\" Journal of Acquired Immune \nDeficiency Syndromes: JAIDS 81 (1): e10-e14. \nCovidence systematic review sof",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Ends a reference list of studies on urine LAM testing and Xpert Ultra for TB diagnosis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-379",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 379",
      "pageStart": 379,
      "pageEnd": 379,
      "target": "../sections/page-379.md",
      "narrative": null,
      "content": "_(no text on this page)_",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A blank page with no extracted text.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-380",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 380",
      "pageStart": 380,
      "pageEnd": 380,
      "target": "../sections/page-380.md",
      "narrative": null,
      "content": "Web Annex B.9. Systematic Review of Economic Evidence Around \nLow and Moderate Complexity Nucleic Acid Amplification Tests \nfor TB Diagnosis and Detection of Resistance \n  \n  \n \n \n Prepared for WHO Global TB Programme, Guideline Development Group Meeting  \nMay 6th to 11th 2024  \n  \n  \n \n  \n  \nFuneka Bango1, Suvesh Shrestha2, Pushpita Samina3, Olivia D’Silva2, Alice Zwerling2 \n1 Health Systems Research Unit, South African Medical Research Council, South Africa \n2 School of Epidemiology & Public Health, University of Ottawa, Ottawa, Canada \n3 Faculty of Health Science, McMaster University, Hamil",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title page of Web Annex B.9, a systematic review of economic evidence on low- and moderate-complexity nucleic acid amplification tests for TB diagnosis and drug resistance, prepared for a WHO guideline meeting in May 2024, with authors and affiliations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-381",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 381",
      "pageStart": 381,
      "pageEnd": 381,
      "target": "../sections/page-381.md",
      "narrative": null,
      "content": "Abbreviations \nART \nAntiretroviral therapy  \nCLHIV \nChildren living with HIV \nCXR \nChest X-ray \nDALYs \nDisability-adjusted life years  \nDR \nDrugs-resistant  \nDS \nDrug-susceptible  \nDST \nDrugs susceptibility testing  \nFIND \nFoundation for Innovative New Diagnostics \nHIV \nHuman immunodeficiency viruses \nHT \nHigh throughput systems  \nICER \nIncremental cost effectiveness ratio \nLC-NAATs \nLow complexity automated nucleic acid amplification tests  \nLF-LAM \nLateral flow urine lipoarabinomannan assay  \nLMIC \nLow- and middle-income countries  \nLT \nLow throughput systems  \nMDR-TB \nMulti-drug resistant T",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A list of abbreviations used in the economic review of TB nucleic acid amplification tests.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-382",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 382",
      "pageStart": 382,
      "pageEnd": 382,
      "target": "../sections/page-382.md",
      "narrative": null,
      "content": "1. \nBACKGROUND \n \nIn 2022, Tuberculosis (TB) was the second leading cause of death from a single infectious agent after \nCOVID-19 globally(1).  TB resulted in 1.3 million deaths in 2022 with more than 10.6 million people who \nfell ill with TB, and of the 6.2 million people diagnosed with pulmonary TB worldwide, 63% were \nbacteriologically confirmed. The World Health Organization (WHO) recommends using rapid molecular \ndiagnostic tests as the initial diagnostic test in individuals with signs and symptoms of TB. Molecular-based \ntesting offers rapid identification of the Mycobacterium tuberculos",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives background noting that TB caused 1.3 million deaths in 2022 and that WHO recommends rapid molecular tests as initial diagnostics. It explains that access remains limited in high-burden areas because of costs, infrastructure and staff shortages.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-383",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 383",
      "pageStart": 383,
      "pageEnd": 383,
      "target": "../sections/page-383.md",
      "narrative": null,
      "content": "Low-complexity automated nucleic acid amplification tests (LC-aNAATs) are a class of TB diagnostics \nincluded in the mWRDs that can be used for TB diagnosis. They include Xpert® MTB/RIF (Xpert), Xpert® \nMTB/RIF Ultra (Ultra) (Cepheid, USA), and TrueNat™ (11). These tests provide initial detection of TB and \nresistance to rifampicin detection. \n \nModerate complexity automated NAATs (MC-aNAATs) offer a high sample throughput for the initial \ndetection of TB and resistance to rifampicin and isoniazid, providing more options for early diagnosis of \nTB and rifampicin-resistant TB but also addressin",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the classes of NAATs, including Xpert, Xpert Ultra and Truenat, moderate-complexity automated tests and TB-LAMP. It notes that economic evidence for the newer tests is limited and has not been recently reviewed.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-384",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 384",
      "pageStart": 384,
      "pageEnd": 384,
      "target": "../sections/page-384.md",
      "narrative": null,
      "content": "The overall aim of the present review is to identify, evaluate, and summarize the findings of available \neconomic evidence on three classes of NAATs: \n• Low complexity automated nucleic acid amplification tests (LC-aNAATs) for TB \ndiagnosis and detection of drug resistance.   \n• Low-complexity manual NAAT (LC-mNAAT) TB-LAMP for TB diagnosis. \n• Moderate complexity automated NAATs (MC-aNAATs).  \n \n2. \nMETHODS \n \nThis systematic review protocol has been developed according to the recommendations from the \nPreferred Reporting Items for Systematic Review and Meta-Analysis Protocols (PRISMA-P) 2015",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "States the review's aim of summarising economic evidence on three NAAT classes and begins describing the methods. It sets out inclusion criteria and excludes studies of active case-finding or triage before testing.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-385",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 385",
      "pageStart": 385,
      "pageEnd": 385,
      "target": "../sections/page-385.md",
      "narrative": null,
      "content": "For the remaining classes of NAATs (LC-mNAATs and MC-aNAATs), the inclusion criteria for each class of \nNAATs included in the review are specified below. \n \n2.2. Study Population & Specimen Type: \nThe study population may include any of the following:  \n \nAll people who screen positive, with signs and symptoms of TB, or with abnormal chest X-rays, people \nliving with HIV (PLHIV), and children with signs or symptoms of TB. Specimen types may include respiratory \nor non-respiratory isolates, including stool and urine. \n \nComparator: \nWe were not restricting by diagnostic comparator or microbiolo",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Specifies the study populations, specimen types and economic outcomes eligible for the review, such as cost per test and cost per QALY gained. It introduces the search strategies developed with information specialists.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-386",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 386",
      "pageStart": 386,
      "pageEnd": 386,
      "target": "../sections/page-386.md",
      "narrative": null,
      "content": "reviewers screened titles and abstracts retrieved using the search strategies according to the inclusion \ncriteria. Following this, full texts of studies assessed to meet the inclusion criteria were independently \nassessed for eligibility by two reviewers. All disagreements were resolved through discussion and the \ninvolvement of a third reviewer if a consensus could not be reached.   \n \nOverall, the search strategy contained three overarching concepts: concept 1 (tuberculosis), concept 2 \n(nucleic acid amplification test specified for each NAAT), and concept 3 (costs and cost analysis). \n \n F",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes dual independent screening and data extraction, with disagreements resolved by a third reviewer. It lists the key economic outcomes extracted, such as cost per patient tested and incremental cost-effectiveness ratios.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-387",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 387",
      "pageStart": 387,
      "pageEnd": 387,
      "target": "../sections/page-387.md",
      "narrative": null,
      "content": "were inflated to 2024 in the local currency and converted to 2024 USD (United States Dollars) using the \nonline currency exchange rate (17).   \n \nAssessing the Quality of Evidence:  \nThe quality of reporting for complete economic evaluation studies included in the review were assessed \nusing a modified version of the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) \nchecklist (18). As some items do not apply to partial economic evaluation or costing studies, we assessed \nthese studies using a version of the CHEERS checklist, which focused on appropriate items for assessing",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains that costs were converted to 2024 US dollars and that reporting quality was assessed with a modified CHEERS checklist. Studies were also summarised by population, HIV status, type of TB, level of care and sample type.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-388",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 388",
      "pageStart": 388,
      "pageEnd": 388,
      "target": "../sections/page-388.md",
      "narrative": null,
      "content": "3.1. Systematic review of LC aNAATs (Xpert, Ultra and TrueNat) \n \nSummary of Findings: \nA total of 1534 studies were identified in the original search, after removing duplicates, 736 potentially \nrelevant studies were screened. Of these, 107 were assigned for full-text review and were evaluated \nagainst the inclusion and exclusion criteria, and 29 studies were included in the final systematic review \n(Figure 1). Of the 29 included studies, 22 (76%) assessed Xpert, six (21%) Ultra, and one study (3%) \nevaluated TrueNat. Ten of the studies evaluating Xpert were cost-effectiveness analyses, and 1",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that 29 studies on LC-aNAATs were included from 1,534 records identified, and shows the PRISMA flow diagram with reasons for exclusion.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-389",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 389",
      "pageStart": 389,
      "pageEnd": 389,
      "target": "../sections/page-389.md",
      "narrative": null,
      "content": "General characteristics of included full economic evaluation studies.  \n \n \n \nXpert cost-effectiveness studies (n=10) \nTen cost-effectiveness analyses evaluated Xpert (19–28). These were conducted across various settings \ncovering all income levels, including one study from the USA, one each from Brazil and South Africa, two \nfrom Thailand, and Uganda and one each from Columbia, Egypt, and Ethiopia (Table 1).  \n \nMost of the studies included adult populations, with only one study including both adults and children \n(29). Three studies included HIV-infected adults (21,23,26).  With one study in",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the ten cost-effectiveness studies of Xpert across countries at all income levels and the studies of Xpert Ultra. Some compared Ultra with smear microscopy and Xpert, while others assessed parallel use with LAM or whole-genome sequencing.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-390",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 390",
      "pageStart": 390,
      "pageEnd": 390,
      "target": "../sections/page-390.md",
      "narrative": null,
      "content": "TrueNat (n=1) \nOne full economic evaluation study conducted in India evaluated the cost-effectiveness of TrueNat in \ndesignated microscopy centers compared to smear and Xpert in children below 15 years of age for drug-\nsensitive and resistant TB (40).",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes that one study in India evaluated the cost-effectiveness of Truenat in children compared with smear and Xpert.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-391",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 391",
      "pageStart": 391,
      "pageEnd": 391,
      "target": "../sections/page-391.md",
      "narrative": null,
      "content": "386\nTable 1. General characteristics of included full economic evaluation studies.  \nCitation  \nCountry \nStudy \npopulation  \nType of TB \ndisease \nStudy Setting  Diagnostic \nIntervention \nDiagnostic \nComparator \nType of sample \nXpert \n Chitpim et al., \n2022(41) \nThailand \nAdults (≥15 \nyears old) \nDS and DT-\nTB \nNot reported \nXpert MTB/RIF \nInitial \n(As an initial \ndiagnostic test in \nall patients or \ninstead of SSM) \n \nXpert MTB/RIF \nAdd-on \n(In smear-\nnegative patients, \nas a secondary \ndiagnostic or \nfollow-on test to \nSSM rather than \nculture) \nSSM + \nCulture/DST \n \nTB-LAMP Initial \n(As an i",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 1 begins listing the included full economic evaluations, with country, population, type of TB, setting, intervention, comparator and sample type.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-392",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 392",
      "pageStart": 392,
      "pageEnd": 392,
      "target": "../sections/page-392.md",
      "narrative": null,
      "content": "387\nCadavid et al., \n2022(29) \nColombia \nChildren and \nadults \nDS and DR-\nTB \nSpecialised \nlab \nXpert \nSSM \nMGIT culture \nRespiratory \nsamples \nDa Silva et al., \n2021(43) \nBrazil \nAdults (≥18 \nyears old) \nDR-TB \nUnclear \nXpert \nBactec MGIT 960 \nSystem \nSputum \nEsmail et al., \n2020(30,31) \nSouth \nAfrica, \nTanzania, \nZambia, \nand \nZimbabwe \nAdults (≥18 \nyears old)  \nUnspecified Hospital  \nConcurrent LAM \n+ Xpert \nSequential Xpert \nfollowed by LAM \nSequential LAM \nfollowed by Xpert \nXpert alone \nLAM alone \nSputum culture \nas a reference \nstandard \nSputum \nHickey et al., \n2022 (31) \nUnited \nStates",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues Table 1 with economic evaluations of Xpert from Colombia, Brazil, several African countries, Ethiopia and Uganda.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-393",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 393",
      "pageStart": 393,
      "pageEnd": 393,
      "target": "../sections/page-393.md",
      "narrative": null,
      "content": "388\nKhumsri et al., \n2020(46) \nThailand \nAdults (≥18 \nyears old) \nDS and DR-\nTB \nHospital  \nXpert  \nSputum AFB \nsmear \nSputum \nSedrak et al., \n2022(47) \nEgypt \nHIV-negative \nadults \nDS and DR-\nTB \nHospital  \nXpert + sputum \nmicroscopy \nSputum \nmicroscopy and \nculture \nSputum \nUltra \nFekadu, Wang & \nYou, 2023 (48) \nSouth \nAfrica \nAdult HIV-\ninfected \nindividuals with \nsigns and \nsymptoms of \nTB disease. \nDS and DR -\nTB \nOutpatients \nand \ninpatients \nUltra \nUltra with \nAlereLAM \nUltra with \nFujiLAM \nXpert \nXpert with \nAlereLAM \nXpert with \nFujiLAM \nSputum smear \nmicroscope (SSM) \nSputum \nUrine",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues Table 1 with studies from Thailand and Egypt and economic evaluations of Xpert Ultra in South Africa, several other countries, Ethiopia and Indonesia.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-394",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 394",
      "pageStart": 394,
      "pageEnd": 394,
      "target": "../sections/page-394.md",
      "narrative": null,
      "content": "389\nMugwagwa, \nAbubakar & \nWhite, 2021(50) \nEngland \nand \nWales \nSouth-Asian and \nBlack-African \nethnic/social \ngroups \nDS and DR-\nTB \nInpatient and \noutpatient \nUltra \nSimultaneous \nUltra and WGS \n Xpert  \nReplacing the \nculture-based \nDST with WGS \nSimultaneously \ndoing WGS and \nXpert \nSputum \nPooran et al., \n2022 (51) \nSouth \nAfrica \nSuspected TB \ncases \nDS-TB \nPrimary care \nclinics \nUltra \nSmear \nmicroscopy  \nMGIT \nAdenosine \ndeaminase  \nInterferon-\ngamma Release \nImmuno-\nSuspension Assay \nor (IRISA-TB) \nPleural fluid \nThompson et al., \n2023 (52) \nUganda \nAdults ≥18 \nyears of age \nDS-TB \nP",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes Table 1 with Ultra studies from England and Wales, South Africa and Uganda and the Truenat study in India.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-395",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 395",
      "pageStart": 395,
      "pageEnd": 395,
      "target": "../sections/page-395.md",
      "narrative": null,
      "content": "390\nGeneXpert MTB \nin Designated \nMicroscopy \nCentre",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A table fragment naming Xpert in designated microscopy centres as a comparator.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-396",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 396",
      "pageStart": 396,
      "pageEnd": 396,
      "target": "../sections/page-396.md",
      "narrative": null,
      "content": "3.1.1. Methodological characteristics of included full economic evaluation studies. \n \nXpert cost-effectiveness studies (n=10) \nSeven of the ten studies were done from the provider perspective (20–24,27,28), one from a clinical trial \nperspective (26) and two from a societal perspective (19,25). All full economic evaluation studies \nperformed a variation of cost-effectiveness analyses (CEA), four presenting utility outcomes, three of \nwhich presented Quality-adjusted life years (QALYs) (19,25,26)and one disability adjusted life years \n(DALYs) (27).   \n \nUltra cost-effectiveness studies (n=6)",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarises the methods of the full economic evaluations, most of which took a provider perspective and used cost-effectiveness or cost-utility analysis with sensitivity analyses. The Truenat study used a decision tree model with DALYs as the outcome.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-397",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 397",
      "pageStart": 397,
      "pageEnd": 397,
      "target": "../sections/page-397.md",
      "narrative": null,
      "content": "392\nTable 2. Methodological characteristics of full economic evaluation studies  \nCitation  \nPerspective  \nType of \neconomic \nanalysis  \nYear of \nanalysis \nCurrency \nPrimary outcome \nPrimary \ncosting  \nModelling  \nType of model used \nSensitivity \nanalysis \nXpert  \nChitpim et \nal., 2022(41) \nSocietal  \nCUA \n2021 \n฿ (Thai Baht) \nQALYs \nNo  \nYes \nA hybrid decision \ntree Markov model \nOne-way \nsensitivity \nanalysis and \nprobability \nsensitivity \nanalysis  \n Ejalu et al., \n2022 (42) \nHealthcare \nProvider \nCEA \nmid-\n2018 \nUS$ \nAdditional PTB \ntest diagnosed \nwhen the Xpert \nMTB/Rif test or \nXpert Om",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 2 begins listing the methodological characteristics of the economic evaluations, including perspective, analysis type, currency, outcome and model.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-398",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 398",
      "pageStart": 398,
      "pageEnd": 398,
      "target": "../sections/page-398.md",
      "narrative": null,
      "content": "393\nthe sixth month \nof follow-up. \nEsmail et al., \n2020 (30) \nHealthcare \nprovider \nCEA \n2018 \nUS$  \nCulture-positive \ncase diagnosed \nand started on \ntreatment \nUnclear \nYes \nNot specified \nOne-way \nHickey et \nal., 2022 \n(31) \nHealthcare \nprovider \nCEA \nUnspecifi\ned \nUS$ \nHospital length-\nof-stay \nUnclear \nYes \nDecision-analytic \nmodel in a decision \ntree with three arms \ncomparing 3 AFB \nsmears (3-AFB), a \nsingle Xpert test (1-\nXpert), and 2 Xpert \ntests (2-Xpert). \nCost scenario \nanalysis  \nKaso & \nHailu, 2021 \n(44) \nHealthcare \nprovider \nCost and \nCEA \n2018 \n US$ \nCase detected \nYes \nNo",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues Table 2 with the methods of economic studies of Xpert, including outcomes, models and sensitivity analyses.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-399",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 399",
      "pageStart": 399,
      "pageEnd": 399,
      "target": "../sections/page-399.md",
      "narrative": null,
      "content": "394\nSedrak et \nal., 2022 \n(47) \nHealthcare \nprovider \nCEA \n2017 \nE£ \n DALYs \ndiagnostic \ncosts, \ndevices and \ntest costs, \ntreatment \ncosts, \npatient costs \nYes \nDecision tree \nanalysis model \nDeterministic \nand probabilistic \nsensitivity \nanalysis \nUltra \n \n \n \n \n \n \n \n \n \nFekadu, \nWang & \nYou, 2023 \n(48) \nHealthcare \nprovider \nCUA \n2023 \nUS$ \nDALYs \nNo \nYes \nDecision tree model \nOne-way, \nprobabilistic, and \nscenario analysis \nIem et al., \n2023 (49) \nHealthcare \nprovider \nCEA/CMA \n2021 \nUS$ \nBacteriologically \nconfirmed TB \ncase detected.  \nYes \nNo \nN/A \nOne way \nMafirakurev\na et al., \n2022",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues Table 2 with the methods of studies from Egypt and of Xpert Ultra evaluations, including decision tree and transmission models.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-400",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 400",
      "pageStart": 400,
      "pageEnd": 400,
      "target": "../sections/page-400.md",
      "narrative": null,
      "content": "395\ndiagnosed and \ninitiated on \ntreatment \nThompson \net al., 2023 \n(52) \nHealthcare \nprovider \nCost + CEA \n2019 \nUS$ \nCost outcomes: \nper-patient and \nper-test cost  \nCost-\neffectiveness \noutcomes: \nincremental \npatient with \nmicrobiologically \nconfirmed \ntuberculosis \ndisease initiating \ntreatment within \n14 days of initial \npresentation \nYes \nNo \nModelling used for \nstatistical analysis \nOne-way and \nprobabilistic \nTrueNat \nBhatia et al., \n2023 (53) \nHealthcare \nprovider \nCUA \n2020 \nUS$ \nDALYs averted \nNo \nYes \nDecision tree \nNot reported",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes Table 2 with methods of a Ugandan Ultra study and the Indian Truenat cost-utility study.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-401",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 401",
      "pageStart": 401,
      "pageEnd": 401,
      "target": "../sections/page-401.md",
      "narrative": null,
      "content": "Quality assessment \n \nTwo reviewers independently evaluated the reporting quality of each included study, and any \ndisagreement was resolved by discussion. An 18-item modified Consolidated Health Economic Evaluation \nReporting Standards (CHEERS) statement was used. The CHEERS statement is not a tool for assessing the \nquality of methods and does not provide an overall score. Instead, it assesses the quality of reporting. \nHowever, we estimated an overall score from the sum of items reported by each study compared to those \nindicated in the modified checklist. Each item that was met was rated a",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes quality assessment with an 18-item modified CHEERS checklist. It begins reporting cost-effectiveness results, including ICERs for Xpert and LAM strategies in hospitalized people with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-402",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 402",
      "pageStart": 402,
      "pageEnd": 402,
      "target": "../sections/page-402.md",
      "narrative": null,
      "content": "Xpert. For parallel LAM + Xpert, the ICER was $2338. Despite this, parallel testing was the preferred \nstrategy for identifying more cases. \n \nKaso & Hailu, 2021 compared Xpert to smear, and the ICER was $28 per additional case diagnosed (44). \n \nCadavid et al., 2022 compared Xpert to smear and culture in adults & children and found Xpert to be more \ncost-effective with an ICER of $1017 compared to $1039 for culture (29).  \n \nHickey et al., 2022, compared one or two Xpert tests to smear and culture. One Xpert was found to be \ncost-saving at $6811 per patient (31). \n \nXpert cost-effectiveness s",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports further cost-effectiveness results, such as Xpert's low cost per additional case compared with smear in Ethiopia and in Colombia. It also reports that pooling sputum for Ultra or Xpert reduced costs per test and per confirmed case.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-403",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 403",
      "pageStart": 403,
      "pageEnd": 403,
      "target": "../sections/page-403.md",
      "narrative": null,
      "content": "pooling. The study concluded that pooled sputum testing could be a cost-effective strategy for diagnosing \nTB, resulting in significant resource savings, as it could increase testing capacity and affordability in \nresource-limited settings. \n \nPooran et al., 2022 determined the cost-effectiveness of five different testing strategies [Smear \nmicroscopy, MGIT, Adenosine deaminase, Interferon-gamma Release Immuno-Suspension Assay or \n(IRISA-TB)] for pleural TB diagnosis in South Africa (51). The Ultra baseline test was $13.28. IRISA-TB was \nthe most cost-effective strategy, as the cost per pleura",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that pooled sputum testing could save resources and summarises studies of Ultra for pleural TB and in children's stool in Ethiopia and Indonesia, with ICERs per DALY averted. It introduces a study of whole-genome sequencing in England and Wales.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-404",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 404",
      "pageStart": 404,
      "pageEnd": 404,
      "target": "../sections/page-404.md",
      "narrative": null,
      "content": "sensitivity testing, compared to using Ultra or the simultaneous use of WGS and Ultra for routine TB \ndiagnosis (50). The study showed that WGS shortens the time to drug-sensitivity testing and treatment \nmodification where necessary, reducing treatment and hospitalisation costs, with an Incremental Net \nbenefit (INB) of £11.1M. Molecular testing shortens the time to TB diagnosis and treatment. While this \ncauses an increase in annual costs of treatment, it averts transmissions and future active TB disease, \nresulting in cost savings and health benefits to achieve an INB of £12.07M (GeneXpert",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that whole-genome sequencing could shorten time to drug susceptibility results and save costs. It also states that Truenat was found more cost-effective than Xpert and smear microscopy in India.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-405",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 405",
      "pageStart": 405,
      "pageEnd": 405,
      "target": "../sections/page-405.md",
      "narrative": null,
      "content": "400\nTable 3. Cost components included in full economic evaluation studies  \nClass of LC-\naNAAT \nCitation \nCountry \nCosts included in the analysis \nCost per test (2024 \nUS$) \n \nStaff \nLab \nspace \nEquipment \nConsumables \nOverhead  \nTransport \n  \nXpert \nPrimary costing \nDa Silva et al., \n2021(43) \nBrazil \n✓ \n✓ \n✓ \n✓ \nX \nX \nXpert with subsidy = \n$41.46 \nXpert without \nsubsidy $217.03 \nEjalu et al., 2022 \n(42) \nUganda \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \nTotal cost per test \nXpert MTB/Rif $17.65 \nXpert Omni $19.10 \nEsmail et al., \n2020 (30) \nSouth Africa, \nTanzania, \nZambia and \nZimbabwe \n✓ \nX \n✓ \n✓ \nX \nX \nCost per",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 3 begins listing cost components and costs per test in 2024 US dollars for Xpert studies in Brazil, Uganda and southern and eastern African countries.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-406",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 406",
      "pageStart": 406,
      "pageEnd": 406,
      "target": "../sections/page-406.md",
      "narrative": null,
      "content": "401\nKaso & Hailu, \n2021. (44) \nEthiopia \n✓ \n✓ \n✓ \n✓ \n✓ \nX \nCost per patient \ntested  \nXpert $28.06 \n(ranging from $27.60 \nto $28.76) \nKhumsri et al., \n2020 (46) \nThailand \n✓ \n✓ \n✓ \n✓ \n✓ \nX \nXpert $53.78 \nSedrak et al., \n2022 (47) \nEgypt \n✓ \nX \n✓ \n✓ \nX \nX \nSecondary cost analyses \nChitpim et al., \n2022 (41) \nThailand \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \nXpert unit costs \n$1,001.61 (SE \n$100.16) \nKeim-Malpass et \nal., 2023 (45) \nUganda \n✓ \n✓ \n✓ \n✓ \nX \nX \nXpert MTB/Rif $23.06 \nHickey et al., \n2022 (31) \nUS \nUnclear \nUnclear methods \n Cadavid et al., \n2022 (29) \nColombia \n✓ \n✓ \n✓ \n✓ \nX \nX \nXpert MTB/Rif $28.34 \nUlt",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues Table 3 with costs per test for Xpert in Ethiopia, Thailand, Egypt, Uganda and Colombia, and for pooled Ultra testing.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-407",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 407",
      "pageStart": 407,
      "pageEnd": 407,
      "target": "../sections/page-407.md",
      "narrative": null,
      "content": "402\nPooran et al., \n2022 (51,52) \nSouth Africa \n✓ \n✓ \n✓ \n✓ \nX \nXpert ULTRA baseline \ntest - $12.19 \nThompson et al., \n2023 (52) \nUganda \n✓ \n✓ \n✓ \n✓ \n✓ \nCentralised = $20.69 \nDecentralised = \n$28.05 \nSecondary cost analyses \nFekadu et al., \n2023 (48) \nSouth Africa \n✓ \n✓ \n✓ \n✓ \nX \nX \n$27.00 \nMugwagwa et al., \n2021 (50) \nEngland and \nWales \n✓ \n✓ \n✓ \n✓ \nX \nX \n$167.52 \nMafirakureva et \nal., 2022 (39) \nEthiopia \nIndonesia \n✓ \n✓ \nX \n✓ \nX \nX \nEthiopia = $46.39 \nIndonesia = $24.31",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes Table 3 with costs per Ultra test in South Africa, Uganda, England and Wales, Ethiopia and Indonesia.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-408",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 408",
      "pageStart": 408,
      "pageEnd": 408,
      "target": "../sections/page-408.md",
      "narrative": null,
      "content": "Xpert cost analyses (n=12) \n \nTwelve studies assessed the cost per test of Xpert (54–65). Most of these studies (46%) were done in \nlower-middle-income countries. Most studies (92%) were done from a provider perspective, and only \none was done from a societal perspective. The studies were conducted in various settings, and Xpert \nwas compared to several different comparators. Table 4 outlines the study characteristics of the 12 \ncosting studies with details on the country of analysis, study population, type of TB disease, study \nsetting, intervention, comparator, and analysis perspective.",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes twelve cost analyses of Xpert, mostly from lower-middle-income countries and a provider perspective, and the cost components they included.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-409",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 409",
      "pageStart": 409,
      "pageEnd": 409,
      "target": "../sections/page-409.md",
      "narrative": null,
      "content": "404 \nTable 4. Description of included cost analyses  \nCitation  \nCountry \nStudy population  \nType of TB \ndisease  \nStudy Setting  \nDiagnostic Intervention \nDiagnostic Comparator \nXpert \n  \n  \n  \n  \n  \n  \nAdjobimey et al., 2022 \n(66) \nBenin \nPregnant woman  \nDS-TB \nHealth centers \nXpert  \nNo comparator \nCapeding et al., 2022 \n(67) \nPhilippines \nNo specified \npopulation \nDS and DR-TB \nPrimary care facilities and hospitals \nXpert  \nSmear microscopy ZN \nCassim et al., 2021 \n(68,69) \nSouth Africa \nHigh-burden peri-\nmining \ncommunities \nDS and DR-TB \nperi-mining community setting \nMobile Xpert \nLabo",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 4 begins describing the included Xpert cost analyses in countries such as Benin, the Philippines, South Africa and Brazil.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-410",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 410",
      "pageStart": 410,
      "pageEnd": 410,
      "target": "../sections/page-410.md",
      "narrative": null,
      "content": "405 \nMuniyandi et al., 2021 \n(72) \nIndia \nUnclear \nDS TB and DR-\nTB \nPrimary care facilities, outpatients, \nand inpatients \nZN, x-ray, and Xpert  \n ZN, x-ray, Xpert and LPA \nFM, x-ray and Xpert  \nFM, x-ray, Xpert and LPA \nZN \nZN, x-ray and LPA \nZN and x-ray \nFM \nFM and x-ray \nFM, x-ray and LPA \nNadjib et al., 2022 (73) \nIndonesia \nAdults (≥15 \nyears)  \nDS-TB and DR-\nTB \nTB diagnostics facilities \nXpert  \nSputum smear microscopy  \nSSM + chest x-ray  \nSantos et al., 2021 \n(74) \nBrazil \nIncarcerated \nadult males (≥18 \nyears) \nDS-TB \nPrison \nStrategy 1: Xpert for all \nparticipants \nStrategy 2: Sym",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues Table 4 with Xpert cost analyses in India, Indonesia, Brazil and other settings, including centralized and decentralized testing.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-411",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 411",
      "pageStart": 411,
      "pageEnd": 411,
      "target": "../sections/page-411.md",
      "narrative": null,
      "content": "406 \nNote:  SSM (Sputum smear microscopy), Xpert (GeneXpert MTB/RIF), FM (Fluorescence microscopy), ZN (Ziehl–Neelsen microscopy), LED‑FM (light‑emitting diode‑fluorescent microscopy), DST (drug susceptibility \ntesting)",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes defining the abbreviations used in Table 4.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-412",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 412",
      "pageStart": 412,
      "pageEnd": 412,
      "target": "../sections/page-412.md",
      "narrative": null,
      "content": "407\nTable 5. Cost components and unit costs \nCitation \nAnalysis Perspective  \nCosts included in the analysis \nCost per test (2024US$) \n \n  \nStaff \nBuilding/ \nInfrastructure  \nEquipment \nConsumables \nOverhead  \nTransport \n  \n Xpert \n \n  \n  \n  \n  \n  \n  \n  \nAdjobimey et al., \n2022 (66) \nHealth system  \n✓ \nX \n✓ \n✓ \n✓ \n✓ \n Xpert MTB/RIF (per sample) $23.14 \nCapeding et al., \n2022 (67) \nHealth system  \n✓ \n✓ \n✓ \n✓ \n✓ \nX \nXpert MTB/RIF testing $24.69 (23.29–\n25.95) Bottom-up, and $28.4 (35.95–\n40.13) Top-down  \nCassim et al., 2021 \n(68) \nHealth system  \n✓ \n? \n✓ \n✓ \n? \n✓ \n \nThe cost per result  \nTradit",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 5 begins listing cost components and unit costs from Xpert cost analyses, with costs per test mostly between about $13 and $32 and costs per case detected in Cambodia.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-413",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 413",
      "pageStart": 413,
      "pageEnd": 413,
      "target": "../sections/page-413.md",
      "narrative": null,
      "content": "408\nCost per screening $13 \nCost per Xpert test $22.79 \nCost per TB case detected (Xpert) $434.24 \nLopes et al., 2020 \n(71) \nHealth system  \nUnclear \n  \n  \n  \n  \n  \nXpert  \nMuniyandi et al., \n2021 (72) \nHealth system  \n✓ \n✓ \n✓ \n✓ \n✓ \n? \n \nGeneXpert MTB/RIF $ 19.88 \nNadjib et al., 2022 \n(73) \nSocietal \n✓ \n✓ \n✓ \n✓ \n? \n? \n Ave cost/person tested \nHealth system$33.04 \nSocietal - Lab only + direct non-medical \n$47.06 \nSocietal - Lab + direct non-medical + \nindirect $85.16 \nSantos et al., 2021 \n(74) \nHealth system  \n✓ \n? \n✓ \n✓ \n? \n? \nCost per participant  \nXpert $23.73 \n \nCost per case \nStrategy 1:",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues Table 5 with Xpert costs per test or person from Tajikistan, India, Indonesia and Brazil, including societal costs.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-414",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 414",
      "pageStart": 414,
      "pageEnd": 414,
      "target": "../sections/page-414.md",
      "narrative": null,
      "content": "409\nYadav, Verma & \nKaushik, 2022 (77) \nLaboratory Perspective \n✓ \n? \n✓ \n✓ \n✓ \n? \nCost per test - $18.07",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes Table 5 with a laboratory-perspective Xpert cost per test from India.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-415",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 415",
      "pageStart": 415,
      "pageEnd": 415,
      "target": "../sections/page-415.md",
      "narrative": null,
      "content": "3.2.  Systematic review of LC-mNAAT (TB-LAMP): \n \n3.2.1. Summary of Findings: \nAfter eliminating 638 duplicates, our search found 1,990 studies. Six studies were included in the final \nsystematic review, as seen in Figure 2. Studies that did not involve people with TB, used LAMP as a \ndiagnostic intervention, or did not contain cost data were excluded.  \n \nOf the six included studies, one performed a cost-utility analysis (41), and two performed a cost-\naffordability analysis (79,80). The three other studies (81–83) estimated the cost of TB-LAMP.  \n \nFigure 2. PRISMA diagram for LC-mNAATs \n \n3",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that six studies were included in the review of TB-LAMP after screening, and shows the PRISMA flow diagram with reasons for exclusion.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-416",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 416",
      "pageStart": 416,
      "pageEnd": 416,
      "target": "../sections/page-416.md",
      "narrative": null,
      "content": "(81) and Cameroon (79), respectively. The studies were conducted over a range of years, from 2014 to \n2021. \n \nThe studies were conducted across various settings, including outpatient departments at health centers \n(83), peripheral laboratories (80), a laboratory setting for the development of modified TB-LAMP (82), and \nall prisons and villages involving inmates and refugees (79). Jaroenram et al. (2020) used the sputum \nsample of known cases of TB (82), and Baikunje et al. (2019) used fine needle aspiration (FNAC) of lymph \nnode samples of HIV-positive patients with TB lymphadenitis (81). Th",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the countries, years, settings and specimen types of the TB-LAMP studies, which mostly used sputum from presumptive TB cases.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-417",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 417",
      "pageStart": 417,
      "pageEnd": 417,
      "target": "../sections/page-417.md",
      "narrative": null,
      "content": "412\nTable 6: General characteristics of included full economic evaluation studies \nManus\ncript        \nCountr\ny \nsetting  \nSetting \nCurren\ncy  \nYear of \ncost \nvaluati\non  \nStudy \npopula\ntion \nDiagno\nstic \nstrateg\nies \nRefere\nnce \ndiagno\nstic \nstrateg\nies  \nAnalysi\ns \nperspe\nctive \n(health \ncare or \nsocieta\nl) \n Type \nof \necono\nmic \nevaluat\nion \nType of \nmodel \nif \napplica\nble  \nPrimar\ny \noutco\nme \nSecond\nary \noutco\nme(s) \nSource \nof \ncosting \nBaikunj\ne 2019 \n(81) \nIndia \nNot \nreporte\nd \nUSD \nJuly \n2015 to \nSeptem\nber \n2016 \nHIV \npositiv\ne \npatient\ns with \nlympha\ndenitis \nTB-\nLAMP \nMPCR, \nGeneX",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 6 lists general characteristics of the TB-LAMP economic studies, such as setting, currency, population, strategies, perspective and type of analysis. The extracted table text is broken up.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-418",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 418",
      "pageStart": 418,
      "pageEnd": 418,
      "target": "../sections/page-418.md",
      "narrative": null,
      "content": "413\nabove \n15 \nyears \n( with \nsympto\nms or \nabnor\nmal \nchest \nXray) \n-on \ntest to \nSSM)  \nTB-\nLAMP \nInitial \n (If \npositiv\ne, the \nfirst-\nline \nDST \n(FL-\nDST) \nand \nsecond\n-line \nDST \n(SL-\nDST) \nwould \nXpert \nMTB/RI\nF Add-\non \n \nXpert \nMTB/RI\nF Initial \nICER \nper \nQALY \ngained",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A fragment of Table 6 describing TB-LAMP and Xpert testing strategies and ICER per QALY outcomes.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-419",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 419",
      "pageStart": 419,
      "pageEnd": 419,
      "target": "../sections/page-419.md",
      "narrative": null,
      "content": "414\nbe \nutilized\n.) \nDonken\ng-Donf\nack \n2022 \n(79) \nCamer\noon \nAll \ninmate\ns and \nrefuge\nes in 34 \nprisons \nand 3 \nvillages \nUSD \n2019 \nPresum\nptive \nTB \ncases \nfrom \nprisons \nand \nrefuge\ne \ncamps. \nTB \nLAMP, \nfollowe\nd by \nGene \nXpert \n Gene \nXpert \nonly \nNot \nmentio\nned  \nCost \nBenefit \nAnalysi\ns \nN/A \nAmoun\nt of \nUSD \nsaved  \n \nliteratu\nre, \nGDF, \nNTP \nJaroenr\nam \n2020 \n(82) \nThailan\nd \nUSD \n2019 \nKnown \nTB \ncases  \nTB- \nLAMP \npotenti\nostat \n(TB \nLAMP-\nEC) \nCulture \nand \npreviou\nsly \ndevelo\nped \nLAMP \nassay \nNot \nmentio\nned  \nCosting  N/A \nCost of \nLAMP-\nEC \nNot \nmentio\nned",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues Table 6 with TB-LAMP studies in Cameroonian prisons and refugee villages and in Thailand.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-420",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 420",
      "pageStart": 420,
      "pageEnd": 420,
      "target": "../sections/page-420.md",
      "narrative": null,
      "content": "415\nNliwasa \n2016 \n(83) \nMalawi LAMP \nwas \ndone at \nprimar\ny care \nfacility \nwherea\ns Xpert \nand \nmicros\ncopy in \na \nresearc\nh \nlaborat\nory. \nUSD \nOctobe\nr 2013 \nto \nMarch \n2014 \nPresum\nptive \nTB \ncases \nabove \n15 \nyears  \nTB-\nLAMP \nCulture\n, SSM,  \nXpert \nMTB/RI\nF \nNot \nmentio\nned  \nCosting  N/A \nCost of \nTB-\nLAMP \nLiterat\nure  \nSohn \n2019 \n(80) \nMalawi \nand \nVietna\nm \nperiph\neral \nlaborat\nories  \nUSD \nFebrua\nry to \nApril of \n2015 \nPresum\nptive \nTB \ncases \nTB \nLAMP \nGeneX\npert \nHealth \ncare \nperspe\nctive \nCost \nand \nafforda\nbility \nN/A \nUnit \ncost of \nTB \nLAMP \nand \nPrimar\ny cost \ndata",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues Table 6 with TB-LAMP costing studies in Malawi and Viet Nam.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-421",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 421",
      "pageStart": 421,
      "pageEnd": 421,
      "target": "../sections/page-421.md",
      "narrative": null,
      "content": "416\nanalysi\ns \nXpert \ntest",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A table fragment naming Xpert testing.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-422",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 422",
      "pageStart": 422,
      "pageEnd": 422,
      "target": "../sections/page-422.md",
      "narrative": null,
      "content": "3.1.3. Economic Evidence Around LC-mNAAT (TB-LAMP): \n \nDiagnostic Cost per Sample:  \nAccording to the costing studies, the cost per test ranged from $1 to $19 (all values in 2024 USD currency). \nThe lowest cost of $1 did not provide any details on the cost calculation (81), while the second lowest cost \nof $2 included the cost of equipment, reagents, and consumables in a laboratory setting (82). The study \nintegrated LAMP and Screen-printed graphene electrodes (SPGE). The TB-LAMP-EC components used in \nthe study include LAMP, Redox probe, and SPGE, with various other components such as primers",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that TB-LAMP costs per test ranged from $1 to $19, with the lowest estimates giving little or partial costing detail. It describes a Thai cost-utility study's methods, including direct medical and non-medical costs.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-423",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 423",
      "pageStart": 423,
      "pageEnd": 423,
      "target": "../sections/page-423.md",
      "narrative": null,
      "content": "The conventional method, SSM followed by Culture/DST (Smear, Culture, and Sensitivity Testing), was \nestablished as the baseline for comparison. This approach incurred a cost of $212.90 and resulted in a \nQALY of 24.24. \n \nNext, the study evaluated alternative strategies: \n• \nSSM followed by TB-LAMP and DST costed $237.20 with a QALY of 25.02. The ICER was calculated \nto be $29.93 per additional QALY gained compared to the baseline. \n• \nTB-LAMP followed by DST costed $204.20 and a QALY of 25.17. \n• \nSSM followed by Xpert MTB/RIF incurred a higher cost of $277.60 but yielded a QALY of 24.77. \nH",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports Thai findings that TB-LAMP followed by drug susceptibility testing was cost-saving compared with smear and culture, alongside ICERs for other strategies. It begins describing a Cameroonian cost-benefit analysis of mass screening.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-424",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 424",
      "pageStart": 424,
      "pageEnd": 424,
      "target": "../sections/page-424.md",
      "narrative": null,
      "content": "printer, air conditioners, consumables, maintenance, and human resources. It revealed that combined TB-\nLAMP and Xpert MTB/RIF significantly reduced turnaround time (TAT) by 73.23% in prisons and 74.92% in \nvillages. Moreover, the algorithm demonstrated cost benefits from the second year of implementation: \nabout $16,986 and $18,654 were saved for years 2 and 3, respectively, in prison (15,699 prisoners) and \n$1879 and $2068 were saved for years 2 and 3, respectively, in the villages (30,611villagers). \n \nSohn et al. (2019) conducted a cost and cost-affordability analysis of TB-LAMP and Xpert",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that combining TB-LAMP and Xpert in Cameroon cut turnaround time by about three-quarters and saved money from the second year. It also compares TB-LAMP and GeneXpert budget needs relative to national TB programme budgets in Malawi and Viet Nam.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-425",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 425",
      "pageStart": 425,
      "pageEnd": 425,
      "target": "../sections/page-425.md",
      "narrative": null,
      "content": "420\nTable 7: Cost of TB LAMP \nAuthor \nIncluded cost  \nUnit test cost (USD) \n \nEquipment \nConsumables  \nStaff \nOverhead \nEquipment \nwarranty \nNliwasa et al. \n2016 (83) \nYes \nYes \nYes \nYes \nYes \nMalawi \n2 tests per batch \n $13 \n4 tests per batch \n$10 \n10 tests per batch \n$8 \n14 tests per batch \n$8 \nSohn et al. \n2019 (80) \nYes \nYes \nYes \nYes \nYes \nMalawi \n   \n2 tests per \nbatch \n$15 \n4 tests per \nbatch \n$11 \n10 tests per \nbatch \n$8 \n14 tests per \nbatch \n$8 \nVietnam \n \n2 tests per batch \n$19 \n4 tests per batch \n$14 \n10 tests per \nbatch \n$11 \n14 tests per \nbatch \n$11",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 7 gives TB-LAMP unit costs in Malawi and Viet Nam by batch size, falling from $13–19 for two tests to $8–11 for 14 tests.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-426",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 426",
      "pageStart": 426,
      "pageEnd": 426,
      "target": "../sections/page-426.md",
      "narrative": null,
      "content": "421\nBaikunje et al.  \n2019 (81) \nNo \nYes \nNo \nNo \nNo \n$1 \nJaroenram et \nal. 2020 (82) \nYes \nYes \nNo \nNo \nNo \n$2 \n \nTable 8: Cost of TB LAMP and other algorithms from Chitipim et al. 2022(41) \n  \nCost ($USD) \nQALY \nICER per QALY gained \nSSM followed by   Culture/DST \n$212.9 \n24.24 \nComparator \nSSM followed by TB-LAMP if \npositive DST \n$237.2 \n25.02 \n$29.93 \nTB-LAMP followed by DST  \n$204.2 \n25.17 \nCost saving \nSSM followed by Xpert MTB/RIF \n$277.6 \n24.77 \n$118.7 \nOnly Xpert MTB/RIF  \n$218.1 \n25.07 \n$5.94",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes Table 7 and presents Table 8 on costs, QALYs and ICERs of TB-LAMP and Xpert algorithms from the Thai study.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-427",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 427",
      "pageStart": 427,
      "pageEnd": 427,
      "target": "../sections/page-427.md",
      "narrative": null,
      "content": "3.2.4. Quality of included studies: \nOur quality of health economic studies extraction form was used to evaluate the quality of included \nmanuscripts. This contained 18 questions and was adapted from a modified CHEERS, as described earlier. \nQuality scores varied from 11/18 to 17/18, with an average quality score of 14/21 among.  \n \n3.3 Systematic review of MC-aNAAT \n \n3.3.1. Summary of Findings: \nWe identified 108 records from database searches, 26 were duplicates, leaving 82 for screening. Of these, \n4 underwent full text review. Only one study was included in the final systematic review (84",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports quality scores of the TB-LAMP studies and that only one study was included in the review of moderate-complexity automated NAATs.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p428-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 428,
      "pageEnd": 428,
      "target": "../images/img-p428-1.webp",
      "narrative": "draft",
      "content": "PRISMA flow diagram (Figure 3) of the MC-aNAAT economic evaluation review in three stages. Identification: 123 studies from databases (Embase 60, PubMed 33, Web of Science 15, Scopus 15), with 31 references removed. Screening: 82 studies screened and 78 excluded (66 not an economic evaluation, 12 not on MC-aNAAT); 4 sought for retrieval and 4 assessed for eligibility, of which 3 were excluded because they did not evaluate cost-effectiveness of MC-aNAAT. Included: 1 study.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-428",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 428",
      "pageStart": 428,
      "pageEnd": 428,
      "target": "../sections/page-428.md",
      "narrative": null,
      "content": "3.3.2. General characteristics of included full economic evaluation studies  \nThe systematic review included a study conducted by Malhotra et al. in 2023, which estimated the costs \nassociated with two moderate complexities automated NAATs with low throughput (LT) and high \nthroughput (HT) systems (85). The selected study was conducted in a hypothetical laboratory in resource-\nlimited setting in a low- or middle-income country (LMIC). Detail on the characteristic of included study is \nshown on Table 9. Here LT system is defined as platforms with the capacity to test up to 24 batch per \nsamples",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the single included study, which estimated costs of low- and high-throughput moderate-complexity NAAT systems in a hypothetical laboratory in a low- or middle-income country.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-429",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 429",
      "pageStart": 429,
      "pageEnd": 429,
      "target": "../sections/page-429.md",
      "narrative": null,
      "content": "equipment, test kit (consumables), staff, building, and overhead costs captured in 2019 US dollars.  The \nsources of costing information included time-and-motion (TAM) studies conducted by Foundation for \nInnovative New Diagnostics (FIND), manufacturer’s data, and certain assumptions derived from earlier \nFIND studies in LMICs. \n  \n3.3.3. Cost Estimates and sensitivity analysis: \nIn the base case the per-test cost of was $22.48 (range: $16.74 - $49.40) for LT NAAT and $18.66 (range: \n$11.66 - $45.40) for HT NAAT after adjusting for inflation (Table 10). In this study the cost of equipment \nwas",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains the study's cost components and data sources and reports base-case costs per test for low- and high-throughput systems. It notes a quality score of 17 out of 18.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-430",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 430",
      "pageStart": 430,
      "pageEnd": 430,
      "target": "../sections/page-430.md",
      "narrative": null,
      "content": "425\nTable 9: Characteristics of included study on systematic revise of MC-aNAAT \n \n \nManuscript \nCountry \nsetting \nSetting \nCurrency \nYear of \ncost \nvaluation \nStudy \npopulation \nDiagnostic \nstrategies \nReference \ndiagnostic \nstrategies \nAnalysis \nperspective \nType of \neconomic \nevaluation \nType of \nmodel if \napplicable \nPrimary \noutcome \nSecondary \noutcome(s) \nSource of \ncosting \nMalhotra et \nal., 2023 \n(85) \nHypothetical \nresource \nlimited setting \nin LMIC \nLaboratory \nsetting \nUS$ \n2019 \nTB \nindividuals \nLT MC-NAAT \n(< 24 tests \nper run) \n \nHT MC-NAAT \n(up to 90+ \ntests per \nrun) \nNon \nHeal",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 9 describes the included moderate-complexity NAAT study, and Table 10 gives its unit test costs of $22.48 for low-throughput and $18.66 for high-throughput systems.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-431",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 431",
      "pageStart": 431,
      "pageEnd": 431,
      "target": "../sections/page-431.md",
      "narrative": null,
      "content": "4. \nDISCUSSION \n \nLC-aNAAT discussion: \nThrough a systematic review of published literature on LC aNAATs, we identified 29 studies that met our \ninclusion criteria. These included seventeen cost-effectiveness analyses and twelve cost analyses. Of the \nseventeen cost-effectiveness analyses, ten were on Xpert, six on Ultra and one on TrueNat. All twelve cost \nanalyses were on Xpert. We did not find any cost analyses for Ultra or TrueNat.  \nThe studies included in our review were diverse, conducted across various settings and covered all income \nlevels. This broad spectrum of research provides a",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Discusses the 29 LC-aNAAT studies, including 17 cost-effectiveness analyses and 12 cost analyses, all of the latter on Xpert. It contrasts this with an earlier review that found no Ultra studies and notes that only one Truenat study exists.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-432",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 432",
      "pageStart": 432,
      "pageEnd": 432,
      "target": "../sections/page-432.md",
      "narrative": null,
      "content": "LC-nNAAT discussion: \n \nThe present systematic review provides an in-depth analysis of the economic and cost-effectiveness of \nimplementing TB-LAMP in diverse settings. Through a range of economic analyses, including cost-utility, \ncost-benefit, and cost-affordability assessments, this study contributes valuable insights into the potential \nrole of TB-LAMP in tuberculosis diagnostics.  \n \nThe reviewed studies reveal that factors such as batch scenarios and larger settings influence the per-test \ncosts, with the cost per test decreasing in specific scenarios. Testing volumes, location, and oper",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Discusses TB-LAMP evidence, noting that batch size and setting affect costs. It adds that a recent fall in TB-LAMP's price may improve its cost-effectiveness and affordability.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-433",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 433",
      "pageStart": 433,
      "pageEnd": 433,
      "target": "../sections/page-433.md",
      "narrative": null,
      "content": "In conclusion, the collective findings suggest that TB-LAMP holds promise as a cost-effective and efficient \ndiagnostic tool for tuberculosis when integrated into broader diagnostic algorithms, particularly in \nresource-constrained settings. The results of this study have important implications for policymakers and \nhealthcare professionals involved in the diagnosis and management of tuberculosis. \n \nMC-aNAAT discussion: \nThis systematic review is the first of its kind to investigate the economic aspects of using a new technology \ncalled MC-aNAATS for diagnosing drug-resistant tuberculosis (DR",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Concludes that TB-LAMP shows promise within diagnostic algorithms in resource-limited settings and that evidence on moderate-complexity NAATs is limited to one study. It begins the overall conclusion that Xpert was generally cost-effective compared with smear or culture.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-434",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 434",
      "pageStart": 434,
      "pageEnd": 434,
      "target": "../sections/page-434.md",
      "narrative": null,
      "content": "heterogeneity in the cost-effectiveness results regarding Ultra, it was generally cost-effective compared \nto sputum smear microscopy.  TrueNat cost-effectiveness was reported in only one study but was found \nto be cost-effective compared to Xpert amongst children in India. In general LC aNAATs are likely to be \ncost-effective across a variety of settings when compared with SSM and/or culture. \n \nFew studies assessed parallel testing but when assessed parallel testing was the preferred strategy among \nPLHIV in South Africa using Ultra & LAM or Xpert & LAM compared with stand-alone testing. Fur",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Concludes that Ultra was generally cost-effective compared with smear and that Truenat was cost-effective in one study, so LC-aNAATs are likely cost-effective in many settings. The reference list then begins.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-435",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 435",
      "pageStart": 435,
      "pageEnd": 435,
      "target": "../sections/page-435.md",
      "narrative": null,
      "content": "10. \nGarcía JI, Allué-Guardia A, Tampi RP, Restrepo BI, Torrelles JB. New Developments and Insights in the \nImprovement of Mycobacterium tuberculosis Vaccines and Diagnostics Within the End TB Strategy. Curr \nEpidemiol Rep. 2021 Apr 7;8(2):33–45.  \n11. \nWorld Health Organization. Update on the use of nucleic acid amplification tests to detect TB and drug \nresistant TB: rapid communication. Geneva ; 2021.  \n12. \nNandlal L, Perumal R, Naidoo K. Rapid Molecular Assays for the Diagnosis of Drug-Resistant Tuberculosis. \nInfect Drug Resist. 2022 Aug;Volume 15:4971–84.  \n13. \nWorld Health Organizatio",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Part of the reference list, citing TB diagnostics research, WHO guidance and economic evaluations of Xpert in various countries.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-436",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 436",
      "pageStart": 436,
      "pageEnd": 436,
      "target": "../sections/page-436.md",
      "narrative": null,
      "content": "Available from: https://ejhm.journals.ekb.eg/ \n28. \nHickey AJ, Cummings MJ, Zafari Z, Louh IK, Li J, O’Donnell MR. Evaluation of screening strategies for \npulmonary tuberculosis among hospitalized patients in a low-burden setting: Cost-effectiveness of \nGeneXpert MTB/RIF compared to smear microscopy. Infect Control Hosp Epidemiol. 2022 Jul \n21;43(7):892–7.  \n29. \nCadavid C, Realpe T, Mejía GI, Zapata E, Hernández M, Robledo J. Contribución del uso de XPERT \nMTB/RIF y su costo-efectividad en el diagnóstico de tuberculosis pulmonar y la resistencia a rifampicina: \nuna comparación con métodos dia",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list with economic evaluations of Xpert and Ultra in countries such as the US, Colombia and Uganda.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-437",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 437",
      "pageStart": 437,
      "pageEnd": 437,
      "target": "../sections/page-437.md",
      "narrative": null,
      "content": "BMJ Open. 2022 Aug 23;12(8):e059823.  \n43. \nSilva SCA da, Vater MC, Ramalho DM de P, Almeida IN de, Miranda SS de, Kritski A. Cost-effectiveness of \nXpert®MTB/RIF in the diagnosis of tuberculosis: pragmatic study. Rev Soc Bras Med Trop. 2021;54.  \n44. \nKaso AW, Hailu A. Costs and cost-effectiveness of Gene Xpert compared to smear microscopy for the \ndiagnosis of pulmonary tuberculosis using real-world data from Arsi zone, Ethiopia. PLoS One. 2021 Oct \n25;16(10):e0259056.  \n45. \nKeim-Malpass J, Heysell SK, Thomas TA, Lobo JM, Mpagama SG, Muzoora C, et al. Decision Analytic \nModeling for Global",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list with cost and cost-effectiveness studies of Xpert and other TB tests from Brazil, Ethiopia, India and elsewhere.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-438",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 438",
      "pageStart": 438,
      "pageEnd": 438,
      "target": "../sections/page-438.md",
      "narrative": null,
      "content": "costs incurred under the National Tuberculosis Elimination Programme: A costing study from Tamil \nNadu, South India. Int Health. 2021 Nov 1;13(6):536–44.  \n61. \nNadjib M, Dewi RK, Setiawan E, Miko TY, Putri S, Hadisoemarto PF, et al. Cost and affordability of scaling \nup tuberculosis diagnosis using Xpert MTB/RIF testing in West Java, Indonesia. PLoS One. 2022 Mar \n1;17(3 March).  \n62. \nSantos ADS, De Oliveira RD, Lemos EF, Lima F, Cohen T, Cords O, et al. Yield, Efficiency, and Costs of \nMass Screening Algorithms for Tuberculosis in Brazilian Prisons. Clinical Infectious Diseases. 2021 Mar \n1",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Part of a reference list for a review of costs of tuberculosis nucleic acid amplification tests (NAATs), citing costing and affordability studies of Xpert MTB/RIF, line probe assays and screening algorithms in countries such as India, Indonesia, Brazil and Uganda.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-439",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 439",
      "pageStart": 439,
      "pageEnd": 439,
      "target": "../sections/page-439.md",
      "narrative": null,
      "content": "diagnosis of presumptive multidrug-resistant tuberculosis patients. J Glob Infect Dis. 2022;14(3):93.  \n78. \nJo Y, Kagujje M, Johnson K, Dowdy D, Hangoma P, Chiliukutu L, et al. Costs and cost-effectiveness of a \ncomprehensive tuberculosis case finding strategy in Zambia. PLoS One. 2021 Sep 9;16(9):e0256531.  \n79. \nDonkeng-Donfack VF, Tchatchueng-Mbougua JB, Abanda NN, Ongboulal SM, Djieugoue YJ, Kamdem \nSimo Y, et al. A cost–benefit algorithm for rapid diagnosis of tuberculosis and rifampicin resistance \ndetection during mass screening campaigns. BMC Infect Dis. 2022 Dec 4;22(1):219.  \n80. \nS",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The end of the reference list for the TB NAAT costing review, citing cost and cost-effectiveness studies of case finding, TB-LAMP, Xpert and moderate-complexity drug susceptibility testing.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-440",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 440",
      "pageStart": 440,
      "pageEnd": 440,
      "target": "../sections/page-440.md",
      "narrative": null,
      "content": "435\nAnnex \nAnnex Table 1. Quality assessment (Part 1) \nCitation \nIs the \nstudy \npopulati\non \nclearly \ndescribe\nd? \nAre \ncompetin\ng \nalternativ\nes clearly \ndescribed\n?  \nIs a well-\ndefined \nresearch \nquestion \nposed in \nanswerab\nle form?  \nIs the \neconomic \nstudy \ndesign \nappropria\nte to the \nstated \nobjective\n?  \nIs the \nchosen time \nhorizon \nappropriate \nto include \nrelevant \ncosts and \nconsequenc\nes?  \nIs the \nchosen \nperspectiv\ne \nappropriat\ne?  \nAre all \nimportan\nt and \nrelevant \ncosts for \neach \nalternati\nve \nidentifie\nd?  \nAre all \ncosts \nmeasured \nappropriat\nely in \nphysical \nunits?  \nA",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The start of an annex table assessing the quality of economic studies of TB NAATs, scoring studies of TrueNat and Xpert Ultra against checklist questions on population, comparators, design, time horizon, perspective and costing.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-441",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 441",
      "pageStart": 441,
      "pageEnd": 441,
      "target": "../sections/page-441.md",
      "narrative": null,
      "content": "436\nMafirakure\nva et al., \n2022 \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \nMugwagw\na, \nAbubakar \n& White, \n2021 \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \nThompson \net al., 2023 \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \nPooran et \nal., 2022 \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \nXpert \n  \n  \n  \n  \n  \n  \n  \n  \n  \nCadavid et \nal., 2022 \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \nChitpim et \nal., 2022 \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \nDa Silva et \nal., 2021 \n✓ \n✓ \n✓ \n✓ \nx \n✓ \n✓ \n✓ \n✓ \nEjalu et al., \n2022 \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the quality assessment table (part 1) for economic studies of TB NAATs, listing further Ultra and Xpert studies with mostly positive ratings.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-442",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 442",
      "pageStart": 442,
      "pageEnd": 442,
      "target": "../sections/page-442.md",
      "narrative": null,
      "content": "437\nEsmail et \nal., 2020 \n✓ \n✓ \n✓ \n✓ \nx \n✓ \n✓ \n✓ \n✓ \nHickey et \nal., 2022 \n✓ \n✓ \n✓ \n✓ \nx \nx  \n✓ \n✓ \n✓ \nKaso & \nHailu, 2021 \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \nKeim-\nMalpass et \nal., 2023 \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \nKhumsri et \nal., 2020 \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \nSedrak et \nal., 2022 \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the quality assessment table (part 1) for Xpert economic studies, with a few studies falling short on time horizon or perspective.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-443",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 443",
      "pageStart": 443,
      "pageEnd": 443,
      "target": "../sections/page-443.md",
      "narrative": null,
      "content": "438\n \nAnnex Table 2. Quality assessment (Part 2) \nCitation \nAre all \nimporta\nnt and \nrelevant \noutcome\ns for \neach \nalternati\nve \nidentifie\nd? \nAre all \noutcomes \nmeasured \nappropriat\nely in \nphysical \nunits? \nAre \noutcomes \nvalued \nappropriate\nly? \nIs an \nincremen\ntal \nanalysis \nof costs \nand \noutcome\ns of \nalternativ\nes \nperforme\nd? \nAre all \nfuture costs \nand \noutcomes \ndiscounted \nappropriate\nly? \nAre all \nimportant \nvariables, \nwhose \nvalues are \nuncertain, \nappropriat\nely \nsubjected \nto \nsensitivity \nanalysis? \nDo the \nconclusio\nns follow \nfrom the \ndata \nreported\n? \nDoes the \nstudy \ndis",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The start of the second part of the quality assessment of TB NAAT economic studies, rating outcomes, incremental analysis, discounting, sensitivity analysis, conclusions, generalizability and conflicts of interest.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-444",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 444",
      "pageStart": 444,
      "pageEnd": 444,
      "target": "../sections/page-444.md",
      "narrative": null,
      "content": "439\nFekadu, \nWang & \nYou, 2023 \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \nMafirakur\neva et al., \n2022 \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \nMugwagw\na, \nAbubakar \n& White, \n2021 \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \nThompson \net al., \n2023 \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \nPooran et \nal., 2022 \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \nXpert  \nCadavid et \nal., 2022 \n✓ \n✓ \n✓ \n✓ \n✓ \nx \n✓ \n✓ \n✓ \nChitpim et \nal., 2022 \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the part 2 quality assessment table for TrueNat, Ultra and Xpert economic studies, with most criteria met.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-445",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 445",
      "pageStart": 445,
      "pageEnd": 445,
      "target": "../sections/page-445.md",
      "narrative": null,
      "content": "440\nDa Silva et \nal., 2021 \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \nEjalu et \nal., 2022 \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \nEsmail et \nal., 2020 \n✓ \n✓ \n✓ \nx \n✓ \n✓ \n✓ \n✓ \n✓ \nHickey et \nal., 2022 \n✓ \n✓ \n✓ \n✓ \nx \n✓ \n✓ \n✓ \n✓ \nKaso & \nHailu, \n2021 \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \nKeim-\nMalpass et \nal., 2023 \n✓ \n✓ \n✓ \n✓ \nx  \n✓ \n✓ \n✓ \n✓ \nKhumsri et \nal., 2020 \n✓ \n✓ \n✓ \n✓ \nx  \n✓ \n✓ \n✓  \n✓ \nSedrak et \nal., 2022 \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n✓ \n \n \nAnnex Table 3. Quality assessment LC-mNAAT",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the part 2 quality assessment table for Xpert economic studies and introduces a quality assessment table for low-complexity manual NAAT studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-446",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 446",
      "pageStart": 446,
      "pageEnd": 446,
      "target": "../sections/page-446.md",
      "narrative": null,
      "content": "441\n \n  \nStudy \npopulation \nCompeting \nalternatives \nResearch \nquestion \nStudy design \nTime horizon \nStudy \nperspective \nRelevant costs \nCosts measured \nCosts valued \nDiscounting \nSensitivity \nanalysis \nConclusions \nGeneralizability \nConflict of \ninterest \nTotal score \nNliwasa 2016 \nYes \nYes \nYes \nYes \nN/A \nYes \nYes \nYes \nYes \nNo \nNo \nYes \nYes \nYes \n85% \nBaikunje 2019 \n  \n  \n  \nNo \nN/A \n  \nNo \nNo \nNo \nNo \nNo \nYes \nNo \nYes \n46% \nSohn 2019 \nYes \nYes \nYes \nYes \nYes \nYes \nYes \nYes \nYes \nYes \nYes \nYes \nYes \nYes \n100% \nJaroenram 2020 \nYes \nYes \nYes \nYes \nN/A \nYes \nNo \nYes \nYes \nNo \nNo \nYes \nNo \nYes",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Quality assessment tables for economic studies of low-complexity manual NAATs and moderate-complexity automated NAATs, giving yes/no ratings per criterion and total scores ranging from 46% to 100%.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-447",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 447",
      "pageStart": 447,
      "pageEnd": 447,
      "target": "../sections/page-447.md",
      "narrative": null,
      "content": "442 \n \nAnnex 5: MC-aNAAT systematic review search strategy \n \nConcept #1 \n\"tuberculosis\"[MeSH Terms] OR \"tuberculosis\"[Title/Abstract] OR \"TB\"[Title/Abstract] OR \n\"tuberculous\"[Title/Abstract] OR \"mycobacterium tuberculosis\"[MeSH Terms] \n \nConcept #2 \n\"Abbott RealTime MTB\"[All Fields] OR \"Abbott RealTime MTB RIF/INH\"[All Fields] OR \"FluoroType \nMTBDR\"[All Fields] OR \"FluoroType MTB\"[All Fields] OR \"Hain Lifescience\"[All Fields] OR \"BD MAX MDR-\nTB\"[All Fields] OR \"Becton Dickinson\"[All Fields] OR \"cobas MTB\"[All Fields] OR \"cobas MTB-RIF/INH\"[All \nFields] OR \"Roche\"[All Fields] OR \"Moderate com",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The literature search strategy for the systematic review of costs of moderate-complexity automated TB NAATs, combining terms for tuberculosis, named test products and manufacturers, and cost analysis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-448",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 448",
      "pageStart": 448,
      "pageEnd": 448,
      "target": "../sections/page-448.md",
      "narrative": null,
      "content": "443 \n \nWeb Annex B.10. Nucleic acid amplification tests (NAATs) for \ntuberculosis and tuberculosis drug resistance: A qualitative \nevidence synthesis of user perspectives \n \nDate: 07 April 2024 \n \n \nTo:   Dr. Alexei Korobitsyn \n         World Health Organization (WHO)  \n         Avenue Appia 20 \n         1211 Geneva \n         Switzerland \n         korobitsyna@who.int \n \n \n \nNames, affiliations, and address of the contact person: \n         Eleanor Ochodo \n         Kenya Medical Research Institute, Center for Global Health Research, Kisumu, Kenya \n         Stellenbosch University, Center for Evi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The cover note of Web Annex B.10, a qualitative evidence synthesis of user perspectives on TB NAATs dated April 2024, giving WHO and author contact details.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-449",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 449",
      "pageStart": 449,
      "pageEnd": 449,
      "target": "../sections/page-449.md",
      "narrative": null,
      "content": "444 \n \nNucleic acid amplification tests (NAATs) for tuberculosis \nand tuberculosis drug resistance: A qualitative evidence \nsynthesis of user perspectives \n \n \nOchodo EA1, 2, Mulaku NM1, 2, 3, Karanja PW4, Engel N5, \n1. Centre for Global Health Research, Kenya Medical Research Institute, Kisumu, Kenya \n2. Centre for Evidence-Based Health Care, Department of Global Health, Faculty of Medicine & \nHealth Sciences, Stellenbosch University, Cape Town, South Africa \n3. Department of Pharmacology, Clinical Pharmacy and Pharmacy Practice, Faculty of Health \nSciences, University of Nairobi, Nairobi, Ke",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title page of the qualitative evidence synthesis on user perspectives of NAATs for TB and drug-resistant TB, listing authors from Kenya, South Africa and the Netherlands and noting it was prepared for a WHO guideline meeting in May 2024.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-450",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 450",
      "pageStart": 450,
      "pageEnd": 450,
      "target": "../sections/page-450.md",
      "narrative": null,
      "content": "445 \n \nList of Abbreviations \n \nCB-NAAT      Cartridge Based- Nucleic Acid Amplification Test \nCENTRAL \nCochrane Central Register of Controlled Trials \nCERQual \nConfidence in the Evidence from Reviews of Qualitative research \nDR-TB            Drug Resistant Tuberculosis \nEPPI \nEvidence for Policy and Practice Information and coordinating Centre tool  \nENTREQ        Enhancing transparency in reporting the synthesis of qualitative research \nGRADE \nGrading of Recommendations, Assessment, Development, and Evaluations \nGDT \nGuideline Development Tool \nLAM \nUrine Lipoarabinomannan \nLAMP \nLoop-mediat",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A list of abbreviations used in the qualitative evidence synthesis on TB NAATs, covering test types, databases and review methods such as GRADE-CERQual and PRISMA.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-451",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 451",
      "pageStart": 451,
      "pageEnd": 451,
      "target": "../sections/page-451.md",
      "narrative": null,
      "content": "446 \n \nExecutive Summary \n \nBackground \nThe WHO has introduced different classes of Nucleic Acid Amplification Tests (NAATs) depending on their \ninfrastructure needs and mode of conduct. Latter NAATs are replacing standard Xpert MTB/RIF as initial \ndiagnostic tests for Tuberculosis (TB) due to their high accuracy, rapid results and impact on patient \nimportant outcomes. Evidence on perspectives on the use, feasibility, acceptability and equity of newer \nNAATs in particular will inform an update to the WHO consolidated guidelines on TB diagnosis. \nObjectives \nTo synthesize provider and end user",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Opens the executive summary of the qualitative synthesis, explaining that newer NAATs are replacing standard Xpert MTB/RIF as initial TB tests and that evidence on their feasibility, acceptability and equity will inform WHO guidelines. It states the objectives and notes that 49 qualitative studies were included, 17 from an updated search.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-452",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 452",
      "pageStart": 452,
      "pageEnd": 452,
      "target": "../sections/page-452.md",
      "narrative": null,
      "content": "447 \n \nsettings in Africa, Asia and Eastern Europe. Two studies were on perspectives about Xpert Ultra and the \nrest about Xpert MTB/RIF or rapid NAATs. The studies about Xpert Ultra were conducted in Africa and \nEastern Europe and focused on all people with presumptive TB, drug resistant TB (DR-TB) and \nextrapulmonary TB (EPTB). All studies focused on perspectives about single tests and neither covered \nparallel testing nor moderate complexity NAATs. \nWe identified 19 review findings from all included NAATs, nine of which were also about Xpert Ultra.  We \ngrouped the findings into three overa",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the executive summary, noting that included studies came from high TB burden settings in Africa, Asia and Eastern Europe, mostly on Xpert MTB/RIF, with two on Xpert Ultra and none on parallel testing or moderate-complexity NAATs. It summarizes 19 review findings grouped into what users value, related challenges and equity concerns, including Ultra's speed and accuracy but also infrastructure problems and fear of false results.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-453",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 453",
      "pageStart": 453,
      "pageEnd": 453,
      "target": "../sections/page-453.md",
      "narrative": null,
      "content": "448 \n \nConcerns for access and equity \nFindings about Xpert MTB/RIF demonstrated concerns about sustainable funding, restrictions by donors, \nmaintenance of the testing platforms, poor referral systems and dependence on outreach workers limiting \naccess. Also protracted diagnostic delays, test underutilization, unavailability of community TB diagnostic \nfacilities and too many eligibility restrictions limited access to testing. \nSimilar to Xpert MTB/RIF, the reported concerns about equity for Xpert Ultra were its limited availability \nin health facilities, as well as high costs incurred by pat",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarizes access and equity concerns in the executive summary, such as funding, platform maintenance, referral weaknesses and costs. It concludes that qualitative evidence on newer NAATs is limited, that challenges mirror those of Xpert MTB/RIF, and recommends studies on parallel testing.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-454",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 454",
      "pageStart": 454,
      "pageEnd": 454,
      "target": "../sections/page-454.md",
      "narrative": null,
      "content": "449 \n \nBackground \nTuberculosis (TB) continues to be among the world’s top causes of mortality while drug resistant TB (DR-\nTB) is the leading cause of death due to anti-microbial resistant disease and poses a global security risk.1 \n2 According to the WHO Global Tuberculosis Report published in 2023,1 about 10.3 million people \ndeveloped TB and 410,000 people developed multidrug rifampicin resistant TB (MDR-TB/RR-TB) globally. \nThere were an estimated 1.3 million deaths attributed to TB. Deaths due to TB were higher in HIV-negative \npeople (1.16 million deaths) compared to HIV-positive people",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Sets out the background to the review, citing the 2023 Global TB Report figures of 10.3 million cases and 1.3 million deaths and the need for prompt, accurate diagnosis. It describes WHO's classes of NAATs and introduces parallel testing as an alternative to sequential algorithms.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-455",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 455",
      "pageStart": 455,
      "pageEnd": 455,
      "target": "../sections/page-455.md",
      "narrative": null,
      "content": "450 \n \ndemonstrated increased sensitivities. For example, Fan and colleagues 8 showed that parallel test \nalgorithms using culture plus Xpert MTB/RIF plus Simultaneous Amplification and Testing method for TB \n(SAT-TB), culture plus Xpert, or culture plus SAT-TB achieved higher sensitivities compared with all three \nsingle test models. A cross sectional study by Esmail and colleagues 9  evaluating five different diagnostic \nstrategies (Xpert alone, Urine Lipoarabinomannan (LAM) alone, sequential testing with Xpert and LAM \nand parallel testing with Xpert and LAM) also concluded that in sputum-e",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives examples of studies where parallel testing, such as Xpert with LAM or culture, increased sensitivity, and notes End TB Strategy targets. It explains that the review was updated to cover newer NAAT classes and parallel testing in children and people with HIV ahead of the WHO guideline meeting.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-456",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 456",
      "pageStart": 456,
      "pageEnd": 456,
      "target": "../sections/page-456.md",
      "narrative": null,
      "content": "451 \n \n \nReview objectives  \n• \nTo synthesize provider and end user perspectives on low complexity and moderate complexity \nNAATs for diagnosis of tuberculosis and tuberculosis drug resistance (DR-TB). \n• \nTo synthesize provider and end user perspectives on parallel use of NAATs or of different \nspecimens on NAATs in children and PLHIV. \n \nMethods \nThe methods for study inclusion, data collection and analysis were based on the Cochrane Guidance for \nQualitative Research.12 We built on the methodology of our previously published Cochrane qualitative \nreview on user perspectives of low complexit",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "States the review objectives and explains that methods followed Cochrane guidance for qualitative research, building on an earlier Cochrane review of low-complexity NAATs. It sets out inclusion criteria for qualitative and mixed-methods studies and the types of publications excluded.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-457",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 457",
      "pageStart": 457,
      "pageEnd": 457,
      "target": "../sections/page-457.md",
      "narrative": null,
      "content": "452 \n \ncommentaries, or non-research editorials.  \nPopulation  \nWe included studies about users or potential users of low and moderate complexity NAATs. Users at \nrecipient level were children, adolescents, adults and people living with HIV with signs and symptoms of \nTB and/or screened positive for TB or their caregivers. Users at provider level were health care providers \n(clinicians, nurses, laboratory technicians), TB programme officers or managers or other staff involved in \nordering, maintaining or acting on results of TB diagnostic tests. We did not limit inclusion by race, \nsocioeconom",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Defines the review's population of test recipients and providers and the interventions covered, describing classes of low-complexity manual and automated NAATs and the products in each.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-458",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 458",
      "pageStart": 458,
      "pageEnd": 458,
      "target": "../sections/page-458.md",
      "narrative": null,
      "content": "453 \n \nrifampicin, isoniazid and 2nd line anti-TB \nagents) (LC-aNAATs)* \nModerate complexity automated NAATs for \nthe detection of MTB-complex (MTBC) and its \nresistance to the anti-tuberculosis (TB) drugs \nfluoroquinolones \n(FLQ; \nlevofloxacin \nand \nmoxifloxacin), amikacin (AMK), ethambutol \n(EMB), and linezolid (LZD)) (MC-aNAATs) \nThe LiquidArray MTB-XDR VER 1.0. \n*Xpert MTB/RIF belongs to the LC-aNAAT class but has been superseded by the new tests presented in the table. \nTarget condition \nDrug susceptible or Drug-resistant Pulmonary or Extrapulmonary TB. \nOutcome measures \nViews, perceptio",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes moderate-complexity automated NAATs for detecting resistance to second-line drugs, the target conditions and outcome measures on user views. It notes reporting standards and that literature searches were run in October 2023.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-459",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 459",
      "pageStart": 459,
      "pageEnd": 459,
      "target": "../sections/page-459.md",
      "narrative": null,
      "content": "454 \n \nusing the search terms described in Appendix 1: Cochrane Central Register of Controlled Trials (CENTRAL), \nincluded in the Cochrane Library, issue 10, October 2023; MEDLINE (Ovid); Embase (Ovid); Social Science \nCitation Index (ISI Web of Knowledge); WHO Global Index Medicus; and CINAHL (EBSCOhost). Though this \nwas an update, we did not limit the search date from the date of the last search in our previous Cochrane \nreview on low complexity NAATs. This was done to avoid missing studies on parallel testing with NAATs \nand moderate complexity NAATs. \n \nSearching other resources \nThe auth",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists the databases searched for the qualitative synthesis and explains that no date limit was applied, plus reference checking. It describes how studies were selected and sampled, including Xpert MTB/RIF studies as supporting evidence when direct evidence on newer tests was lacking.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-460",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 460",
      "pageStart": 460,
      "pageEnd": 460,
      "target": "../sections/page-460.md",
      "narrative": null,
      "content": "455 \n \nThree review authors (EO, MM and PK) independently performed data extraction in duplicate using the \nCovidence platform.17 We resolved any disagreements by discussion and consensus. The platform for data \nextraction was designed and piloted on at least two studies before use. We extracted data findings in a \nnarrative form including themes and supporting quotations. \n \nWe extracted data about: \nStudy characteristics:  the first author, publication year, journal, language, participant group (type of and \ncadre of user), setting (country, rural/urban, type of health facility). \n \nInterven",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes duplicate data extraction in Covidence, the study, intervention and finding details extracted, and the approach to appraising methodological limitations and ethics reporting.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-461",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 461",
      "pageStart": 461,
      "pageEnd": 461,
      "target": "../sections/page-461.md",
      "narrative": null,
      "content": "456 \n \nwork, we ensured that participants could not be recognized by readers. \n \nWe extracted the following information across the five domains: \n \n Rigor in sampling: \n• \nWas the sampling strategy appropriate to the questions posed in the study? (e.g. was the \nstrategy well-reasoned and justified?) \n• \nWere attempts made to obtain a diverse sample of the population in question?  (Considering \npopulations excluded and those with different perspectives) \n• \nWere characteristics of the sample critical to the understanding of the study context and \nfindings presented (for example participants in",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists the appraisal questions used to assess rigour in sampling, data collection and analysis in the included qualitative studies, with attention to participant anonymity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-462",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 462",
      "pageStart": 462,
      "pageEnd": 462,
      "target": "../sections/page-462.md",
      "narrative": null,
      "content": "457 \n \npreliminary results to participants, asking a colleague to review the data, or reflexivity; in \nquantitative research this may be done by, for example, significance testing). \n \nExtent to which findings are grounded in/supported by the data: \n• \nWere enough data presented to show how the authors arrived at their findings? \n• \nWere the data presented fit the interpretation/support claims about patterns in data? \n• \nWere the data presented to illustrate the findings? \n• \nFor qualitative studies, were quotes numbered or otherwise identified and the reader could see that \nthey did not just",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the appraisal criteria, covering how well findings are grounded in data and the breadth and depth of findings. It introduces the GRADE-CERQual approach used to rate confidence in each review finding.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-463",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 463",
      "pageStart": 463,
      "pageEnd": 463,
      "target": "../sections/page-463.md",
      "narrative": null,
      "content": "458 \n \nassesses confidence in the review evidence on four main components including methodological \nlimitations of included studies, coherence of review findings, adequacy of data contributing to review \nfindings and relevance of the included studies to the review question.  \n \nWe rated our confidence of the evidence as either high (not downgraded), moderate, low, or very low for \neach of the four domains and made a judgment about the overall confidence in the evidence supporting \nthe review finding.26 Our final assessment was based on consensus among the review authors.  We \nsummarized this i",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains how GRADE-CERQual confidence ratings were made across four components and summarized in tables. It describes the thematic synthesis method and the reviewers' steps to manage reflexivity and bias.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-464",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 464",
      "pageStart": 464,
      "pageEnd": 464,
      "target": "../sections/page-464.md",
      "narrative": null,
      "content": "459 \n \n \nEO is a public health physician and methodologist. She has 12 years' experience in evidence synthesis \nspecializing in methodology, systematic reviews, and meta-analysis of diagnostic tests. She has conducted \ndiagnostic test accuracy and qualitative systematic reviews on tuberculosis tests, some of which have \ninformed WHO guidelines on tuberculosis tests. She is also an academic editor with the Cochrane \nInfectious Disease Group. \n \nPK has experience with qualitative evidence synthesis for tuberculosis diagnostics. She has co-authored a \nqualitative review of diagnostic tests which",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives reflexivity statements describing the review authors' backgrounds and potential influences on their views. It begins the results, reporting 49 included studies, 17 newly identified and 32 from the earlier Cochrane review.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-465",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 465",
      "pageStart": 465,
      "pageEnd": 465,
      "target": "../sections/page-465.md",
      "narrative": null,
      "content": "460 \n \nstudies were about Xpert Ultra published in 2020 and 2022 and the rest were about Xpert MTB/RIF or \nrapid NAATs published between 2012 and 2023. No qualitative studies on moderate complexity NAATs or \nother newer low complexity NAATs (apart from Xpert Ultra) were identified. An overview of included and \nexcluded studies can be found in Annex 3 and 4 respectively.  \n \nDescription of the studies \n \nStudies about Xpert Ultra \nThese two studies were conducted in Africa (Kenya and Swaziland)30 and Eastern Europe (Belarus),31 \ncountries identified as high burden TB settings. Mwaura et al was",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that only two included studies concerned Xpert Ultra and none addressed moderate-complexity or other newer low-complexity NAATs. It describes the Ultra studies in Kenya, Eswatini and Belarus and the settings and uses of the Xpert MTB/RIF studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-466",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 466",
      "pageStart": 466,
      "pageEnd": 466,
      "target": "../sections/page-466.md",
      "narrative": null,
      "content": "461 \n \nperspectives of single tests and not parallel testing. \n \nMethodological limitations of the studies \nA summary of our assessment for methodological quality can be found in Annex 5. Briefly: \nThe two studies30 31 about Xpert Ultra were overall of very good quality having undertaken several steps \nor fairly thorough attempt towards methodological quality across the assessed components. \nStudies about Xpert MTB/RIF or rapid NAATs in general were of overall good quality with more than half \nof them (n=32) having largely undertaken several steps or fairly thorough attempt towards \nmethodolog",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarizes the methodological quality of included studies, judged very good for the Ultra studies and mostly good for others. It introduces the 19 review findings and describes what providers and laboratory staff valued about Xpert, such as accuracy, speed and low patient costs.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-467",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 467",
      "pageStart": 467,
      "pageEnd": 467,
      "target": "../sections/page-467.md",
      "narrative": null,
      "content": "462 \n \nstaff satisfaction compared to sputum microscopy. People with TB valued an accurate diagnosis, \navoiding diagnostic delays and low costs associated with diagnostic testing. \nCompared to Xpert MTB/RIF, providers valued Xpert Ultra‘s capacity to improve TB case detection \namong these hard-to-diagnose patients (EPTB, pediatric TB and TB patients who are co-infected with \nCOVID-19) and detecting more cases. By building on existing Xpert platforms, its ease of implementation \nand integration with testing for other diseases was valued. \n \n \n \n \nDetailed review findings for Xpert Ultra \nSummar",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes what patients and providers valued about Xpert and Ultra, including Ultra's better detection in hard-to-diagnose groups such as children and extrapulmonary TB. It summarizes challenges limiting these benefits, such as weak systems, outdated guidelines and over-reliance on test results.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-468",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 468",
      "pageStart": 468,
      "pageEnd": 468,
      "target": "../sections/page-468.md",
      "narrative": null,
      "content": "463 \n \nprocesses also led to underutilization. \n \nProviders’ acceptability of Xpert Ultra seemed high but there was uncertainty about its accuracy, \npotentially leading to reduced trust and litigation should there be a false diagnosis. Like Xpert MTB/RIF, \nits implementation could also be hindered by Infrastructural issues such as power outages, staff \nshortages. Limited sputum transportation and limited availability of Xpert testing platforms in health \nfacilities.  \n \nDetailed review findings for Xpert Ultra \nSummary of Qualitative Findings table: Findings 3-7, Annex 6 \n  \nFinding 3: Patient",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes uncertainty among providers about Ultra's accuracy and infrastructure barriers to its use. It presents detailed findings on patient reluctance to test due to stigma, staffing and training gaps, and workload, illustrated with participant quotations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-469",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 469",
      "pageStart": 469,
      "pageEnd": 469,
      "target": "../sections/page-469.md",
      "narrative": null,
      "content": "464 \n \nFinding 7: Health workers accept a measure of uncertainty of tests but fear that false diagnosis can \nstill lead to litigation and diminished trust in test results (Moderate confidence, Mwaura 202030).  \nThe reduced specificity of Xpert Ultra included the risk of litigation from patients for a false diagnosis \nand reduced trust in the healthcare system by sharing through word of mouth, within a patient’s \nnetwork, impacting future diagnostic services. \nConcerns for access and equity  \n \nSummary  \nFindings about Xpert MTB/RIF demonstrated concerns about sustainable funding, restrictions",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents the finding that health workers accept some test uncertainty but fear litigation and loss of trust from false diagnoses with Ultra. It then summarizes access and equity concerns and begins the discussion, restating the 19 findings in three categories.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-470",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 470",
      "pageStart": 470,
      "pageEnd": 470,
      "target": "../sections/page-470.md",
      "narrative": null,
      "content": "465 \n \naccess and equity.  Of these findings, nine also applied to Xpert Ultra. Overall, we had high confidence in \nthe review findings.  \nThe strengths of our review included utilizing a robust study search and selection process and building on \na previously utilized framework for our thematic analysis. Studies were screened and appraised in \nduplicate to minimize bias. We also strengthened the limited evidence on Xpert Ultra by adding evidence \nabout the similar standard Xpert MTB/RIF test.14 \nAlthough Xpert MTB/RIF has been superseded by newer rapid NAATs, most of the available qualitative",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Discusses the review's high overall confidence, its strengths such as duplicate screening, and its main limitation, the scarcity of direct evidence on newer NAATs. It compares findings with other literature and concludes that newer NAATs are valued for accuracy and integration with existing GeneXpert platforms.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-471",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 471",
      "pageStart": 471,
      "pageEnd": 471,
      "target": "../sections/page-471.md",
      "narrative": null,
      "content": "466 \n \nand their ease in integrating with rapid testing for other diseases. Overall, we had high confidence in the \nreview findings. Challenges limiting the realization of these values for latter NAATs are similar to Xpert \nMTB/RIF and comprise weak infrastructure, fragmented systems, heavy workload and limited availability \nof NAATs and their supplies. \nAcknowledgements \nWe acknowledge the efforts of the information specialist, Dr. Vittoria Lutje for designing the search \nstrategies. \nFunding \nThe conduct of this review is funded by WHO. \nCompeting interests \nThe review authors have no confli",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Concludes that challenges for newer NAATs mirror those of Xpert MTB/RIF, followed by acknowledgements, WHO funding and a no-conflicts statement.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-472",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 472",
      "pageStart": 472,
      "pageEnd": 472,
      "target": "../sections/page-472.md",
      "narrative": null,
      "content": "467 \n \nReferences \n1. WHO. Global Tuberculosis Report. Geneva, Switzerland: World Health Organization.2023. Available at \nhttps://www.who.int/teams/global-tuberculosis-programme/tb-reports/global-tuberculosis-\nreport-2023. \n2. WHO. Status Update: Reaching the targets in the political declaration of the United Nations General \nAssembly High-level Meeting on the fight against tuberculosis. September 2023. Geneva, \nSwitzerland: World Health Organization.2023. Available at \nhttps://www.who.int/publications/m/item/WHO-UCN-TB-2023.4. \n3. WHO. The End TB Strategy. Geneva, Switzerland: World Health Or",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The start of the reference list for the qualitative synthesis on TB NAATs, citing WHO TB reports and strategies and Cochrane methods guidance.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-473",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 473",
      "pageStart": 473,
      "pageEnd": 473,
      "target": "../sections/page-473.md",
      "narrative": null,
      "content": "468 \n \n13. Tong A, Flemming K, McInnes E, et al. Enhancing transparency in reporting the synthesis of \nqualitative research: ENTREQ. BMC Med Res Methodol 2012;12:181. doi: 10.1186/1471-2288-\n12-181 \n14. Murad MH, Chang SM, Fiordalisi CV, et al. Improving the utility of evidence synthesis for decision \nmakers in the face of insufficient evidence. J Clin Epidemiol 2021;135:170-75. doi: \n10.1016/j.jclinepi.2021.02.028 \n15. StopTBPartnership. Xpert MTB/RIF and Ultra, Technical Information Note. Geneva: Stop TB \nPartnership, 2019. Available at \nhttps://stoptb.org/assets/documents/gdf/drugsupply/Xpe",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list of the qualitative synthesis, citing qualitative reporting standards, PRISMA and GRADE-CERQual methods papers.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-474",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 474",
      "pageStart": 474,
      "pageEnd": 474,
      "target": "../sections/page-474.md",
      "narrative": null,
      "content": "469 \n \n30. Mwaura M, Kao K, Wambugu J, et al. Situating trade-offs: Stakeholder perspectives on \novertreatment versus missed diagnosis in transition to Xpert MTB/RIF Ultra in Kenya and \nSwaziland. PLoS One 2020;15(2):e0228669. doi: 10.1371/journal.pone.0228669 \n31. Sereda Y, Korotych O, Klimuk D, et al. Tuberculosis Co-Infection Is Common in Patients Requiring \nHospitalization for COVID-19 in Belarus: Mixed-Methods Study. Int J Environ Res Public Health \n2022;19(7) doi: 10.3390/ijerph19074370 \n32. An Y, Teo AKJ, Huot CY, et al. Barriers to childhood tuberculosis case detection and management i",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list of the qualitative synthesis with included studies on Xpert Ultra and Xpert MTB/RIF implementation in countries such as Kenya, Belarus, Cambodia and Uganda.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-475",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 475",
      "pageStart": 475,
      "pageEnd": 475,
      "target": "../sections/page-475.md",
      "narrative": null,
      "content": "470 \n \n46. Oga-Omenka C, Boffa J, Kuye J, et al. Understanding the gaps in DR-TB care cascade in Nigeria: A \nsequential mixed-method study. J Clin Tuberc Other Mycobact Dis 2020;21:100193. doi: \n10.1016/j.jctube.2020.100193 \n47. Creswell J, Codlin AJ, Andre E, et al. Results from early programmatic implementation of Xpert \nMTB/RIF testing in nine countries. BMC Infect Dis 2014;14:2. doi: 10.1186/1471-2334-14-2 \n48. England K, Masini T, Fajardo E. Detecting tuberculosis: rapid tools but slow progress. Public Health \nAction 2019;9(3):80-83. doi: 10.5588/pha.19.0013 \n49. Engel N, Krumeich A. Valu",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list of the qualitative synthesis, citing studies on drug-resistant TB care and Xpert roll-out in countries including Nigeria and India.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-476",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 476",
      "pageStart": 476,
      "pageEnd": 476,
      "target": "../sections/page-476.md",
      "narrative": null,
      "content": "471 \n \n63. Rendell NL, Bekhbat S, Ganbaatar G, et al. Implementation of the Xpert MTB/RIF assay for \ntuberculosis in Mongolia: a qualitative exploration of barriers and enablers. PeerJ 2017;5:e3567. \ndoi: 10.7717/peerj.3567 \n64. Vijayageetha M, Kumar AM, Ramakrishnan J, et al. Tuberculosis screening among pregnant women \nattending a tertiary care hospital in Puducherry, South India: is it worth the effort? Glob Health \nAction 2019;12(1):1564488. doi: 10.1080/16549716.2018.1564488 \n65. Medina-Marino A, de Vos L, Bezuidenhout D, et al. \"I got tested at home, the help came to me\": \nacceptability",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list of the qualitative synthesis, citing implementation studies of molecular TB testing in Mongolia, India, South Africa, Nepal and elsewhere.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-477",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 477",
      "pageStart": 477,
      "pageEnd": 477,
      "target": "../sections/page-477.md",
      "narrative": null,
      "content": "472 \n \n78. Stime KJ, Garrett N, Sookrajh Y, et al. Clinic flow for STI, HIV, and TB patients in an urban infectious \ndisease clinic offering point-of-care testing services in Durban, South Africa. BMC Health Serv \nRes 2018;18(1):363. doi: 10.1186/s12913-018-3154-2 \n79. Adam P, Pai M. Implementation of Xpert((R)) MTB/RIF in high-burden countries: voices from the field \nmatter. Public Health Action 2019;9(3):78-79. doi: 10.5588/pha.19.0055 \n80. Adepoyibi T, Lilis L, Greb H, et al. Which attributes within target product profiles for tuberculosis \ndiagnostics are the most important to focus on? In",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list of the qualitative synthesis, citing studies on clinic flow, Xpert implementation lessons and health system factors in TB management.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-478",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 478",
      "pageStart": 478,
      "pageEnd": 478,
      "target": "../sections/page-478.md",
      "narrative": null,
      "content": "473 \n \n95. Dominique JK, Ortiz-Osorno AA, Fitzgibbon J, et al. Implementation of HIV and Tuberculosis \nDiagnostics: The Importance of Context. Clin Infect Dis 2015;61Suppl 3(Suppl 3):S119-25. doi: \n10.1093/cid/civ552 \n96. Gray CM, Katamba A, Narang P, et al. Feasibility and Operational Performance of Tuberculosis \nDetection by Loop-Mediated Isothermal Amplification Platform in Decentralized Settings: \nResults from a Multicenter Study. J Clin Microbiol 2016;54(8):1984-91. doi: 10.1128/JCM.03036-\n15 \n97. Griffin AM, Caviedes L, Gilman R, et al. Field and laboratory preparedness: challenges to ro",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list of the qualitative synthesis, citing studies on diagnostic implementation, TB-LAMP feasibility and health worker knowledge.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-479",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 479",
      "pageStart": 479,
      "pageEnd": 479,
      "target": "../sections/page-479.md",
      "narrative": null,
      "content": "474 \n \n113. Ntinginya NE, Kuchaka D, Orina F, et al. Unlocking the health system barriers to maximise the \nuptake and utilisation of molecular diagnostics in low-income and middle-income country \nsetting. BMJ Glob Health 2021;6(8) doi: 10.1136/bmjgh-2021-005357 \n114. Ochodo EA, Naidoo S, Schumacher S, et al. Improving the design of studies evaluating the impact of \ndiagnostic tests for tuberculosis on health outcomes: a qualitative study of perspectives of \ndiverse stakeholders. Wellcome Open Res 2019;4:183. doi: 10.12688/wellcomeopenres.15551.1 \n115. Oga-Omenka C, Bada F, Agbaje A, et al. Eas",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list of the qualitative synthesis, citing work on health system barriers to molecular diagnostics and access to drug-resistant TB services.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-480",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 480",
      "pageStart": 480,
      "pageEnd": 480,
      "target": "../sections/page-480.md",
      "narrative": null,
      "content": "475 \n \nAnnexes \nAnnex 1: Search strategies \nSearch methods for identification of studies \nElectronic searches \nThe Cochrane Infectious Diseases Group Information Specialist (VL) performed literature searches \nbetween 9th-13th October 2023, without language restrictions, in the following databases, using the \nsearch terms described in Appendix xx: Cochrane Central Register of Controlled Trials (CENTRAL), \nincluded in the Cochrane Library, issue 10, October 2023; MEDLINE (Ovid); Embase (Ovid); Social Science \nCitation Index (ISI Web of Knowledge); WHO Global Index Medicus; and Cinahl (EBSCOhost)",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Annex 1 of the qualitative synthesis, describing the database searches and beginning the detailed Ovid MEDLINE search strategy with terms on acceptability, equity and barriers.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-481",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 481",
      "pageStart": 481,
      "pageEnd": 481,
      "target": "../sections/page-481.md",
      "narrative": null,
      "content": "476 \n \n29 \npatient retention*.mp.  \n30 \nhealth personnel attitude/ or providers experience*.mp. \n \n31 \nClinical Pathway/ \n \n32 \nfacilitator*.ab. or facilitator*.ti. or (implement* or adoption).tw. \n33 \n(Appropriateness or feasibility or readiness or penetration or reach).tw.  \n34 \nSustainability.tw. or cost effectiveness.mp. or Cost-Benefit Analysis/ \n \n35 \n23 or 24 or 25 or 26 or 27 or 28 or 29 or 30 or 31 or 32 or 33 or 34 \n \n36 \n22 and 35 \n \n37 \ninterview*.mp.  \n38 \nsurvey*.mp. or Health Survey/ or \"Surveys and Questionnaires\".tw. \n \n39 \nQualitative Research/ \n \n40 \nFocus group discussion*.",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the MEDLINE search strategy for the qualitative synthesis, listing terms for implementation outcomes, qualitative methods and named NAAT products.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-482",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 482",
      "pageStart": 482,
      "pageEnd": 482,
      "target": "../sections/page-482.md",
      "narrative": null,
      "content": "477 \n \n21 \n7 or 8 or 9 or 10 or 11 or 12 or 13 or 14 or 15 or 16 or 17 or 18 or 19 or 20 \n \n22 \n6 and 21 \n \n23 \nPatient attitude/ or acceptability.mp. or acceptance.mp. \n24 \nHealth Equity/ or equity.mp. or health care access/ \n \n25 \nPatient Preference/ or preference*.mp.  \n26 \nPatient Satisfaction/ or Attitude to Health/ \n \n27 \nbarrier*.mp. \n \n28 \nchallenge*.mp.  \n29 \npatient retention*.mp.  \n30 \nhealth personnel attitude/ or providers experience*.mp. \n \n31 \nClinical Pathway/ \n \n32 \nfacilitator*.ab. or facilitator*.ti. or (implement* or adoption).tw. \n33 \n(Appropriateness or feasibility or rea",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Part of the literature search strategy for a qualitative review on TB nucleic acid amplification tests, listing database search lines on acceptability, equity, preferences, barriers and implementation, followed by the start of the CINAHL search.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-483",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 483",
      "pageStart": 483,
      "pageEnd": 483,
      "target": "../sections/page-483.md",
      "narrative": null,
      "content": "478 \n \nframework analysis )  \nS5  S3 AND S4  \nS4 \n TX ( patient* and (attitude* or satisfaction or preference* or retention) ) OR TX ( (health personnel \nor healthcare professionals or healthcare workers) and (attitude* or satisfaction or preference*) ) \nOR TX ( Health Equity or equity or health care access ) OR TX ( (Appropriateness or feasibility or \nreadiness or penetration or reach) ) OR TX ( Sustainability or cost effectiveness )  \nS3  S1 AND S2  \nS2 \n TX ( (Xpert* or GeneXpert* or cepheid) ) OR TX ( (Truenat* or Molbio or TruRif or LAM or LAMP) ) \nOR TX ( (STANDARD M10 or SD Biosensor or",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the search strategies for the qualitative review, giving CINAHL and other database queries combining TB test names such as Xpert, Truenat and LAMP with terms on attitudes, equity and feasibility, ending with the WHO Global Index Medicus search.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-484",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 484",
      "pageStart": 484,
      "pageEnd": 484,
      "target": "../sections/page-484.md",
      "narrative": null,
      "content": "479",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A page containing only a page number.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-485",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 485",
      "pageStart": 485,
      "pageEnd": 485,
      "target": "../sections/page-485.md",
      "narrative": null,
      "content": "480 \n \n  \n  \nAnnex 2: PRISMA Flow chart for study selection \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n  \n \n \n \n \n \n \n \n \n \nIdentification \nReferences from other sources (n =0 )   \nCitation searching (n =0) \nGrey literature (n =0)  \nStudies screened (n = 3891) \nStudies sought for retrieval (n = 104) \nStudies assessed for eligibility (n = 104)     \nReferences removed (n = 1)   \nDuplicates identified manually (n = 0) \nDuplicates identified by Covidence (n = 1)  \nMarked as ineligible by automation tools (n = 0) \nOther reasons (n =0 ) \nStudies excluded (n",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A PRISMA flow chart for study selection, showing 3892 records screened, 104 assessed in full text and 49 studies included, two on Xpert Ultra and 47 on Xpert MTB/RIF.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-486",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 486",
      "pageStart": 486,
      "pageEnd": 486,
      "target": "../sections/page-486.md",
      "narrative": null,
      "content": "Click or tap here to enter text. \n \n481 \n \nAnnex 3: Table of included studies \nTable 2: Table of characteristics of included studies  \nStudy ID \nCountry, Income \nclassification, TB \nBurden \nType of health facility \nDiagnostic \ntechnology \nTarget population \nof test \nType & number of \nparticipants \nData collection \nmethods \nLow Complexity automated NAATs – Xpert Ultra test or cartridges \nMwaura 202030 \nKenya, Swaziland \nLower Middle \nIncome \nHigh burden \n \nWorkshop meeting \nwith TB stakeholders \nXpert Ultra \nPeople with \npresumptive TB \nHealth care providers, \nLaboratory \nprofessionals, Program",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The start of a table describing the studies included in the qualitative review, giving country, income and TB burden, facility type, test, population, participants and data collection methods; the page carries a stray form placeholder line.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-487",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 487",
      "pageStart": 487,
      "pageEnd": 487,
      "target": "../sections/page-487.md",
      "narrative": null,
      "content": "Click or tap here to enter text. \n \n482 \n \nBangladesh, \nCambodia,  \nDavids 201571 \nSouth Africa \nUpper middle \nincome \nHigh Burden \nPrimary health care \nclinics \nPOC TB test \nPeople with \npresumptive TB \nPrimary health care \nproviders (n=400) \nInterviews, \nquestionnaire \nand recorded \nanecdotal notes \nDe Camargo 201572 \nBrazil \nUpper middle \nincome \nHigh burden \nClinics \nXpert MTB/RIF \nPresumptive \ncases of TB and \nDR TB \nLaboratory technician \n(11), patients (30) , \nmanagers (unclear) \nInterviews, \ngroup meets to \nproduce \ndiagnostic \nflowcharts \n†Doulla 201942 \nTanzania \nLower-Middle \nIncome",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the table of included qualitative studies on TB molecular testing, covering studies in South Africa, Brazil, Tanzania and elsewhere with their participants and methods.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-488",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 488",
      "pageStart": 488,
      "pageEnd": 488,
      "target": "../sections/page-488.md",
      "narrative": null,
      "content": "Click or tap here to enter text. \n \n483 \n \n†Gwaza 202343 \nLesotho \nLow Income \nHigh Burden \n \nHospital, Clinic \nXpert MTB/RIF \nPeople with \npresumptive TB \nand DR-TB \nHealth care providers \nand Ministry of health \nstaff (n=44) \nInterviews \nHoang 201574 \nVietnam \nLower middle \nincome \nHigh Burden \nDistrict health centre \n/hospital \nXpert MTB/RIF \nPatients at high \nrisk for MDR-TB \nTB provincial staff \nmembers & health staff \nat central, provincial \ndistrict & community \nlevel (n=110) \nInterviews, \nFGDs and \nDocument \nreviews \n†Isangula 202344 \nTanzania, Kenya, \nUganda, and \nEthiopia \nLow Income",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the table of included studies, with entries from Lesotho, Viet Nam, East African countries, China and Nepal among others.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-489",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 489",
      "pageStart": 489,
      "pageEnd": 489,
      "target": "../sections/page-489.md",
      "narrative": null,
      "content": "Click or tap here to enter text. \n \n484 \n \nLower middle \nincome \nprimary health \ncentres, district public \nhealth office \nlaboratory \nliving with HIV \n(PLHIV); severe \nforms of TB; and \nin presumptive \nMDR TB \n†Juliasih 202034 \nIndonesia \nLower-Middle \nIncome \nHigh Burden \nPrimary Health Centre \nRapid molecular \ntest/Unspecified \nNAAT \nPeople with \npresumptive DR-\nTB \nTB Officers (n=not \nreported) \nInterviews, \nObservations \n†Kanakaraju 202035 \nIndia \nLower-Middle \nIncome \nHigh Burden \nHospital, Health \nCentres, Dispensaries \nCB-NAAT \n(unspecified) \nPeople with \npresumptive TB \nHealth Care pro",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the table of included studies, listing studies from Indonesia, India, Tanzania and other countries with their settings, participants and interview or focus group methods.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-490",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 490",
      "pageStart": 490,
      "pageEnd": 490,
      "target": "../sections/page-490.md",
      "narrative": null,
      "content": "Click or tap here to enter text. \n \n485 \n \nMohammed 202051 \nEthiopia \nLow income \nHigh burden \nOPD, HIV clinic, \ndiabetic and maternal \nchild health clinics \nXpert MTB/RIF \nPresumptive TB \npatients \nHead of each health \nfacility and a focal \nhealth care provider \n(n=35) \nInterviews & \ndocument \nreviews \n†Nagar 202036 \nIndia \nLower-Middle \nIncome \nHigh Burden \nHospital, Clinic, Private \nnursing home, \nGovernment medical \ncollege, \nXpert MTB/RIF or \nTB PCR \nPresumptive EPTB \n(musculoskeletal \nTB) \nOrthopedic surgeons \n(n=12) \nInterviews \nNaidoo 201552 \nSouth Africa \nUpper middle \nincome \nHigh Bu",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the table of included studies, covering studies in Ethiopia, India, South Africa and Kenya, including some focused on children and extrapulmonary TB.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-491",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 491",
      "pageStart": 491,
      "pageEnd": 491,
      "target": "../sections/page-491.md",
      "narrative": null,
      "content": "Click or tap here to enter text. \n \n486 \n \nclinicians, laboratory \ntechnologists (n=40) \nOo 201961 \nMyanmar \nLower Middle \nincome \nHigh Burden \nTownship TB centre \nXpert MTB/RIF \nPresumptive DR-\nTB patients \nTownship TB \ncoordinators & lab \ntechs (n=32) \nInterviews \n†Paudel 202137 \nNepal \nLower-Middle \nIncome \nHigh Burden \nHospital \nXpert MTB/RIF \nPeople with \npresumptive TB \nHealth care providers, \nhospital managers, \ndistrict focal persons, \nprovincial team leader, \nprogram coordinators \nand patients (n=27) \nInterviews, \nFGDs \nPhyo 201966 \nMyanmar \nLower middle \nincome \nHigh Burden \nCommunit",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the table of included studies, with entries from Myanmar, Nepal and Cambodia among others.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-492",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 492",
      "pageStart": 492,
      "pageEnd": 492,
      "target": "../sections/page-492.md",
      "narrative": null,
      "content": "Click or tap here to enter text. \n \n487 \n \nHigh Burden \nhospital and health \ncentres \nofficers, nurses, \nlaboratory staff, and TB \nofficers (n=26) \nSaria 202054 \nIndia \nLower-Middle \nIncome \nHigh Burden \nnot clear, some private \nclinics and informal \nproviders \nXpert MTB/RIF \nPresumptive TB \npatients  \nFormal and informal \nhealth providers \nInterviews and \nshadowing \nformal health \ncareproviders \n†Shrestha 202339 \nIndia \nLower-Middle \nIncome \nHigh Burden \nAnti-Retroviral \nTherapy Centres \nCB NAAT or Xpert \nMTB/RIF \nHIV-TB Co-\ninfected patients \nHealth care providers \n(n=14) \nInterviews \nShewad",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The final rows of the included studies table, covering studies from India such as those involving people with HIV and pregnant women, with a note marking studies from the updated search.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-493",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 493",
      "pageStart": 493,
      "pageEnd": 493,
      "target": "../sections/page-493.md",
      "narrative": null,
      "content": "Click or tap here to enter text. \n \n488 \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \nAnnex 4: Table of excluded studies \nTable 3: Excluded studies and reasons for exclusion \nNo. \nStudy ID \nReason for exclusion \n1. \n†Adam 201979 \nStudy type (Editorial) \n2. \nAdepoyibi 201880 \nNot a qualitative study design \n3. \nAlbert 201681 \nStudy type (Review article) \n4. \nArdizzoni 201582 \nData not analysed using qualitative analysis methods  \n5. \nArdizzoni 201582 \nDuplicate",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The start of a table of studies excluded from the qualitative review, listing the first five with reasons such as study type or non-qualitative design.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-494",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 494",
      "pageStart": 494,
      "pageEnd": 494,
      "target": "../sections/page-494.md",
      "narrative": null,
      "content": "Click or tap here to enter text. \n \n489 \n \n6. \n†Boulanger 202083 \nDid not address phenomenon of interest  \n7. \n†Brown 202184 \nStudy type (systematic review) \n8. \n†Cash 201385 \nStudy type (Review) \n9. \nCharoensook 201886 \nData not analysed using qualitative analysis methods \n10. \nChalwa 201687 \nNot a qualitative study design \n11. \n†Clouse 201288 \nStudy type (review)  \n12. \n†Cowan 201389 \nDid not address phenomenon of interest \n13. \n†Daftary 202090 \nStudy type (Letter to the Editor) \n14. \nDa Silva 201491 \nNot a qualitative study design \n15. \n†deHaas 202392 \nNot a qualitative study design \n16. \nD",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the table of excluded studies, giving reasons such as reviews, letters, non-qualitative designs or not addressing the topic.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-495",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 495",
      "pageStart": 495,
      "pageEnd": 495,
      "target": "../sections/page-495.md",
      "narrative": null,
      "content": "Click or tap here to enter text. \n \n490 \n \n24. \n†Kedia 201599 \nStudy type (conference abstract) \n25. \n†Kondo 2021100 \nNot a qualitative study design \n26. \n†Lemaire 2010101 \nStudy type (Debate) \n27. \nLorent 2016102 \nDid not address phenomenon of interest \n28. \nMaraba 2018103 \nDid not address phenomenon of interest \n29. \n†Mbelele 2018104 \nStudy type (Meta-narrative review) \n30. \n†Mistry 2016105 \nDid not address phenomenon of interest \n31. \n†Mohapatra 2011106 \nStudy type (letter to the editor)  \n32. \n†Moyenga 2015107 \nNot a qualitative study design \n33. \n†Moyenga 2015107 \nDuplicate \n34. \nMpagama",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the list of excluded studies with reasons for exclusion, such as conference abstracts and lack of qualitative analysis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-496",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 496",
      "pageStart": 496,
      "pageEnd": 496,
      "target": "../sections/page-496.md",
      "narrative": null,
      "content": "Click or tap here to enter text. \n \n491 \n \n41. \nOchodo 2019114 \nDid not address phenomenon of interest \n42. \n†Oga-Omenka 2020115 \nDid not address phenomenon of interest \n43. \n†Oga-Omenka 2020116 \nStudy type (systematic review)  \n44. \n†Palamountain 2012117 \nStudy type (Review)  \n45. \nPalupi 2019118 \nNot a qualitative study design \n46. \n†Pamba 2022119 \nStudy type (conference abstract) \n47. \n†Parcell 2012120 \nStudy type (Conference abstract) \n48. \nPaudel 202137 \nDuplicate  \n49. \n†Perkel 2009121 \nStudy type (chapter in a textbook etc) \n50. \n†Phyo 201966 \nDid not address phenomenon of interest \n51.",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The last entries in the excluded studies table, with reasons including duplicates, conference abstracts and very thin data.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-497",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 497",
      "pageStart": 497,
      "pageEnd": 497,
      "target": "../sections/page-497.md",
      "narrative": null,
      "content": "492 \n \nAnnex 5: Assessment of methodological quality \nTable 4: Assessment of methodological limitations of the included studies using EPPI -Centre tool \nNo. \nStudy ID \nWere steps taken to \nincrease rigour in \nsampling? \nWere steps taken \nto increase rigour \nin the data \ncollected? \nWere steps taken to \nincrease rigour in the \nanalysis of the data? \nWere the findings of the \nstudy grounded \nin/supported by the \ndata? \nLow complexity NAATs: Xpert Ultra \n1. \nMwaura 202030 \nYes, several steps were \ntaken \nYes, several steps \nwere taken \nYes, several steps were \ntaken \nYes, a fairly thorough \nattem",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A table assessing the methodological rigour of included studies with the EPPI-Centre tool, rating sampling, data collection, analysis and grounding of findings, beginning with the Xpert Ultra studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-498",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 498",
      "pageStart": 498,
      "pageEnd": 498,
      "target": "../sections/page-498.md",
      "narrative": null,
      "content": "493 \n \n13. \n†Pongpeeradech \n202238 \nYes, a few steps were \ntaken \nYes, a few steps \nwere taken \nYes, a few steps were \ntaken \nYes, a few steps were \ntaken \n14. \n†Shrestha 202339 \nYes, a few steps were \ntaken \nYes, a few steps \nwere taken \nYes, a few steps were \ntaken \nYes, several steps were \ntaken \n15. \n†Singh 202140 \nYes, a few steps were \ntaken \nYes, several steps \nwere taken \nYes, several steps were \ntaken \nYes, a fairly thorough \nattempt was made \n16. \n†Velayutham \n202341 \nYes, several steps were \ntaken \nYes, several steps \nwere taken \nYes, a few steps were \ntaken \nYes, several steps were",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the EPPI-Centre quality assessment table, rating rigour in sampling, data collection and analysis for further included studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-499",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 499",
      "pageStart": 499,
      "pageEnd": 499,
      "target": "../sections/page-499.md",
      "narrative": null,
      "content": "494 \n \nwas made \n30. \nMcDowell \n201856 \nYes, a fairly thorough \nattempt was made \nYes, few steps were \ntaken \nYes, a fairly thorough \nattempt was made \nYes, a fairly thorough \nattempt was made \n31. \nMedina-Morino \n202165 \nYes, several steps were \ntaken \nYes, few steps were \ntaken \nYes, several steps were \ntaken \nYes, a fairly thorough \nattempt was made \n32. \nMnyambwa \n201857 \nYes, few steps were taken \nYes, few steps were \ntaken \nNo, not at all/Not \nstated/Can't tell \nNo, not at all/Not \nstated/Can't tell \n33. \nMohammed \n202051 \nYes, few steps were taken \nYes, few steps were \ntaken \nNo, not at",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the methodological quality assessment table for included studies, with ratings ranging from no steps taken to fairly thorough attempts.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-500",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 500",
      "pageStart": 500,
      "pageEnd": 500,
      "target": "../sections/page-500.md",
      "narrative": null,
      "content": "495 \n \n47. \nViyageetha \n201964 \nYes, few steps were taken \nYes, few steps were \ntaken \nYes, few steps were \ntaken \nYes, few steps were \ntaken \n† Studies identified in updated search \n \n \n \n \n \n \n \n \n \nAnnex 6: Summary of review findings (Xpert Ultra) \nTable 5: Summary table of qualitative findings for Xpert Ultra studies \nFinding \n# \nReview finding \nMethodological \nlimitations \nCoherence \nRelevance \nAdequacy \nCERQual \nassessment \nof \nconfidence \nin the \nevidence \nExplanation of \nCERQual assessment \nStudies \ncontributing to \nreview finding \nCritical aspects users value \n1 \nCompared to existing",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Finishes the quality assessment table and begins a table of qualitative review findings on Xpert Ultra, giving CERQual confidence ratings and explanations, starting with what users value.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-501",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 501",
      "pageStart": 501,
      "pageEnd": 501,
      "target": "../sections/page-501.md",
      "narrative": null,
      "content": "496 \n \n(n=13) with \nalbeit a good \nvariety of users \nChallenges to realizing these values \n3 \nPatients can be \nreluctant to test for \nTB/MDR-TB because of \nstigma related to the  \nassociation of TB  with \nlow socioeconomic \nstatus \nNo/very minor \nconcerns- One \nstudy which had \nfairly good \nmethodological \nquality across all \nfour components \nNo concerns -  \ngood fit of \nfinding with \nprimary study  \nModerate \nconcerns, varied \nparticipants from \nhospital set-up, \nfrom one country \n(upper-middle \nincome) with high \nDR-TB burden, \nperceptions about \npatients were \nstated by health \nworkers  \nMi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the Xpert Ultra findings table with CERQual assessments, covering challenges such as patients' reluctance to test because of stigma linked to TB and low socioeconomic status.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-502",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 502",
      "pageStart": 502,
      "pageEnd": 502,
      "target": "../sections/page-502.md",
      "narrative": null,
      "content": "497 \n \n6 \nThe introduction of \nXpert Ultra increases \nworkload of laboratory \nstaff if added onto \nexisting work without \nadjusting staffing \narrangements, \ndecreasing motivation \nof health workers who \nmay be hesitant to \naccept testing with \nXpert Ultra \nNo/very minor \nconcerns - only \none study which \ntook several steps \ntowards \nmethodological \nquality across all \nthe four domains] \nNo/very minor \nconcerns - \nfinding captures \nthe primary \nstudy well, not \naccepting CB-\nNAAT because \nof workload \nconcerns and \nlack of \nmotivation \namongst the \nhealthcare \nworkers was \nmentioned in \nthis st",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the Xpert Ultra findings table, including a finding that adding Xpert Ultra without adjusting staffing increases laboratory workload and lowers health worker motivation.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-503",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 503",
      "pageStart": 503,
      "pageEnd": 503,
      "target": "../sections/page-503.md",
      "narrative": null,
      "content": "498 \n \n9 \nLack of Xpert testing in \nsurrounding health \nfacilities hamper access \nto prompt and accurate \ntesting and treatment \nparticularly for \nvulnerable groups. \nNo or very minor \nconcerns - only \none study \ncontributed to \nthis review finding \nhowever, the \nstudy took several \nsteps towards \nincreasing quality \nNo concerns - \ncaptures the \ndata from \nprimary studies  \nModerate \nconcerns- good \nvariety of users, \nonly hospital set-\nup, one high DR-TB \nburden country \n(upper-middle \nincome), the study \nwas conducted in \nbetter equipped \nhospitals for TB \ntesting compared \nto other hospital",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Ends the Xpert Ultra findings with lack of testing in nearby facilities limiting access for vulnerable groups, then introduces a table of findings for all included NAATs integrated with an earlier Cochrane qualitative review.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-504",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 504",
      "pageStart": 504,
      "pageEnd": 504,
      "target": "../sections/page-504.md",
      "narrative": null,
      "content": "499 \n \nFinding # Review finding\nMethodological limitations\nCoherence\nRelevance\nAdequacy\nCERQual assessment \nof confidence in the \nevidence\nEplanation of CERQual assessment\nStudies contribut\nreview finding\n1\nPatients in high-burden TB settings \nvalue 1) getting an accurate diagnosis \nand reaching diagnostic closure (finally \nknowing what is wrong with me), 2) \navoiding diagnostic delays as they \nexacerbate existing financial hardships \nand emotional and physical suffering \nand make patients feel guilty for \ninfecting others (especially children), 3) \nhaving accessible facilities and 4) \nreducin",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A table of qualitative findings for all included NAATs with CERQual confidence ratings, opening with what patients in high-burden settings value, such as an accurate diagnosis and avoiding delays. Further findings cover what health workers and programme staff value from rapid molecular testing. Each finding lists methodological, coherence, relevance and adequacy concerns and contributing studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-505",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 505",
      "pageStart": 505,
      "pageEnd": 505,
      "target": "../sections/page-505.md",
      "narrative": null,
      "content": "500 \n \n7\nPatients can be reluctant to test for \nTB/MDR-TB because of stigma related to \nMDR-TB or related to having interrupted \ntreatment in the past or relatd to \nassociation of TB with low \nsocioeconomic status, because of fears \nof side effects, the failure to recognize \nsymptoms, the inability to produce \nsputum and the cost, distance and travel \nconcerns related to (repeat) clinic visits. \nno/very minor concerns- six \nout of ten included studies \nhave fairly good \nmethodological quality across \nall four components\nno concerns -  good fit of \nfinding with primary study \nno/very minor conc",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the NAAT qualitative findings table, including patients' reluctance to test owing to stigma, fear of side effects, inability to produce sputum and travel costs. It also covers how clinic workflows and patient flows affect use of the tests. Parts of the extracted text are garbled, with words broken into fragments.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-506",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 506",
      "pageStart": 506,
      "pageEnd": 506,
      "target": "../sections/page-506.md",
      "narrative": null,
      "content": "501 \n \n11\nThe lack of sufficient resources and of \nensuring maintenance (i.e. stock-outs; \nunreliable logistics; lack of funding, \nelectricity, space, air conditioners, and \nsputum containers; dusty environment, \nand delayed or absent local repair \noption) leads to higher test failure rates \nand underutilization and negatively \ninfluences uptake and impact of LC-\nNAAT. \nminor concerns - of 25 studies \ncontributing to the finding, \nabout 13 were of fairly good \nor very quality while the \nother half took a few steps \ntowards methodological \nquality]\nno concerns - captures the \ndata from primary",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the NAAT findings table, covering how stock-outs, poor logistics, lack of power, space and maintenance raise test failure and underuse. It also includes findings on access and equity concerns. Most findings are rated with moderate or high confidence.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-507",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 507",
      "pageStart": 507,
      "pageEnd": 507,
      "target": "../sections/page-507.md",
      "narrative": null,
      "content": "502 \n \n٭ Studies on Xpert Ultra contributed to reviews findings in yellow rows. \n \n \n \n \n \n \n \n \n \n \n \n \n \nList of Tables \nTable 1: Classes of NAATs for detection of TB and DR-TB \nTable 2: Table of characteristics of included studies  \nTable 3: Excluded studies and reasons for their exclusion  \nTable 4: Assessment of methodological limitations of included studies using the EPPI-Centre tools  \nTable 5: Summary table of qualitative findings for Xpert Ultra studies  \nTable 6: Summary table of qualitative findings for all included NAATs  \n \nList of Figures \nFig 1: PRISMA flow chart showing study s",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes that studies on Xpert Ultra contributed to highlighted findings and gives lists of the review's tables and figures, with some repeated table text.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-508",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 508",
      "pageStart": 508,
      "pageEnd": 508,
      "target": "../sections/page-508.md",
      "narrative": null,
      "content": "503",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A page containing only a page number.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-509",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 509",
      "pageStart": 509,
      "pageEnd": 509,
      "target": "../sections/page-509.md",
      "narrative": null,
      "content": "504 \n \nWeb Annex B.11. Moderate complexity automated NAATs: Diagnostic \naccuracy for TB detection and detection of resistance to rifampicin and \nisoniazid. A systematic review and meta-analysis \n \nDecember 2020 \n \n \n \n \n \n \n \nAuthors: \nMikashmi Kohli1,2; Emily MacLean1,2; Madhukar Pai1,2; Claudia M. Denkinger3,4; Samuel G. \nSchumacher3 \n \n \nCorresponding author: \nSamuel Schumacher PhD \nEvidence & Policy Lead at FIND \nGeneva, Switzerland \nEmail: Samuel.Schumacher@finddx.org \n \n \n \n \n \n \n \n1 Department of Epidemiology, Biostatistics and Occupational Health, McGill University, Montreal, Canada.",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title page of Web Annex B.11, a December 2020 systematic review of the diagnostic accuracy of moderate-complexity automated NAATs for TB and rifampicin and isoniazid resistance, with authors and affiliations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-510",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 510",
      "pageStart": 510,
      "pageEnd": 510,
      "target": "../sections/page-510.md",
      "narrative": null,
      "content": "505 \n \n \n \nEXECUTIVE SUMMARY \n \nBackground \nThe World Health Organization (WHO) estimates that in 2017, 10 million people became ill with TB \nglobally. Drug-resistant TB is an enormous threat. In 2017, an estimated 558,000 people were newly \ndiagnosed with rifampicin-resistant TB, 468,720 of whom had multidrug-resistant TB (MDR-TB). \nMDR-TB is caused by infection with M. tuberculosis bacteria that are resistant to at least rifampicin and \nisoniazid. The introduction and rollout of nucleic acid amplification tests (NAATs) has revolutionized \nthe area of TB diagnosis. WHO recommends Xpert, LPA,",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The opening of the executive summary for the review of moderate-complexity NAATs, giving the global TB and drug-resistant TB burden in 2017 and describing the review methods, including pooled meta-analysis in Stata.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-511",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 511",
      "pageStart": 511,
      "pageEnd": 511,
      "target": "../sections/page-511.md",
      "narrative": null,
      "content": "506 \n \n \n \nA total of 36 studies contributed to 39 unique datasets (four studies provided data for more than one \nindex test). All the studies were performed in central laboratories as the index tests require sophisticated \nlaboratory infrastructure. For most studies, samples were included in the study on arrival into the \nlaboratory, as a result there was limited data on the demographics of the included patient population, \nsuch as age, HIV status, past TB history. \n \nA total of 29 studies with 13852 specimens (4767 TB positive specimens) provided data for evaluating \nTB detection from the fi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarises the review results: 36 studies from central laboratories gave 39 datasets, with limited patient data, and the tests showed accuracy for TB, rifampicin and isoniazid resistance detection comparable to WHO-recommended molecular tests.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-512",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 512",
      "pageStart": 512,
      "pageEnd": 512,
      "target": "../sections/page-512.md",
      "narrative": null,
      "content": "507 \n \n \n \n \nBackground \n \nTuberculosis (TB) is caused by Mycobacterium tuberculosis (Mtb) bacteria. TB causes tremendous \nsuffering worldwide and has surpassed HIV/AIDS as the world’s leading infectious cause of death. The \nWorld Health Organization (WHO) estimates that in 2017, 10 million people became ill with TB \nglobally. Approximately, 1.3 million HIV-negative people and 300,000 HIV-positive people died from \nTB (WHO 2018). Drug-resistant TB is an enormous threat. In 2017, an estimated 558,000 people were \nnewly diagnosed with rifampicin-resistant TB, 468,720 of whom had multidrug-resist",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives background on the TB and MDR-TB burden and on molecular tests already recommended by WHO, such as Xpert, line-probe assays, LAMP and Truenat.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-513",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 513",
      "pageStart": 513,
      "pageEnd": 513,
      "target": "../sections/page-513.md",
      "narrative": null,
      "content": "508 \n \n \n \nRecently, several companies have developed molecular tests for TB and RIF/INH resistance detection \non centralized platforms, many of which have been established as multi-disease platforms, primarily for \ndetection of Human immunodeficiency virus (HIV), Human Papillomavirus (HPV) and Hepatitis C \nVirus. On these high-throughput platforms a large number of patient specimens can be tested in a shorter \namount of time. Some assays can generate around 1000 results in just 8 hours (Cobas 8800 platform). \nThis could help reduce the current problem of diagnostic delays and delay in treatme",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes high-throughput centralised molecular platforms for TB and resistance detection, such as Abbott RealTime and Roche Cobas, and the Bruker-Hain FluoroType assays with their extraction and detection systems.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-514",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 514",
      "pageStart": 514,
      "pageEnd": 514,
      "target": "../sections/page-514.md",
      "narrative": null,
      "content": "509 \n \n \n \nBecton Dickinson (BD) has developed a multiplexed real-time PCR (BD MAXTM MDR-TB, CE-marked \nsince 2018) assay for the detection of Mtb and resistance to both RIF and INH. The platform utilizes 5- \ncolour detection (UNITAID 2017). For Mtb detection, this assay detects multicopy genomic targets \nIS6110 and IS1081 as well as a single copy genomic target. For resistance, the assay detects rpoB, inhA \npromoter and katG genes. The LoD of BD Max MDR-TB assay has been reported as 0.5 cfu/ml for Mtb \ndetection and 6 cfu/ml for resistance detection. The assay uses a fully-integrated, automat",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the BD MAX MDR-TB assay and its genetic targets and detection limits, then sets out the review's eligible study types, including case-control studies and data from a WHO public call.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-515",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 515",
      "pageStart": 515,
      "pageEnd": 515,
      "target": "../sections/page-515.md",
      "narrative": null,
      "content": "510 \n \n \n \nTypes of participants \nPatients of all age groups presumed of having or confirmed with pulmonary TB or MDR-TB, in all \nsettings and all countries, were included. Specimen types were limited to pulmonary TB specimens (such \nas sputum, bronchoalveolar lavage, tracheal aspirate etc.). \n \nIndex tests \n1. Abbott RealTime MTB and Abbott RealTime MTB RIF/INH \n2. Bruker-Hain FluoroType MTB and FluoroType MTBDR \n3. BD Max MDR-TB assay \n4. Roche Cobas MTB assay and Roche Cobas MTB RIF/INH assay \n \n \nTarget conditions \nWe considered the following three target conditions: resistance of Mtb to r",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Sets out the review's participants, the four index test families, the target conditions and the reference standards, including a composite standard combining phenotypic testing and sequencing.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-516",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 516",
      "pageStart": 516,
      "pageEnd": 516,
      "target": "../sections/page-516.md",
      "narrative": null,
      "content": "511 \n \n \n \nlow-level heteroresistance, below the limit of detection of the sequencing technologies used; e.g. Sanger). \nThe composite reference standard was constructed without knowledge of the index test results. \n \nOutcome measures \n \nFor Mtb detection, sensitivity refers to the proportion of specimens or isolates with a culture positive \nfor Mtb that were detected as Mtb-positive by the index test. \nFor RIF- or INH-resistance detection, sensitivity refers to the proportion of specimens or isolates with \nthe phenotypic or composite reference standard separately, demonstrating RIF/INH-resista",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains how discordant composite reference results were handled, defines sensitivity and specificity outcomes, and describes the search and screening by two review authors.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-517",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 517",
      "pageStart": 517,
      "pageEnd": 517,
      "target": "../sections/page-517.md",
      "narrative": null,
      "content": "criteria (Screen 2). In Screen 2, any discrepancies were resolved by discussion between the review \nauthors, and for any study that were not resolved by the review authors, were then decided by a third \nreview author (SS). If a study contributed data to more than one analysis (e.g. two different index tests \nin one study, on the same or different specimens), it was considered as two or more datasets. \n \nData extraction \nWe created a data extraction form, piloted it with a subset of eligible studies, and then finalized the form \n(Appendix B). Two review authors (MK and EM) independently extract",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes resolution of screening disagreements, data extraction and quality assessment, noting that publication bias tests were not done and unpublished data were included.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-518",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 518",
      "pageStart": 518,
      "pageEnd": 518,
      "target": "../sections/page-518.md",
      "narrative": null,
      "content": "Meta-analysis was after pooling the dataset from all the index tests together. We also visually saw the \nresults from each index test by forest plots. Bivariate random effects meta-analyses were performed (Chu \n2006; Reitsma 2005) using the metandi programme in Stata for index tests with enough studies that \nincluded data to calculate sensitivity, specificity, and 95% confidence intervals for each of these. \nSummary and individual estimates were also presented graphically with the 95% Confidence Interval \n(CI) interval and prediction region. \n \nForest plots were visually assessed for heterogen",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the bivariate random-effects meta-analysis methods in Stata, investigation of heterogeneity and the use of GRADE to assess certainty of evidence.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-519",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 519",
      "pageStart": 519,
      "pageEnd": 519,
      "target": "../sections/page-519.md",
      "narrative": null,
      "content": "Summary of Findings tables. As recommended, we rated the quality of evidence as high (no points \nsubtracted), moderate (one point subtracted), low (two points subtracted), or very low greater than two \npoints subtracted) based on five factors: risk of bias, indirectness, inconsistency, imprecision, and other \nconsiderations. We subtracted one point when there was a serious issue identified or two points when \nthere was a very serious issue identified in any of the factors used to judge the quality of evidence. Two \nreview authors (MK and EM) discussed and applied the GRADE criteria. \n \nResults",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains how evidence quality was rated using GRADE, then reports the number of studies for each index test for TB detection, all using culture as the reference.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p520-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 520,
      "pageEnd": 520,
      "target": "../images/img-p520-1.webp",
      "narrative": "draft",
      "content": "Stacked horizontal bar chart (Figure 2) of QUADAS-2 judgements across 29 studies, showing the proportion rated high (red), unclear (yellow) or low (green) for risk of bias and applicability concerns by domain. For risk of bias, patient selection is 7 high, 9 unclear and 13 low; index test is 29 low; reference standard is 3 high, 6 unclear and 20 low; flow and timing is 1 high, 2 unclear and 26 low. For applicability, patient selection is 13 high, 3 unclear and 13 low, and index test and reference standard are 29 low.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-520",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 520",
      "pageStart": 520,
      "pageEnd": 520,
      "target": "../sections/page-520.md",
      "narrative": null,
      "content": "Figure 2: QUADAS-2 assessment for TB detection \n \nPatient selection \n \n \nRisk of bias \n \nOf the total 29 studies, 16 (55%) had high or unclear risk of bias as they either did prior testing before \nincluding specimens in the study or used convenience sampling or the method of participant selection \nwas not reported. \n \nApplicability concern \n \nOnly 45% of the studies were conducted in high TB/MDR TB burden country. \n \nIndex test \n \nAs all the index test assays were automated and thresholds were pre-defined, there is no scope of \nsubjective interpretation of results. We judged all of them at low",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarises QUADAS-2 risk of bias for TB detection, noting that over half of the 29 studies had high or unclear risk in patient selection, and reports pooled sensitivity of 93.0% and specificity of 97.7%.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p521-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 521,
      "pageEnd": 521,
      "target": "../images/img-p521-1.webp",
      "narrative": "draft",
      "content": "Forest plot (Figure 3) of sensitivity and specificity for TB detection for 29 studies, listing TP, FP, FN and TN counts with sensitivity and specificity (95% CI) and plots of each estimate on a 0–1 scale. Most sensitivity estimates lie between 0.79 (Berhanu 2018) and 1.00; most specificity estimates are 0.92 or higher, with lower values for Obasanya 2017 (0.60), Shaheed 2020 (0.68), Paradis 2018 (0.72) and Wang 2016 (0.84).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-521",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 521",
      "pageStart": 521,
      "pageEnd": 521,
      "target": "../sections/page-521.md",
      "narrative": null,
      "content": "Figure 3: Forest plot of included studies for TB detection \nSubgroup analysis: Smear status \n \nThere were 26 studies that provided information for TB detection stratified by smear status. A total of \n2880 smear positive and 9447 smear negative specimens were evaluated. \nThe sensitivity for smear positive specimens ranged from 93% to 100%. The pooled sensitivity was \n98.8% (93.7 to 99.7) (Table 2b) (Figure 4).",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A forest plot caption followed by the smear-positive subgroup analysis, with pooled sensitivity of 98.8%.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p522-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 522,
      "pageEnd": 522,
      "target": "../images/img-p522-1.webp",
      "narrative": "draft",
      "content": "Forest plot (Figure 4) of sensitivity for TB detection in smear-positive specimens for 26 studies, listing TP, FP, FN and TN counts with sensitivity (95% CI) and a plot of each estimate on a 0–1 scale. Sensitivity ranges from 0.93 (Shaheed 2020) to 1.00, with most studies at 0.97 or above.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p522-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 522,
      "pageEnd": 522,
      "target": "../images/img-p522-2.webp",
      "narrative": "draft",
      "content": "Forest plot (Figure 5) of sensitivity and specificity for TB detection in smear-negative specimens for 27 studies, listing TP, FP, FN and TN counts with sensitivity and specificity (95% CI) and plots of each estimate on a 0–1 scale. Studies are ordered by an 'S- C+' column (from 65.12 to 3.14); sensitivity varies widely, from 0.22 (Howlett 2020) to 1.00, while most specificity estimates are 0.91 or higher.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-522",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 522",
      "pageStart": 522,
      "pageEnd": 522,
      "target": "../sections/page-522.md",
      "narrative": null,
      "content": "Figure 4: Forest plot of included studies for TB detection in smear positive specimens. \n \nFor smear negative the sensitivity ranged from 22% to 100%, and specificity ranged from 68% to 100%. \nMeta-analysis of the data provided a pooled sensitivity of 86.1% (73.4 to 93.2) and the pooled specificity \nwas 96.4% (93.6 to 98.0) (Table 2b) (Figure 5). \n \n \n \n \nFigure 5: Forest plot of included studies for TB detection in smear negative specimens.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports pooled sensitivity of 86.1% and specificity of 96.4% for TB detection in smear-negative specimens, with forest plot captions.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p523-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 523,
      "pageEnd": 523,
      "target": "../images/img-p523-1.webp",
      "narrative": "draft",
      "content": "Paired forest plots (part of Figure 6, head-to-head comparisons) for index-test TB detection in Berhanu 2018, Scott 2017 and Wang 2016 (upper panel) and for Xpert on the same specimens (lower panel), each listing TP, FP, FN, TN and sensitivity and specificity (95% CI). Index-test sensitivity is 0.79, 0.85 and 1.00 versus 0.82, 0.92 and 0.97 for Xpert; specificity is 0.97, 0.92 and 0.84 versus 1.00, 0.98 and 0.90.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-523",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 523",
      "pageStart": 523,
      "pageEnd": 523,
      "target": "../sections/page-523.md",
      "narrative": null,
      "content": "Sensitivity analysis: Manufacturer driven \n \nThere were four studies which were manufacturer driven in this dataset for TB detection. We excluded \nthese studies and performed a sensitivity analysis to evaluate the diagnostic accuracy of these class of \ntests for TB detection. The definition of manufacturer driven was that we removed studies which were \nprovided by the manufacturers’ during the public call. Additionally, we reviewed the included published \nstudies and if the authors of any study were manufacturers’ and they were involved in the design, analysis \nor running the study, we also ex",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes a sensitivity analysis excluding four manufacturer-driven studies for TB detection and introduces head-to-head comparisons of the tests with Xpert on the same specimens.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p524-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 524,
      "pageEnd": 524,
      "target": "../images/img-p524-1.webp",
      "narrative": "draft",
      "content": "Paired forest plots (part of Figure 6, head-to-head comparisons) for index-test TB detection in Berhanu 2018 and Howlett 2020 (upper panel) and for Xpert Ultra on the same specimens (lower panel), each listing TP, FP, FN, TN and sensitivity and specificity (95% CI). Index-test sensitivity is 0.79 and 0.81 versus 0.89 and 0.93 for Ultra; specificity is 0.97 in both studies versus 0.96 and 0.97.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p524-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 524,
      "pageEnd": 524,
      "target": "../images/img-p524-2.webp",
      "narrative": "draft",
      "content": "Forest plot (part of Figure 6) comparing two rows from one study: Obasanya 2017 (sensitivity 0.89 [0.79, 0.95], specificity 0.60 [0.53, 0.66]) and Obasanya 2017-Xpert (sensitivity 0.79 [0.67, 0.87], specificity 0.94 [0.90, 0.97]), with TP, FP, FN and TN counts.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p524-4",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 524,
      "pageEnd": 524,
      "target": "../images/img-p524-4.webp",
      "narrative": "draft",
      "content": "Paired forest plots (part of Figure 6, head-to-head comparisons) for index-test TB detection in Paradis 2018 and Shah 2019 (upper panel) and for a comparator on the same specimens (lower panel), each listing TP, FP, FN, TN and sensitivity and specificity (95% CI). Index-test sensitivity is 0.95 and 0.93 and specificity 0.72 and 0.97; for the comparator, sensitivity is 0.95 and 0.90 and specificity 0.73 and 0.98.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p524-8",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 524,
      "pageEnd": 524,
      "target": "../images/img-p524-8.webp",
      "narrative": "draft",
      "content": "Paired forest plots (part of Figure 6) for one study, de Vos 2020: 'Cobas 6800/8800 Tb detection' has sensitivity 0.92 [0.88, 0.95] and specificity 0.96 [0.92, 0.98], and the comparator has sensitivity 0.90 [0.85, 0.94] and specificity 0.95 [0.90, 0.98].",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-524",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 524",
      "pageStart": 524,
      "pageEnd": 524,
      "target": "../sections/page-524.md",
      "narrative": null,
      "content": "BD Max \n \n \nCobas MTB \n \n \n \nFigure 6: Head- to head comparisons of cNATs and Xpert for TB detection \n \nIn the study by Wang 2016, a lower specificity was observed for both Abbott RealTime MTB assay \n(84%) and the Xpert assay (90%) for pulmonary TB specimens. \nObasanya 2017 used Xpert in the study for pulmonary specimens. The study reported lower sensitivity \nof 79% for Xpert in comparison to the FluoroType (sensitivity of 89%), while the specificity of the Xpert \nFluoroType MTB \nXpert \nXpert",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents head-to-head comparisons with Xpert for TB detection, noting low specificity for both Abbott RealTime and Xpert in one study and differing results for FluoroType and Xpert in another.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p525-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 525,
      "pageEnd": 525,
      "target": "../images/img-p525-1.webp",
      "narrative": "draft",
      "content": "Stacked horizontal bar chart (Figure 7) of QUADAS-2 judgements across 18 studies, showing the proportion rated high (red), unclear (yellow) or low (green) for risk of bias and applicability concerns by domain. For risk of bias, patient selection is 2 high, 6 unclear and 10 low; index test is 18 low; reference standard is 1 high, 3 unclear and 14 low; flow and timing is 1 high, 1 unclear and 16 low. For applicability, patient selection is 6 high, 3 unclear and 9 low, and index test and reference standard are 18 low.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-525",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 525",
      "pageStart": 525,
      "pageEnd": 525,
      "target": "../sections/page-525.md",
      "narrative": null,
      "content": "was substantially improved (94% versus 60% for the FluoroType). Considering the possible reasons \nmentioned about for the low specificity of the FluoroType, and from the head-to-head comparison results, \nit is possible that the manual DNA extraction method used in this study for performing FluoroType MTB \nassay could have led to substantially high cross-contamination than a fully automated Xpert assay. \nFor Paradis 2018, Xpert was also performed on the sputum specimens. It was observed that the specificity \nwas low (73%) with Xpert as well. This suggests that the reasons for the false positive",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Discusses possible cross-contamination from manual extraction as a reason for low FluoroType specificity, and reports pooled rifampicin resistance detection with sensitivity of 96.7% and specificity of 98.9%.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p526-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 526,
      "pageEnd": 526,
      "target": "../images/img-p526-1.webp",
      "narrative": "draft",
      "content": "Forest plot (Figure 8) of sensitivity and specificity for rifampicin resistance detection with phenotypic DST as reference standard for 19 studies, listing TP, FP, FN and TN counts with sensitivity and specificity (95% CI) and plots of each estimate on a 0–1 scale. Sensitivity is mostly 0.88–1.00 (not estimable for Cieslelczuk 2020 and Luukinen 2020) and specificity is 0.95–1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p526-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 526,
      "pageEnd": 526,
      "target": "../images/img-p526-2.webp",
      "narrative": "draft",
      "content": "Forest plot (Figure 9) of sensitivity and specificity for isoniazid resistance detection with phenotypic DST as reference standard for 19 studies, listing TP, FP, FN and TN counts with sensitivity and specificity (95% CI) and plots of each estimate on a 0–1 scale. Studies are ordered by a column headed '%RIF resistant' (56.85 to 0.0); sensitivity ranges from 0.58 (Hofmann-Thiel 2020) to 1.00, while specificity is 0.94–1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-526",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 526",
      "pageStart": 526,
      "pageEnd": 526,
      "target": "../sections/page-526.md",
      "narrative": null,
      "content": "Figure 8: Forest plot for rifampicin resistance detection with phenotypic DST as the reference standard \n \n \n3. PICO question: Among people with bacteriologically-confirmed PTB (adults, PLHIV, children) \nshould cNAT assays on sputum be used to detect isoniazid resistance using phenotypic DST as \nreference standard ? \n \nIsoniazid resistance \n \nThe overall sensitivity for rifampicin resistance in these 18 studies ranged from 58% to 100% and the \nspecificity ranged from 94% to 100%. \nThe pooled sensitivity (95% CI) was 86.4 (82.8 to 89.3) and the pooled specificity was 99.2 (98.1 to \n99.7). Table",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Covers isoniazid resistance detection against phenotypic drug susceptibility testing, reporting pooled sensitivity of 86.4% and high specificity, with forest plot captions.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p527-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 527,
      "pageEnd": 527,
      "target": "../images/img-p527-1.webp",
      "narrative": "draft",
      "content": "Forest plot (Figure 10) of sensitivity and specificity for rifampicin resistance detection with a composite reference standard for 9 studies, listing TP, FP, FN and TN counts with sensitivity and specificity (95% CI) and plots of each estimate on a 0–1 scale. Sensitivity ranges from 0.92 to 1.00 and specificity from 0.95 to 1.00; de Vos 2020 has a very wide specificity interval (0.03 to 1.00).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p527-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 527,
      "pageEnd": 527,
      "target": "../images/img-p527-2.webp",
      "narrative": "draft",
      "content": "Forest plot (Figure 11) of sensitivity and specificity for isoniazid resistance detection with a composite reference standard for 8 studies, listing TP, FP, FN and TN counts with sensitivity and specificity (95% CI) and plots of each estimate on a 0–1 scale. Sensitivity ranges from 0.58 (Hofmann-Thiel 2020) to 0.92 and specificity from 0.99 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-527",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 527",
      "pageStart": 527,
      "pageEnd": 527,
      "target": "../sections/page-527.md",
      "narrative": null,
      "content": "Pooled Analysis: Composite reference standard \n \n4. PICO question: Among people with bacteriologically-confirmed PTB (adults, PLHIV, children) \nshould cNAT assays on sputum be used to detect rifampicin resistance using composite reference \nstandard ? \n \nRifampicin resistance \n \nThe overall sensitivity for rifampicin resistance in these 9 studies ranged from 92% to 100% and the \nspecificity ranged from 95% to 100%. \nThe pooled sensitivity (95% CI) was 96.7 (92.9 to 98.5) and the pooled specificity was 98.7 (97.1 to \n99.4). Table 3b (Figure 10) provides the overall pooled estimates for all tests",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports rifampicin and isoniazid resistance detection against a composite reference standard, with pooled rifampicin sensitivity of 96.7% and specificity of 98.7%.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-528",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 528",
      "pageStart": 528,
      "pageEnd": 528,
      "target": "../sections/page-528.md",
      "narrative": null,
      "content": "Sensitivity analysis: Manufacturer driven \nThere were three studies which were manufacturer driven in this dataset for TB detection. We performed \na sensitivity analysis excluding these studies driven by manufacturers to assess whether this would affect \nthe results. We excluded Kostera 2016, Kostera 2018 and Paradis 2018. \nThe pooled estimate for this sensitivity analysis for rifampicin was sensitivity of 97.6% (93.3 to 99.2), \nspecificity of 98.6% (97.0 to 99.4). Similarly, for isoniazid the sensitivity and specificity was 85.7% \n(81.3 to 89.2) and 99.4% (98.6 to 99.7) respectively. Table 4",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports a sensitivity analysis excluding three manufacturer-driven studies for resistance detection, with similar results, followed by acknowledgements.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-529",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 529",
      "pageStart": 529,
      "pageEnd": 529,
      "target": "../sections/page-529.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Do not distribute further. \n524 \n \n \n \n \nTABLES \n \n \nTable 1: Number of included studies for each index test \n \n \nINDEX TEST \nPURPOSE \nNUMBER OF STUDIES \nABBOTT REALTIME MTB \nTB detection \n12 \nABBOTT REALTIME RIF/INH \nResistance detection \n9 \nFLUOROTYPE MTB \nTB detection \n6 \nFLUROTYPE MTBDR \nTB detection \n4 \nFLUROTYPE MTBDR \nResistance detection \n3 \nBD MAX MDR-TB \nTB detection \n5 \nBD MAX MDR-TB \nResistance detection \n5 \nCOBAS 6800/8800 MTB \nTB detection \n2 \nCOBAS 6800/8800 MTB-DR \nResistance detection \n2",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The first table of the review, giving the number of included studies for each index test and purpose, marked as prepared for the WHO Global TB Programme and not for further distribution.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-530",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 530",
      "pageStart": 530,
      "pageEnd": 530,
      "target": "../sections/page-530.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Do not distribute further. \n525 \n \n \n \n \n \nTable 2a: Diagnostic accuracy of each index test- TB detection \n \nParameter \nPoint estimate \nCI \nTB+ \nTB - \nOverall (N = 29) \nOverall sens \n93.0 \n90.9 to 94.7 \n4767 \n9085 \nOverall spec \n97.7 \n95.6 to 98.8 \nAbbott RealTime MTB (N= 12) \nPooled sens \n96.1 \n89.9 to 98.5 \n1274 \n4162 \nPooled spec \n97.6 \n94.8 to 98.9 \nFluoroType MTB (N= 6) \nPooled sens \n91.3 \n89.0  to 93.0 \n755 \n2502 \nPooled spec \n98.4 \n91.9 to 99.7 \nFluoroType MTBDR (N= 4) \nPooled sens \n91.7 \n86.2 to 95.1 \n976 \n677 \nPooled spec",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Tabulates pooled accuracy of centralized nucleic acid assays for detecting tuberculosis across 29 studies, overall about 93% sensitivity and 98% specificity, with results for each assay such as Abbott RealTime MTB, FluoroType and BD Max. The page carries a note that the report was prepared for WHO's Global TB Programme and is not for further distribution.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-019",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 2a: Diagnostic accuracy of each index test- TB detection",
      "pageStart": 530,
      "pageEnd": 530,
      "target": "../tables/table-2a.csv",
      "narrative": "draft",
      "content": "Diagnostic accuracy of moderate-complexity automated NAATs for TB detection, overall and for each index test (Abbott RealTime MTB, FluoroType MTB, FluoroType MTBDR, BD Max MTB and Roche Cobas MTB). For each, the table gives sensitivity and specificity point estimates with confidence intervals and the numbers of TB-positive and TB-negative specimens.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/table-020",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 2b: Diagnostic accuracy for TB detection by smear status",
      "pageStart": 530,
      "pageEnd": 531,
      "target": "../tables/table-2b.csv",
      "narrative": "draft",
      "content": "Diagnostic accuracy of moderate-complexity automated NAATs for TB detection by smear status, overall and for each index test. Columns give point estimates and confidence intervals for smear-positive and smear-negative specimens; specificity is not reported for smear-positive specimens.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-531",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 531",
      "pageStart": 531,
      "pageEnd": 531,
      "target": "../sections/page-531.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Do not distribute further. \n526 \n \n \n \n \n \n \n \nParameter \nPoint estimate \nCI \nPoint estimate \nCI \nSmear positive (N = 26; 2880 specimens) \nSmear negative (N = 26; 9947 specimens) \nOverall sens \n98.8% \n93.7 to 99.7 \n86.1% \n73.4 to 93.2 \nOverall spec \n- \n- \n96.4% \n93.6 to 98.0 \nAbbott RealTime MTB(N= 11) \nPooled sens \n99.7% \n97.3 to 99.9 \n85.7% \n65.5 to 95.0 \nPooled spec \n- \n- \n99.2% \n96.4 to 99.1 \nFluoroType MTB (N= 4) \nPooled sens \n87.0% \n83 to 91 \n84.0% \n80 to 89 \nPooled spec \n- \n- \n97.0% \n96 to 98 \nFluoroType MTBDR (N= 4) \nPooled",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Tabulates the accuracy of the centralized assays for tuberculosis detection by smear status, with near-complete sensitivity in smear-positive and about 86% in smear-negative specimens, broken down by assay.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-532",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 532",
      "pageStart": 532,
      "pageEnd": 532,
      "target": "../sections/page-532.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Do not distribute further. \n527 \n \n \n \n \nTable 3a: Diagnostic accuracy of each index test - resistance detection with phenotypic DST as reference standard \n \n \nOverall Pooled RIF (N = 18) \nOverall Pooled INH (N = 18) \nPooled Sens \n96.7 (93.1 to 98.4) \nPooled sens \n86.4  (82.8 to 89.3) \nPooled Spec \n98.9 (97.5 to 99.5) \nPooled spec \n99.2 (98.1 to 99.7) \nAbbott (RIF) N = 9 \nAbbott (INH) N = 9 \nPooled sens \n94% (91 to 96) \nPooled sens \n89% (86 to 92) \nPooled spec \n99% (99 to 100) \nPooled spec \n98% (98 to 100) \nFluoroType MTBDR (RIF) N",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Tabulates the accuracy of each centralized assay for detecting rifampicin and isoniazid resistance against phenotypic drug susceptibility testing, with pooled sensitivity of about 97% for rifampicin and 86% for isoniazid and high specificity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-021",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 3a: Diagnostic accuracy of each index test - resistance detection with phenotypic DST as reference standard",
      "pageStart": 532,
      "pageEnd": 532,
      "target": "../tables/table-3a.csv",
      "narrative": "draft",
      "content": "Diagnostic accuracy of moderate-complexity automated NAATs for detecting rifampicin (left) and isoniazid (right) resistance against phenotypic DST, overall and for each test (Abbott, FluoroType MTBDR, BD Max MDR-TB and Roche Cobas MTB RIF/INH). Each block gives pooled or study-level sensitivity and specificity with 95% CIs.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-533",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 533",
      "pageStart": 533,
      "pageEnd": 533,
      "target": "../sections/page-533.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Do not distribute further. \n528 \n \n \n \nTable 3b : Diagnostic accuracy of each index test - resistance detection with composite reference standard \n \n \nOverall Pooled RIF (N = 9) \nOverall Pooled INH (N = 8) \nPooled sens \n96.7 (92.9 to 98.5) \nPooled sens \n86.4  (82.1 to 89.8) \nPooled spec \n98.7 (97.1 to 99.4) \nPooled spec \n99.8 (98.3 to 99.8) \n \nData are presented as point estimate with 95% CI. Sens: sensitivity; spec: specificity \n \n \n \nTable 4: Sensitivity analysis \n \n \nParameter \nPoint estimate (95% CI) \nTB detection (N = 25) \nOve",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Tabulates resistance detection accuracy against a composite reference standard and gives a sensitivity analysis for tuberculosis, rifampicin and isoniazid resistance detection.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-022",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 3b : Diagnostic accuracy of each index test - resistance detection with composite reference standard",
      "pageStart": 533,
      "pageEnd": 533,
      "target": "../tables/table-3b.csv",
      "narrative": "draft",
      "content": "Overall pooled sensitivity and specificity (95% CI) of moderate-complexity automated NAATs for detecting rifampicin and isoniazid resistance against a composite reference standard.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-534",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 534",
      "pageStart": 534,
      "pageEnd": 534,
      "target": "../sections/page-534.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Do not distribute further. \n529 \n \n \n \n \nTable 5: Summary of findings: TB detection \n \nWhat is the diagnostic accuracy of cNATs for TB detection? \nPatient or population: Participants with presumptive pulmonary TB \nSetting: outpatient and inpatient \nReference standard: culture \nPooled sensitivity: 0.93 (95% CI: 0.91 to 0.95) Pooled specificity: 0.98 (95% CI: 0.96 to 0.99) \n \nTest result \nNumber of results per 1,000 patients tested (95% CrI) \nNumber of \nCertainty of the \nPrevalence 2.5% \nPrevalence 10% \nPrevalence 30% \nparticipants",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A summary-of-findings table on the accuracy of centralized nucleic acid tests for tuberculosis detection, showing expected true and false results per 1000 people at different prevalence levels with GRADE certainty and reasons for downgrading.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-535",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 535",
      "pageStart": 535,
      "pageEnd": 535,
      "target": "../sections/page-535.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Do not distribute further. \n530 \n \n \n \n \nTable 6:  Summary of findings: cNATs  for rifampicin resistance \n \nWhat is the diagnostic accuracy of cNATs for RIF resistance detection? \nPatient or population: Participants with presumptive MDR-TB \nSetting: outpatient and inpatient \nReference standard: culture \nPooled sensitivity: 0.97 (95% CI: 0.93 to 0.98) | Pooled specificity : 0.99 (95% CI: 0.97 to 0.99) \n \n \nTest result \nNumber of results per 1,000 patients tested (95% CI) \nNumber of \nparticipants \n(studies) \nCertainty of the \nEvidenc",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A summary-of-findings table on centralized tests for detecting rifampicin resistance, giving expected results per 1000 people tested at several prevalence levels and the GRADE certainty of evidence.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-536",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 536",
      "pageStart": 536,
      "pageEnd": 536,
      "target": "../sections/page-536.md",
      "narrative": null,
      "content": "531 \n \n \n \n \nTable 7:  Summary of findings: cNATs assay for isoniazid resistance \n \nWhat is the diagnostic accuracy of cNATs assay for INH resistance detection? \nPatient or population: Participants with presumptive MDR-TB \nSetting: outpatient and inpatient \nReference standard: culture \nPooled sensitivity: 0.86 (95% CI: 0.83 to 0.89) | Pooled specificity : 0.99 (95% CI: 0.98 to 1.00) \n \n \nTest result \nNumber of results per 1,000 patients tested (95% CrI) \nNumber of \nparticipants \n(studies) \nCertainty of the \nEvidence (GRADE) \nPrevalence 2% \nTypically seen in \nPrevalence 10% \nTypically seen in",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A summary-of-findings table on centralized tests for detecting isoniazid resistance, with pooled sensitivity of 0.86 and specificity of 0.99, results per 1000 people and notes on GRADE judgements.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-537",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 537",
      "pageStart": 537,
      "pageEnd": 537,
      "target": "../sections/page-537.md",
      "narrative": null,
      "content": "532 \n \n \n \n \nTable 8:  Summary of findings table: cNATs for RIF resistance: CRS \n \nWhat is the diagnostic accuracy of cNATs for RIF resistance detection? \nPatient or population: Participants with presumptive pulmonary TB \nSetting: outpatient and inpatient \nReference standard: CRS: phenotypic DST (solid or liquid) and gene sequencing, where at least one test is resistant \nPooled sensitivity : 0.97 (95% CI: 0.93 to 0.98) | Pooled specificity : 0.99 (95% CI: 0.97 to 0.99) \n \nTest result \nNumber of results per 1,000 patients tested (95% CrI) \nNumber of \nPrevalence 2% \nTypically seen in \nPrevalence",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A summary-of-findings table on rifampicin resistance detection against a composite reference standard of phenotypic testing and sequencing, with results per 1000 people and GRADE certainty.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-538",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 538",
      "pageStart": 538,
      "pageEnd": 538,
      "target": "../sections/page-538.md",
      "narrative": null,
      "content": "533 \n \n \n \n \nTable 9:  Summary of findings table: cNATs for INH resistance: CRS \n \nWhat is the diagnostic accuracy of cNATs for INH resistance detection? \nPatient or population: Participants with presumptive pulmonary TB \nSetting: outpatient and inpatient \nReference standard: CRS: phenotypic DST (solid or liquid) and gene sequencing, where at least one test is resistant \nPooled sensitivity : 0.86 (95% CI: 0.82 to 0.90) | Pooled specificity : 1.00 (95% CI: 0.98 to 1.00) \n \nTest result \nNumber of results per 1,000 patients tested (95% CrI) \nNumber of \nPrevalence 2% \nTypically seen in \nPrevalence",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A summary-of-findings table on isoniazid resistance detection against a composite reference standard, with results per 1000 people and GRADE certainty.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-539",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 539",
      "pageStart": 539,
      "pageEnd": 539,
      "target": "../sections/page-539.md",
      "narrative": null,
      "content": "Figures \n \n \n \n \n \n \n \n \n \n \n \nFigure 1. PRISMA diagram of studies included in the review \nPotentially relevant \ncitations identified from \nelectronic databases: \n878 \nExcluded Screen One: 638 \nReason: Not relevant based \non assessment of title and \nabstract \nFull papers retrieved for \nmore detailed evaluation: \n133 \nExcluded Screen Two: 57 \nReasons: \nInappropriate index test: 48 \nAbstract: 4 \nNot pulmonary specimens: 2 \nCase report: 1 \nCase-control: 1 \nNot diagnostic accuracy study: 1 \nExcluded for non-extractable data with no response \nfrom authors: 1 \nPapers included in the \nsystematic revi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A PRISMA flow diagram showing 878 citations screened, 133 full papers reviewed and 36 studies (39 datasets) included, broken down by assay.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p540-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 540,
      "pageEnd": 540,
      "target": "../images/img-p540-1.webp",
      "narrative": "draft",
      "content": "Risk-of-bias and applicability summary (Figure A) for 29 studies, from Berhanu 2018 to Wang 2016, with a coloured circle per QUADAS-2 domain: red minus for high, yellow question mark for unclear and green plus for low. Index test is rated low in every study. High or unclear ratings occur mainly for patient selection, in both risk of bias and applicability.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-540",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 540",
      "pageStart": 540,
      "pageEnd": 540,
      "target": "../sections/page-540.md",
      "narrative": null,
      "content": "Appendix A: Quallity assessment \n \n \n \n \nFigure A. Risk of Bias and Applicability Concerns summary for QUADAS-2 domains in each study \nevaluating assays for TB detection",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The caption for a figure summarizing QUADAS-2 risk of bias and applicability concerns for studies of assays for tuberculosis detection.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p541-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 541,
      "pageEnd": 541,
      "target": "../images/img-p541-1.webp",
      "narrative": "draft",
      "content": "Risk-of-bias and applicability summary (Figure B) for 18 studies, from Berhanu 2018 to Tam 2017, with a coloured circle per QUADAS-2 domain: red minus for high, yellow question mark for unclear and green plus for low. Index test is rated low in every study. High or unclear ratings occur mainly for patient selection, in both risk of bias and applicability.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-541",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 541",
      "pageStart": 541,
      "pageEnd": 541,
      "target": "../sections/page-541.md",
      "narrative": null,
      "content": "Figure B: Risk of Bias and Applicability Concerns summary for QUADAS-2 domains in each study \nevaluating assays for resistance detection",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The caption for a figure summarizing QUADAS-2 risk of bias and applicability concerns for studies of resistance detection.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-542",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 542",
      "pageStart": 542,
      "pageEnd": 542,
      "target": "../sections/page-542.md",
      "narrative": null,
      "content": "Appendix B. QUADAS-2 Protocol \n \nDomain 1 Patient Selection: \nRisk of Bias: Could the selection of patients have introduced bias? \n• Signaling question 1: Was a consecutive or random sample of patients or specimens enrolled? \no We scored ‘yes’ if the study enrolled a consecutive or random sample of eligible \npatients; ‘no’ if the study selected patients by convenience, and ‘unclear’ if the study did \nnot report the manner of patient selection or this cannot be discerned. \n \n• \nSignaling question 2: Was a case-control design avoided? \no We scored ‘yes’ if the study enrolled only patients presum",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Sets out the review's QUADAS-2 protocol, explaining how each signalling question on patient selection and the index test was scored. It notes that the centralized assays give automated results with prespecified thresholds, leaving no room for subjective interpretation.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-543",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 543",
      "pageStart": 543,
      "pageEnd": 543,
      "target": "../sections/page-543.md",
      "narrative": null,
      "content": "o As the threshold is prespecified in all centralized molecular DST assay in this review, \nwe answered this question \"yes\" for all studies. \n \n \n \nApplicability: Are there concerns that the index test, its conduct, or its interpretation differ from the \nreview question? Variations in test technology, execution, or interpretation may affect estimates of the \ndiagnostic accuracy of a test. \nWe judged ‘low concern’ if the test was done as per recommendation of the manufacturer for PTB \nspecimens. We judged ‘high concern’ it was stated and/or if additional steps were used for sample \npreparation a",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the QUADAS-2 protocol with criteria for applicability of the index test and for judging risk of bias in the reference standard, including blinding to index test results.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-544",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 544",
      "pageStart": 544,
      "pageEnd": 544,
      "target": "../sections/page-544.md",
      "narrative": null,
      "content": "o We added a signaling question for rifampicin resistance detection. We scored \"yes\" if \nthe reference test provided an automated result (for example, MGIT 960), blinding was \nexplicitly stated, or it was clear that the reference standard was performed at a separate \nlaboratory or performed by different people, or both. We scored \"no\" if the study stated \nthat the reference standard result was interpreted with knowledge of the index test result. \nWe scored \"unclear\" if we could not tell. \n \nApplicability: Are there concerns that the target condition as defined by the reference standard does no",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the QUADAS-2 protocol with an added question for rifampicin resistance on blinding of the reference standard, and criteria for applicability and for the flow and timing domain.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p545-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 545,
      "pageEnd": 545,
      "target": "../images/img-p545-1.webp",
      "narrative": "draft",
      "content": "Forest plot (Figure A) of TB detection by the Abbott RealTime MTB assay for 12 studies, listing TP, FP, FN and TN counts with sensitivity and specificity (95% CI) and plots on a 0–1 scale. Sensitivity ranges from 0.79 (Berhanu 2018) to 1.00 and specificity from 0.84 (Wang 2016) to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p545-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 545,
      "pageEnd": 545,
      "target": "../images/img-p545-2.webp",
      "narrative": "draft",
      "content": "Forest plot (Figure B) of TB detection by the FluoroType MTB assay for 6 studies, listing TP, FP, FN and TN counts with sensitivity and specificity (95% CI) and plots on a 0–1 scale. Sensitivity ranges from 0.87 to 0.95; specificity is 0.94–1.00 except Obasanya 2017 (0.60).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p545-3",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 545,
      "pageEnd": 545,
      "target": "../images/img-p545-3.webp",
      "narrative": "draft",
      "content": "Forest plot (Figure C) of TB detection by the FluoroType MTBDR assay for 4 studies, listing TP, FP, FN and TN counts with sensitivity and specificity (95% CI) and plots on a 0–1 scale. Sensitivity ranges from 0.85 to 0.96 and specificity from 0.97 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-545",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 545",
      "pageStart": 545,
      "pageEnd": 545,
      "target": "../sections/page-545.md",
      "narrative": null,
      "content": "analysis and there was no explanation given; and 'unclear ' if not enough information was given \nto assess whether participants were excluded from the analysis \n \nAppendix C: Data by different technologies \n \nTB detection \n1. Abbott RealTime MTB assay \n \n \nFigure  A: Forest plot of TB detection by Abbott RealTime MTB assay \n \n2. FluoroType MTB assay \n \n \n \nFigure  B: Forest plot of TB detection by FluoroType MTB assay \n \n3. FluoroType MTBDR assay \n \n \nFigure  C: Forest plot of TB detection by FluoroType MTBDR assay",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the QUADAS-2 protocol and begins Appendix C, which lists forest plots of tuberculosis detection for each assay.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p546-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 546,
      "pageEnd": 546,
      "target": "../images/img-p546-1.webp",
      "narrative": "draft",
      "content": "Forest plot of TB detection by the BD Max MDR-TB assay (Figure D) in four studies: Ciesielczuk 2020, Hofmann-Thiel 2020, Shah 2019 and Shaheed 2020. Each row gives a study's TP, FP, FN and TN counts and sensitivity and specificity with 95% CI, plotted as squares with CI lines on 0–1 scales. Sensitivity ranges from 0.87 to 0.95; specificity is 0.93–1.00 except 0.68 in Shaheed 2020.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p546-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 546,
      "pageEnd": 546,
      "target": "../images/img-p546-2.webp",
      "narrative": "draft",
      "content": "Forest plot of TB detection by the Cobas MTB assay (Figure E) in two studies, Cobas 2020 and de Vos 2020. Each row gives a study's TP, FP, FN and TN counts and sensitivity and specificity with 95% CI, plotted as squares with CI lines on 0–1 scales. Sensitivity is 0.95 and 0.92, and specificity 0.98 and 0.96.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p546-3",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 546,
      "pageEnd": 546,
      "target": "../images/img-p546-3.webp",
      "narrative": "draft",
      "content": "Forest plots of rifampicin (RIF) resistance detection by the Abbott RealTime RIF/INH assay, in three panels by reference standard: culture (9 studies), sequencing (3 studies) and composite reference standard (CRS, 3 studies). Each row gives a study's TP, FP, FN and TN counts and sensitivity and specificity with 95% CI, plotted as squares with CI lines on 0–1 scales. Estimable sensitivities range from 0.88 to 1.00 and specificities from 0.98 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-546",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 546",
      "pageStart": 546,
      "pageEnd": 546,
      "target": "../sections/page-546.md",
      "narrative": null,
      "content": "4. BD Max MDR-TB \n \n \nFigure  D: Forest plot of TB detection by BD Max MDR-TB assay \n \n \n5. Cobas 6800/8800 MTB assay \n \n \nFigure  E: Forest plot of TB detection by Cobas MTB assay \n \nResistance detection \n \n \n \n \nFigure: RIF resistance detection  by Abbott RealTIme RIF/INH resistance",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists further forest plot captions for tuberculosis detection by BD Max and Cobas assays and for rifampicin resistance detection by the Abbott assay.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p547-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 547,
      "pageEnd": 547,
      "target": "../images/img-p547-1.webp",
      "narrative": "draft",
      "content": "Forest plots of isoniazid (INH) resistance detection by the Abbott RealTime RIF/INH assay, in three panels by reference standard: culture (9 studies), sequencing (3 studies) and CRS (3 studies). Each row gives a study's TP, FP, FN and TN counts and sensitivity and specificity with 95% CI, plotted as squares with CI lines on 0–1 scales. Sensitivities range from 0.79 to 1.00 and specificities from 0.94 to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p547-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 547,
      "pageEnd": 547,
      "target": "../images/img-p547-2.webp",
      "narrative": "draft",
      "content": "Forest plot of rifampicin resistance detection by the FluoroType MTBDR assay in three studies: Haasis 2018, Hillemann 2018 and Shaheed 2020. Each row gives a study's TP, FP, FN and TN counts and sensitivity and specificity with 95% CI, plotted as squares with CI lines on 0–1 scales. Sensitivity is 0.97–1.00 and specificity 0.95–1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p547-3",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 547,
      "pageEnd": 547,
      "target": "../images/img-p547-3.webp",
      "narrative": "draft",
      "content": "Forest plot of isoniazid resistance detection by the FluoroType MTBDR assay in three studies: Haasis 2018, Hillemann 2018 and Shaheed 2020. Each row gives a study's TP, FP, FN and TN counts and sensitivity and specificity with 95% CI, plotted as squares with CI lines on 0–1 scales. Sensitivity is 0.70, 0.92 and 0.79, and specificity 0.99–1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-547",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 547",
      "pageStart": 547,
      "pageEnd": 547,
      "target": "../sections/page-547.md",
      "narrative": null,
      "content": "Figure: INH resistance detection  by Abbott RealTIme RIF/INH resistance \n \n \n \n \nFigure: RIF resistance detection  by FluroType MTBDR resistance \n \n \n \nFigure: INH resistance detection  by FluroType MTBDR resistance",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists forest plot captions for isoniazid and rifampicin resistance detection by the Abbott and FluoroType MTBDR assays.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p548-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 548,
      "pageEnd": 548,
      "target": "../images/img-p548-1.webp",
      "narrative": "draft",
      "content": "Forest plots of rifampicin and isoniazid resistance detection by the BD Max MDR-TB assay, each panel with four studies (Ciesielczuk 2020, Hofmann-Thiel 2020, Shah 2019, Shaheed 2020). Each row gives a study's TP, FP, FN and TN counts and sensitivity and specificity with 95% CI, plotted as squares with CI lines on 0–1 scales. For rifampicin, estimable sensitivity is 0.90–1.00 and specificity 0.95–0.98; for isoniazid, sensitivity is 0.58–1.00 and specificity 0.99–1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p548-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 548,
      "pageEnd": 548,
      "target": "../images/img-p548-2.webp",
      "narrative": "draft",
      "content": "Forest plots of rifampicin and isoniazid resistance detection by the Cobas MTB-DR assay (panels 'Cobas RIF' and 'Cobas INH'), each with two studies, de Vos 2020 and Scott 2020. Each row gives a study's TP, FP, FN and TN counts and sensitivity and specificity with 95% CI, plotted as squares with CI lines on 0–1 scales. Rifampicin sensitivity is 0.91 and 1.00, isoniazid sensitivity 0.90 and 0.80, and all specificities are 0.96–1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-548",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 548",
      "pageStart": 548,
      "pageEnd": 548,
      "target": "../sections/page-548.md",
      "narrative": null,
      "content": "Figure:  RIF and INH resistance detection  by BD Max MDR-TB assay \n \n \nFigure:  RIF and INH resistance detection  by Cobas  MTB-DR assay",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists forest plot captions for rifampicin and isoniazid resistance detection by the BD Max and Cobas assays.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-549",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 549",
      "pageStart": 549,
      "pageEnd": 549,
      "target": "../sections/page-549.md",
      "narrative": null,
      "content": "544 \n \n \n \n \nAppendix D \n \nComposite reference standard \nThree studies (Hofmann Thiel, 2018; Kostera 2016 and Tam 2017) evaluated sequencing and \nprovided information on the mutations to assess results using CRS for RIF resistance. In the study \nby Hofmann Thiel 2018, three specimens which were susceptible by phenotypic DST but resistant \nby Abbott, had a L511P mutation as confirmed by sequencing. As there is limited confidence that \nthis mutation is associated with resistance (Miotto 2017), we interpreted the composite reference \nstandard as sensitive and thus classified these results as fals",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Appendix D explains how the composite reference standard was applied in three studies with sequencing data, including how discordant results tied to particular mutations were classified. It also notes resistance mutations that some assays failed to detect.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-550",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 550",
      "pageStart": 550,
      "pageEnd": 550,
      "target": "../sections/page-550.md",
      "narrative": null,
      "content": "545 \n \n \n \n \nReferences \n \nAbbott 2019a   \nAbbott. Abbott RealTime MTB. https://www.molecular.abbott/int/en/products/infectious- \ndisease/realtime-mtb 2019. \n \nAbbott 2019b   \nAbbott. Abbott RealTime RIF/INH. https://www.molecular.abbott/int/en/products/infectious- \ndisease/realtime-mtb-rif-inh-resistance 2019. \n \nAlmeida Da Silva 2011   \nAlmeida Da Silva PE and Palomino JC. Molecular basis and mechanisms of drug resistance in \nMycobacterium tuberculosis: classical and new drugs. J Antimicrob Chemother \n2011;66(7):1417-1430. \n \nBruker-Hain Diagnostics 2019a   \nBruker-Hain Diagnostics. FluoroTy",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The start of the review's reference list, citing manufacturer product pages and studies on tuberculosis drug resistance and diagnosis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-551",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 551",
      "pageStart": 551,
      "pageEnd": 551,
      "target": "../sections/page-551.md",
      "narrative": null,
      "content": "546 \n \n \n \n \nChu H and Cole SR. Bivariate meta-analysis of sensitivity and specificity with sparse data: a \ngeneralized linear mixed model approach. J Clin Epidemiol 2006;59(12):1331-1332. \n \nCochrane 2008   \nThe Cochrane Collaboration. Cochrane Handbook for Systematic Reviews of Diagnostic Test \nAccuracy. 2008. \n \nDicks 2019   \nDicks KV and Stout JE. Molecular Diagnostics for Mycobacterium tuberculosis Infection \n[Annual Review of Medicine]. 2019;70(1):77-90. \n \nFriedrich 2011 \nFriedrich SO, Venter A, Kayigire XA, Dawson R, Donald PR, Diacon AH. Suitability of Xpert \nMTB/RIF and genotype MTBD",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Part of the review's reference list, citing statistical methods, the Cochrane diagnostic review handbook and studies of molecular tuberculosis diagnostics.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-552",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 552",
      "pageStart": 552,
      "pageEnd": 552,
      "target": "../sections/page-552.md",
      "narrative": null,
      "content": "547 \n \n \n \n \nNathavitharana 2017 \nNathavitharana RR, Cudahy PG, Schumacher SG, Steingart KR, Pai M, Denkinger CM. \nAccuracy of line probe assays for the diagnosis of pulmonary and multidrug-resistant \ntuberculosis: a systematic review and meta-analysis. The European respiratory journal \n2017;49(1):1601075. DOI: 10.1183/13993003.01075-2016 \n \nReitsma 2005   \nReitsma JB, Glas AS, Rutjes AW, Scholten RJ, Bossuyt PM, Zwinderman AH. Bivariate \nanalysis of sensitivity and specificity produces informative summary measures in diagnostic \nreviews. J Clin Epidemiol 2005;58(10):982-990. \n \nSchünemann 202",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Part of the review's reference list, citing meta-analyses of line probe assays, bivariate methods, GRADE papers and studies of false-positive Xpert results.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-553",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 553",
      "pageStart": 553,
      "pageEnd": 553,
      "target": "../sections/page-553.md",
      "narrative": null,
      "content": "548 \n \n \n \n \nin Patients With Previous Tuberculosis: Can We Distinguish True From False Positive Results?. \nClinical Infectious Diseases 2016;62 (8):995–1001. \n \nTheron 2016 \nTheron G, Venter R, Calligaro G, Smith L, Limberis J, Meldau R, et al. Xpert MTB/RIF Results \nin Patients With Previous Tuberculosis: Can We Distinguish True From False Positive Results?. \nClinical Infectious Diseases 2016;62 (8):995–1001. \n \nTheron 2018 \nTheron G, Venter R, Smith L, Esmail A, Randall P, Sood V, et al. False positive Xpert MTB/RIF \nresults in re-tested patients with previous tuberculosis: frequency, profi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The end of the review's reference list, including studies of false-positive Xpert results and WHO tuberculosis reports and guidelines.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-554",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 554",
      "pageStart": 554,
      "pageEnd": 554,
      "target": "../sections/page-554.md",
      "narrative": null,
      "content": "549 \n \n \n \n \n \n \n \n \n \n \n \nWeb Annex B.12. Systematic Literature Review of Economic \nEvidence for Nucleic acid amplification tests (NAATs) to detect TB \nand DR-TB in adults and children \n \n \n \nProduced in Preparation for the WHO guideline development group meeting Dec. 7-18, 2020 \n \n \nFinal Report: PLEASE DO NOT DISTRIBUTE \nDecember 2020 \n \nAlice Zwerling, Sazia Naz & Brianna Empringham \nSchool of Epidemiology & Public Health \nFaculty of Medicine, \nUniversity of Ottawa, Ottawa, Canada \nazwerlin@uottawa.ca",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title page of Web Annex B.12, a systematic review of economic evidence on nucleic acid amplification tests for detecting TB and drug-resistant TB, prepared by University of Ottawa authors for a December 2020 WHO guideline meeting and marked not for distribution.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-555",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 555",
      "pageStart": 555,
      "pageEnd": 555,
      "target": "../sections/page-555.md",
      "narrative": null,
      "content": "550 \n \n \n \n \nBACKGROUND \n \nIn preparation for the upcoming guideline development group meeting “Nucleic acid amplification tests \nto detect TB and DR-TB” scheduled for 7-11 December, 2020, there is a need to summarize the current \neconomic evidence on the following three classes of NAATs: \n \n1) Centralized assays that present end-to-end (E2E) solutions for detection of pulmonary TB \nand resistance to rifampicin and isoniazid; including the following commercially available \nplatforms: Abbott RealTime MTB, Abbott RealTime RIF/INH, FluoroType MTB, \nFluoroType MTDBR, BD Max MDR-TB assay, Roche Cob",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Sets out the background to the economic evidence review, which summarizes cost data on three classes of nucleic acid amplification tests for TB and drug-resistant TB, starting with centralized end-to-end platforms such as Abbott RealTime, FluoroType, BD Max and Roche Cobas. It describes the database search strategy and its cost-related search terms.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-556",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 556",
      "pageStart": 556,
      "pageEnd": 556,
      "target": "../sections/page-556.md",
      "narrative": null,
      "content": "551 \n \n \n \n \n \nDue to limited direct evidence, several subsequent searches were performed in the same 4 databases as \nof late November to identify indirect evidence using these platforms in non-TB disease areas. Search \nstrategies were run with the specific platform names but without the following search term for TB: \n(tuberculosis OR TB OR mycobacterium). Additional searches were also run with additional search terms \nfor Genotype MTBDRsl and MTBDRplus to look for indirect evidence from line probe assays to inform  \nthe evidence around PZA LPA. \n \nStudy selection & data extraction \n \nStudies",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains that because direct evidence was scarce, further searches looked for indirect cost evidence from the same platforms in other diseases and from line probe assays. It describes study selection and data extraction and an unpublished costing study that used time-and-motion observations in a South African laboratory to estimate unit test costs for BD Max and Hain.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-557",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 557",
      "pageStart": 557,
      "pageEnd": 557,
      "target": "../sections/page-557.md",
      "narrative": null,
      "content": "552 \n \n \n \n \nof staff conducting various activities to capture discrete time estimates that can be attributed to each \nactivity and used to construct personnel cost per activity. Those staff costs are then used in the bottom- \nup ingredients approach where component costs (staff, consumables, equipment, etc.) are added up to \ncreate a total unit cost. Several important simplifying assumptions were made that may limit \ngeneralizability of results, including assuming 50% of lab operations were dedicated to TB, a minimum \ndaily throughput of 24 samples/day or the equivalent of one BD MAX run (24",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains the bottom-up ingredients approach to unit test costing and the simplifying assumptions behind the estimates, such as half of laboratory work being for TB and a minimum daily throughput. A table breaks down BD Max costs by overhead, building, staff, equipment and test kits, at about US$ 18.50 per test in the main estimate.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p558-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 558,
      "pageEnd": 558,
      "target": "../images/img-p558-1.webp",
      "narrative": "draft",
      "content": "Stacked bar chart 'Per-test cost (BD)' (Figure 2A) showing per-test cost in 2019 USD for the primary estimate (PE), low and high scenarios of the sensitivity analysis, split into overhead, building, staff, equipment and test kits. Totals are roughly $18 (PE), $14 (low) and $41 (high); test kits are the largest component in the PE and low bars, while equipment makes up a large share of the high bar.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p558-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 558,
      "pageEnd": 558,
      "target": "../images/img-p558-2.webp",
      "narrative": "draft",
      "content": "Stacked bar chart 'Per-test cost (Hain)' (Figure 2B) showing per-test cost in 2019 USD for the primary estimate (PE), low and high scenarios, split into overhead, building, staff, equipment and test kits. Totals are roughly $15 (PE), $10 (low) and $37 (high); test kits dominate the PE and low bars, and equipment is the largest component of the high bar.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-558",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 558",
      "pageStart": 558,
      "pageEnd": 558,
      "target": "../sections/page-558.md",
      "narrative": null,
      "content": "553 \n \n \n \n \nType of test: \nHain \nResource Type \nPE \nLow \nHigh \nOverhead \n$1.95 \n$0.64 \n$5.00 \nBuilding \n$0.33 \n$0.05 \n$1.69 \nStaff \n$1.66 \n$0.69 \n$3.11 \nEquipment \n$2.43 \n$1.62 \n$15.60 \nTest Kits \n$9.00 \n$6.60 \n$12.00 \nTotal \n$15.37 \n$9.61 \n$37.40 \nData courtesy of FIND/H Sohn, unpublished. PE: Primary estimate \n \nThe cost of the test kit was the largest single cost contributor across both tests (~$4 less for Hain kit) \nhowever under conditions with higher annual overhead costs and fewer annual operational days and \noperational hours/day (‘HIGH’er GDP scenario) equipment costs became the larg",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives the cost breakdown for the Hain test, about US$ 15.40 per test in the main estimate, and notes that test kits were the largest cost except in a higher-cost scenario where equipment dominated. It adds that Hain's unit cost would probably fall further if run at full capacity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p559-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 559,
      "pageEnd": 559,
      "target": "../images/img-p559-1.webp",
      "narrative": "draft",
      "content": "Line chart 'Per-test cost by annual testing volume (BD vs. Hain)' (Figure 3), plotting per-test cost (2019 USD) against annual testing volume from 0 to 25,000 tests, with background bands shaded red below 5,000, yellow from 5,000 to 10,000 and green above 10,000. Both lines fall steeply at low volumes and flatten above about 5,000 tests; the Hain line levels off at roughly $12–13 and the BD line ends near 6,000 tests at about $18.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-559",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 559",
      "pageStart": 559,
      "pageEnd": 559,
      "target": "../sections/page-559.md",
      "narrative": null,
      "content": "554 \n \n \n \n \nIn one-way sensitivity analyses, annual testing volumes were varied from <5,000 to >25,000 tests/year \n(Figure 3). Per-test cost was highly sensitive to testing volume when fewer than 5000 tests were \nconducted per year, but unit test costs begin to stabilize between 5,000 and 10,000 tests /year, and \nabove 10,000 test per year, unit cost estimate was robust. When equipment can be multiplexed and \nused at capacity, per test cost can be minimized. \n \nFigure 3. One-way sensitivity analysis of per-test cost and annual testing volume. \n \n \nAdditional literature searches conducted to l",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that per-test cost was very sensitive to annual testing volume below 5000 tests and stabilized above about 10 000 tests. It notes that further searches found little additional data and that evidence on variation across sites and countries is limited.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-560",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 560",
      "pageStart": 560,
      "pageEnd": 560,
      "target": "../sections/page-560.md",
      "narrative": null,
      "content": "555 \n \n \n \n \nNo studies were identified that assessed cost-effectiveness analyses for any of the E2E solutions and \nextrapolation was not appropriate given differences in standard of care, different care cascades and \nassociated costs, operational conditions, testing volume and diagnostic accuracy. Implementation \nconsiderations such as test placement, lab network, and ability of program to initiate treatment quickly \nwill all likely impact unit test cost and cost-effectiveness. Economic modelling is needed across various \nsettings to understand the range cost-effectiveness profiles of E2E sol",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "States that no cost-effectiveness studies were found for the centralized end-to-end platforms and that extrapolating from other tests is not appropriate. It calls for economic modelling across settings and gives the same caution for the Xpert MTB/XDR test.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-561",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 561",
      "pageStart": 561,
      "pageEnd": 561,
      "target": "../sections/page-561.md",
      "narrative": null,
      "content": "556 \n \n \n \n \nSeveral factors that are likely influential on cost-effectiveness of Xpert MTB/XDR include diagnostic \naccuracy which may lead to more or fewer individuals being diagnosed compared with the standard of \ncare, and this will vary by region depending on what the standard of care is across different regions. \nAlong with diagnostic accuracy associated with the test itself, the diagnostic algorithm and placement of \nthe Xpert MTB/XDR test within that algorithm has important implications as well. \n \nThe novel Xpert MTB/XDR provides results in <90 minutes and therefore introduction of thi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Discusses factors likely to shape the cost-effectiveness of Xpert MTB/XDR, including its accuracy relative to the local standard of care, its place in the diagnostic algorithm and its results in under 90 minutes. It notes that total costs and budget impact will depend on testing volume, XDR prevalence and current practice.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-562",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 562",
      "pageStart": 562,
      "pageEnd": 562,
      "target": "../sections/page-562.md",
      "narrative": null,
      "content": "557 \n \n \n \n \n \nDirect costs related to cartridge and machinery are provided from the manufacturer while several \nimportant items related to resource use including staff time, overhead and operational costs associated \nwith implementing Xpert MTB/XDR have not been investigated. Differences in resource use between \nXpert MTB/XDR and existing approaches will vary across settings using different phenotypic and \ngenotypic DST. Important variability exists in costs of staff time and operational costs, such as testing \nvolume across settings. \n \n \nDoes the cost-effectiveness of the intervention favor",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes that manufacturer prices cover cartridges and equipment but not staff, overhead and operating costs of Xpert MTB/XDR, which vary by setting. It then turns to line probe assays for pyrazinamide resistance, citing a multi-country study comparing per-sample costs of line probe assays with other drug susceptibility methods.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-563",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 563",
      "pageStart": 563,
      "pageEnd": 563,
      "target": "../sections/page-563.md",
      "narrative": null,
      "content": "558 \n \n \n \n \nsample for LPA ranged across countries from $33.73 in Moldova to $100.50 in South Africa. LPAs \nbecame less expensive than PSQ if testing volume and batch size were maximized. Key drivers of cost \nincluded test volume, batch size and staff wages, with South African sites driving higher average test \ncosts due to low volume and high staff wages. \n \nLi et al described the performance and costs of the Genotype MTBDRplus line probe assay (LPA) among \ninpatients in China, and found conventional DST cost $50.72 per sample, while the LPA unit test cost was \nestimated to be $108.70(Li et",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that line probe assay costs varied widely between countries and were driven by test volume, batch size and staff wages, and cites a study from China where the assay cost more than conventional testing. It repeats that operating cost data are limited and vary across settings.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-564",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 564",
      "pageStart": 564,
      "pageEnd": 564,
      "target": "../sections/page-564.md",
      "narrative": null,
      "content": "559 \n \n \n \n \nNo cost-effectiveness studies were identified using the Genoscholar PZA-TB II. Extrapolation of cost- \neffectiveness data from other line probe assays is not advised due to differences in diagnostic accuracy, \nresistance prevalence, and the testing and treatment cascade of care. \n \n \nPRINCIPAL FINDINGS \n \n• \nOperational costs likely a large proportion of unit test costs and are largely unknown aside from \nequipment. These costs are also not always included in unit test costs making comparisons across \nestimates, methods and studies difficult. \n \n• \nBatch size and test volume are i",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "States that no cost-effectiveness studies were found for the Genoscholar PZA-TB II assay and lists the review's principal findings. These stress that operational costs are largely unknown, that batch size and volume drive unit costs, and that budget impact depends on who is tested, resistance prevalence and current practice.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-565",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 565",
      "pageStart": 565,
      "pageEnd": 565,
      "target": "../sections/page-565.md",
      "narrative": null,
      "content": "560 \n \n \n \n \nthe additional cost of treating more MDR and XDR patients in scenarios where sensitivity is improved, \nand more patients are initiating anti-TB treatment. \n \nCost-effectiveness studies providing direct evidence were not identified and extrapolation is not \nadvised. Cost-effectiveness may be impacted by changes in diagnostic accuracy, differences in standard \nof care across sites, diagnostic algorithm, reductions in turn-around-time among others. Without proper \neconomic modelling, potential cost-effectiveness of various tests and testing approaches remains \nunknown. \n \nCONCLUSION",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Concludes that direct economic evidence on these new TB diagnostics is very limited and that cost-effectiveness and transmission modelling across settings is needed.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-566",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 566",
      "pageStart": 566,
      "pageEnd": 566,
      "target": "../sections/page-566.md",
      "narrative": null,
      "content": "561 \n \n \n \n \nREFERENCES \n \nAbimbola, T. O., Marston, B. J., Date, A. A., Blandford, J. M., Sangrujee, N., & Wiktor, S. Z. (2012). Cost- \neffectiveness of tuberculosis diagnostic strategies to reduce early mortality among persons with \nadvanced HIV infection initiating antiretroviral therapy. Journal of Acquired Immune Deficiency \nSyndromes, 60(1), 1–7. https://doi.org/10.1097/QAI.0b013e318246538f \nAdelman, M. W., McFarland, D. A., Tsegaye, M., Aseffa, A., Kempker, R. R., & Blumberg, H. M. (2018). \nCost-effectiveness of WHO-Recommended Algorithms for TB Case Finding at Ethiopian HIV Clinics. \nO",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The start of the economic review's reference list, citing cost-effectiveness studies of TB diagnostic strategies, including in people with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-567",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 567",
      "pageStart": 567,
      "pageEnd": 567,
      "target": "../sections/page-567.md",
      "narrative": null,
      "content": "562 \n \n \n \n \nhttps://doi.org/10.1186/1471-2334-13-352 \nWhitfield, M. G., Marras, S. A. E., Warren, R. M., Van Rie, A., Rice, J., Wangh, L. J., & Kreiswirth, B. N. \n(2020). Rapid Pyrazinamide Drug Susceptibility Testing using a Closed-Tube PCR Assay of the Entire \npncA gene. Scientific Reports, 10(1). https://doi.org/10.1038/s41598-020-61286-7 \nZwerling, A. A., Sahu, M., Ngwira, L. G., Khundi, M., Harawa, T., Corbett, E. L., Chaisson, R. E., & Dowdy, \nD. W. (2015). Screening for tuberculosis among adults newly diagnosed with HIV in Sub-Saharan \nAfrica: A cost-effectiveness analysis. Journal of",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The end of the economic review's reference list, citing a pyrazinamide susceptibility assay study and a cost-effectiveness analysis of TB screening in people newly diagnosed with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p568-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 568,
      "pageEnd": 568,
      "target": "../images/img-p568-1.webp",
      "narrative": "draft",
      "content": "Manufacturer flyer titled '10-Color Multiplexing Technology: GeneXpert System Upgrade Options for HBDC Customers', with product photos above three numbered options: 1) convert current GeneXpert modules from 6-color to 10-color; 2) daisy-chain a 10-color satellite instrument to the current GeneXpert System; 3) order a new GeneXpert System with 10-color technology. Each option lists what must be ordered separately and computer requirements.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-568",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 568",
      "pageStart": 568,
      "pageEnd": 568,
      "target": "../sections/page-568.md",
      "narrative": null,
      "content": "563 \n \n \n \n \n \nSUPPLEMENT \n \nCepheid 10-Colour Multiplexing technology Upgrade Flyer with costs",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A supplement heading referring to a Cepheid flyer with costs for a 10-colour multiplexing upgrade.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p569-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 569,
      "pageEnd": 569,
      "target": "../images/img-p569-1.webp",
      "narrative": "draft",
      "content": "Ordering information table from the manufacturer flyer listing catalogue numbers, descriptions and prices (US dollars) for GeneXpert 10-color products, grouped as New Systems with Desktop, New Systems with Laptop, Satellite Instruments + DAISYKIT required, New Module and Accessories. Systems range from 1-module to 16-module configurations; the small print is only partly legible.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-569",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 569",
      "pageStart": 569,
      "pageEnd": 569,
      "target": "../sections/page-569.md",
      "narrative": null,
      "content": "564",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A near-blank page showing only a page number.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-570",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 570",
      "pageStart": 570,
      "pageEnd": 570,
      "target": "../sections/page-570.md",
      "narrative": null,
      "content": "565 \n \n \n \n \n \n \n \n \n \n \nWeb Annex B.13. User perspectives on NAATs to detect TB and \nresistance to anti-TB agents: results from qualitative evidence \nsynthesis (systematic review) \n \n \n \n \nNora Engel1, Eleanor A Ochodo2,3, Perpetua Wanjiku Karanja4, Bey-Marrié Schmidt5, Ricky Janssen1, \nKaren R Steingart6, Sandy Oliver,7,8 \n \nAffiliations: \n1Department of Health, Ethics and Society, Maastricht University, Maastricht, Netherlands \n(Kingdom of the) \n2Centre for Evidence-based Health Care, Department of Global Health, Faculty of Medicine and Health \nSciences, Stellenbosch University, Cape Town,",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title page of Web Annex B.13, a qualitative evidence synthesis of user perspectives on nucleic acid amplification tests for TB and drug resistance, listing the authors and their affiliations and noting it was prepared for a December 2020 WHO guideline meeting.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-571",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 571",
      "pageStart": 571,
      "pageEnd": 571,
      "target": "../sections/page-571.md",
      "narrative": null,
      "content": "566 \n \n \n \n \nBackground \n \nDescription of the topic \n \nTuberculosis (TB) causes 10 million cases and 1.5 million deaths annually and it is estimated that 3 million \ncases go undiagnosed each year (World Health Organization, 2020a). Drug-resistant TB (DR-TB) is a major \nthreat to global TB control. Ending the global TB epidemic will be achievable over the next 20 years only \nif there is intensive action by all countries which have endorsed the End TB Strategy and its ambitious \ntargets (World Health, 2015). Early diagnosis and prompt treatment of all persons of all ages with any form \nof drug-s",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Sets out the background to the qualitative review, noting TB's annual toll of 10 million cases and 1.5 million deaths, the threat of drug-resistant TB and the need for early diagnosis under the End TB Strategy. It explains that user perspectives on diagnostic technologies have rarely been studied and never systematically reviewed for a group of tests.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-572",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 572",
      "pageStart": 572,
      "pageEnd": 572,
      "target": "../sections/page-572.md",
      "narrative": null,
      "content": "567 \n \n \n \n \nreviewed for a group of technologies. Current WHO guidance on automated NAATs of low complexity for \nTB diagnosis is based on systematic reviews of diagnostic accuracy and cost-effectiveness. Accuracy \nstudies do not reveal what users think of or experience with the diagnostic in question. Yet to understand \nwhy diagnostics are utilized, how effective they are and their impact on health equity, it is essential to \nanswer questions around feasibility, added value and experiences which our review findings will provide. \n \nHow the intervention might work \n \nThe promise of automated N",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Argues that accuracy and cost reviews do not reveal users' experiences, which are needed to understand uptake, effectiveness and equity. It explains how automated low-complexity tests might work and lists the classes of tests under review for the WHO guideline meeting.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-573",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 573",
      "pageStart": 573,
      "pageEnd": 573,
      "target": "../sections/page-573.md",
      "narrative": null,
      "content": "568 \n \n \n \n \n• \nIndex test 2: Moderate-complexity automated NAATs for detection of TB and resistance to \nrifampicin and isoniazid; \n• \nIndex test 3: High-complexity hybridization-based NAATs for detection of resistance to \npyrazinamide. \n \nA qualitative evidence synthesis can add value by providing decision makers with additional evidence to \nimprove understanding of intervention complexity, contextual variations, implementation, and stakeholder \npreferences and experiences. \n \nObjectives \n \nTo synthesize user perspectives and experiences with automated NAATs of low complexity for detection \no",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the list of test classes reviewed and explains what a qualitative evidence synthesis adds for decision makers. It states the review's objective and the types of studies eligible, such as operational research and feasibility and acceptability assessments.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-574",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 574",
      "pageStart": 574,
      "pageEnd": 574,
      "target": "../sections/page-574.md",
      "narrative": null,
      "content": "569 \n \n \n \n \nThis review focuses on users and potential users of automated NAATs of low complexity. Users include \npatients and their caregivers, laboratory technicians, healthcare providers, implementers and programme \nofficers who are involved in diagnosing and treating tuberculosis and drug resistant forms of TB as well as \nordering, operating, maintaining diagnostics and acting on diagnostic test results. Potential users include \nusers who do not (yet) utilize the diagnostic for instance because they are unable to access it or make it \nwork within their routines or setting. \nSetting \nWe in",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Defines the users covered by the review, from patients and caregivers to laboratory staff, providers and programme officers, as well as potential users who cannot yet access or use the tests. It describes the settings included, the search and the screening of studies by two authors.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-575",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 575",
      "pageStart": 575,
      "pageEnd": 575,
      "target": "../sections/page-575.md",
      "narrative": null,
      "content": "570 \n \n \n \n \nassessment and their reasons for exclusion in the Characteristics of excluded studies table (appendix 5). We \nillustrated the study selection process in a PRISMA diagram (Moher, Liberati, Tetzlaff, & Altman, 2009). \n \nLanguage translation \n \nWe searched for primary studies irrespective of their language of publication. \n \nSampling of studies \n \nThis qualitative evidence synthesis aims to describe the experiences of those using automated NAATs of \nlow complexity for diagnosis of tuberculosis and DR-TB in a coherent way. After we identified all studies \nthat were eligible for inclus",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes how excluded studies were recorded, that studies in any language were sought, and how eligible studies were sampled for richness and relevance. It lists the factors the authors expected to shape user experiences, such as workflow, resources, confidence in results and equity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-576",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 576",
      "pageStart": 576,
      "pageEnd": 576,
      "target": "../sections/page-576.md",
      "narrative": null,
      "content": "571 \n \n \n \n \nTwo review authors individually extracted data from the same study and resolved any conflicts in a \nconsensus meeting. Authors of primary studies did not extract data from their own study or studies. Instead, \nanother author extracted these data. \n \nAssessing the methodological limitations of included studies \n \nTwo review authors [any pair from NE, BS, PW, EO] independently assessed methodological limitations \nfor each study using the EPPI-Centre tool (Rees, Caird, Dickson, Vigurs, & Thomas, 2014). We resolved \ndisagreements by discussion. Team members who are authors of included",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes data extraction by two authors, assessment of methodological limitations with the EPPI-Centre tool, and safeguards so that authors did not assess their own studies. It explains the thematic approach used to draw out values, feasibility and acceptability considerations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-577",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 577",
      "pageStart": 577,
      "pageEnd": 577,
      "target": "../sections/page-577.md",
      "narrative": null,
      "content": "572 \n \n \n \n \nBased on the key findings extracted from the six rich studies, NE developed a coding scheme which was \ndiscussed with the other review authors. Using the coding scheme, NE coded the key findings of the 6 \nstudies with rich data using NVIVO (version 12) and wrote memos on selected themes. In a second round \nof analysis, data from the 15 studies of medium richness was extracted by NE, EO, PW and BS; and NE \ncoded these summaries in the same way as the 6 rich studies. NE then added the emerging additional \ninsights and data to the existing memos. In a next step, NE generated review f",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes how the analysis proceeded, coding six rich studies first and then adding data from 15 studies of medium richness using NVivo. It explains how review findings were generated and begins a reflexivity statement on the authors' prior views about the tests.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-578",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 578",
      "pageStart": 578,
      "pageEnd": 578,
      "target": "../sections/page-578.md",
      "narrative": null,
      "content": "573 \n \n \n \n \nexist to their implementation, while others might be more hesitant about the presumed automatic benefit of \nintroducing advanced technologies but then not investing in strengthening weak health systems, while again \nsome might wonder about the design process of automated NAATs of low complexity and to what extent \nit took into account perspectives of various users. All would have been in contact with different types of \nusers throughout their research career. We minimized the risk that our perspectives as authors influence the \nanalysis and interpretation, by using refutational an",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reflexivity statement, describing the range of prior views among the authors and the techniques used to limit their influence on the analysis, followed by short biographies of the review team.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-579",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 579",
      "pageStart": 579,
      "pageEnd": 579,
      "target": "../sections/page-579.md",
      "narrative": null,
      "content": "574 \n \n \n \n \nResults \n \nResults of the search \n \nWe found 27 studies that met our inclusion criteria. We sampled 21 of these studies for inclusion in the \nanalysis (see Figure 1: Flow chart Appendix 4). All of the sampled studies were published between 2012 \nand 2020. Six studies remain to be reviewed which were not sampled as we assessed them as particularly \nthin (see section sampling). They do not address participants not already addressed in the sampled studies. \nFor an overview of the studies that were not sampled and studies that were excluded, see Appendix 5. \n \nDescription of the studi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Opens the results of a qualitative evidence synthesis, prepared for the WHO Global TB Programme, on user perspectives of low-complexity automated nucleic acid amplification tests (NAATs) for TB. It reports that 27 studies met the inclusion criteria, of which 21 published between 2012 and 2020 were sampled, all from high-TB-burden countries such as South Africa, India and Viet Nam. It describes the studies as mostly qualitative and descriptive, with few using a theoretical framework, and focused on views of providers, managers and patients.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-580",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 580",
      "pageStart": 580,
      "pageEnd": 580,
      "target": "../sections/page-580.md",
      "narrative": null,
      "content": "and low-complexity automated NAATs use and challenges to their implementation (Cattamanchi, et al., \n2020; Creswell, et al., 2014; Jaroslawski & Pai, 2012; Joshi, et al., 2018; McDowell, et al., 2018; Mwaura, \net al., 2020; Naidoo, et al., 2015; Newtonraj, et al., 2019; Phyo, et al., 2019; R. de Camargo, R. Guedes, \nCaetano, Menezes, & Trajman, 2015; Rendell, et al., 2017; Royce, et al., 2014; Shewade, et al., 2018). A \nsecond set of studies had a more procedural approach where understanding the process of using or \nimplementing diagnostics was the main aim which generated data on the perspect",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the two broad groups of included studies, those exploring perspectives on NAAT use and implementation challenges and those examining the process of using diagnostics, and judges their methodological quality as generally good. It then presents the first finding, that patients value an accurate diagnosis and closure, avoiding delays and low costs. A quote from a parent illustrates the distress caused by late detection of a child's TB.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-581",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 581",
      "pageStart": 581,
      "pageEnd": 581,
      "target": "../sections/page-581.md",
      "narrative": null,
      "content": "576 \n \n \n \n \npatients would recommend family and friends to avoid private sector services and instead immediately go \nto the public primary care facilities despite perceptions of long waiting times, lack of privacy and poor staff \nattitudes associated with public facilities (Naidoo 2015). In a study from Brazil, patients did not struggle \nwith delays or cost, because they either lived close by testing facilities and those who did have to take a \nlong bus trip were on medical leave (i.e. had time) and got the bus ticket subsidized (no extra cost) (R. de \nCamargo, et al., 2015). \n \nFinding 2: Co",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the patient perspective with examples from South Africa, where patients preferred public clinics to private ones, and Brazil, where delays and costs were not a burden. It reports that health professionals value the speed, accuracy, range of sample types and drug-resistance detection of automated NAATs compared with sputum microscopy. It adds that clinicians value the confidence that NAAT results give them in starting treatment and reassuring patients.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-582",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 582",
      "pageStart": 582,
      "pageEnd": 582,
      "target": "../sections/page-582.md",
      "narrative": null,
      "content": "577 \n \n \n \n \nto start treatment, reassure and motivate family and patients (to begin and adhere to treatment) and justify \nactions towards other doctors who do not question low-complexity automated NAAT result. Experience \nwith successful treatment following a positive NAAT result increases that confidence among clinicians and \npediatricians (McDowell, et al., 2018; Oliwa, et al., 2020). Availability of low-complexity automated \nNAAT and fast turnaround times also increased confidence in overall quality of public sector laboratories \nand willingness to collaborate and refer patients among priv",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains that confidence in NAAT results helps clinicians justify decisions and has increased private paediatricians' trust in public laboratories in India. It reports that laboratory technicians value easier, safer and more ergonomic work than with microscopy. It then turns to patient barriers to testing, such as stigma, fear of side effects, inability to produce sputum, distance, cost and inconvenient clinic hours.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-583",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 583",
      "pageStart": 583,
      "pageEnd": 583,
      "target": "../sections/page-583.md",
      "narrative": null,
      "content": "578 \n \n \n \n \nFinding 8: Health workers can be reluctant to test for TB or MDR-TB because of TB associated \nstigma and consequences for their patients, fears of acquiring TB, fear from supervisors when \nreclassifying patients already on TB treatment who turn out to be misclassified, fear of side effects of \ndrugs in children, and community awareness of disease manifestations in children. (high \nconfidence:(Oliwa, et al., 2020; Royce, et al., 2014). \nIn the context of pediatric TB in Kenya, health workers can be reluctant to test for TB because of the \nassociation of TB with being HIV positive.",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that health workers may be reluctant to test for TB or drug-resistant TB because of stigma, fear of infection, fear of supervisors and concerns about side effects in children, illustrated by a Kenyan paediatrician linking TB stigma to HIV. It describes how delays accumulate along the testing pathway through non-adherence to algorithms, late testing and empirical treatment. It also cites slow sample transport and result reporting, inconsistent recording and weak referral follow-up.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-584",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 584",
      "pageStart": 584,
      "pageEnd": 584,
      "target": "../sections/page-584.md",
      "narrative": null,
      "content": "579 \n \n \n \n \ninitiating MDR-TB treatment because of the procedural requirement to determine the MDR-TB treatment \nplan at a weekly consensus meeting (Rendell, et al., 2017). \nIn South Africa, “delays overall were longer for patients in whom initial tests were negative with 1st-line \nTB treatment started on clinical or chest x-ray findings.”(p9) (Naidoo, et al., 2015). Strategies by providers \nto deal with associated delays create new problems such as artificially prolonging turnaround times when \nasking patients to come back later, anticipating delays (Engel, et al., 2015). Passage of time and",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives examples of treatment delays, such as weekly consensus meetings for drug-resistant TB in Moldova and longer delays in South Africa for patients first treated on clinical grounds. It notes that private practitioners in India try several antibiotics before considering TB and that providers' coping strategies can lengthen turnaround. It also describes resource problems such as cartridge stock-outs that delayed diagnoses and discouraged future test orders.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-585",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 585",
      "pageStart": 585,
      "pageEnd": 585,
      "target": "../sections/page-585.md",
      "narrative": null,
      "content": "580 \n \n \n \n \nPoor laboratory infrastructure, including frequent power cuts, lack of air conditioners and/or dusty \nenvironment and lack of adequate rooms or proper furniture, can challenge proper testing (Creswell, et al., \n2014; Joshi, et al., 2018; Nalugwa, et al., 2020) and explain high test failure rates and indeterminate results \n(Joshi, et al., 2018). Yet, differences between types of failed tests are unclear and available data not always \nused. ‘No result’ test results were often caused by a power failure (Creswell, et al., 2014). \nThere was need for more basic office equipment includin",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes infrastructure problems affecting NAAT testing, including power cuts, heat, dust, lack of space and internet, which contribute to test failures and indeterminate results. It covers difficulties in hub-and-spoke sample transport, such as a lack of containers and couriers. It also lists organizational weaknesses like non-standardized systems, poor filing, unclear follow-up, weak public–private links and limited laboratory staff hours.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-586",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 586",
      "pageStart": 586,
      "pageEnd": 586,
      "target": "../sections/page-586.md",
      "narrative": null,
      "content": "581 \n \n \n \n \nLow-complexity automated NAAT use also has implications for patient flows who only have to submit one \nsputum sample (Phyo, et al., 2019) but might find it difficult to find their way through different sites and \ndepartments (Stime, et al., 2018; Vijayageetha, et al., 2019) or where they need to (Oliwa, et al., 2020; \nStime, et al., 2018). \n \nFinding 14: Too much confidence in low-complexity automated NAAT's accuracy can mean blindly \naccepting results without using clinical impressions or for patients to trust low-complexity automated \nNAAT because it is a computer-based result (",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes that NAAT use changes patient flows, requiring only one sputum sample but sometimes confusing movement between departments. It reports a finding that excessive trust in NAAT accuracy can lead clinicians to accept negative results without using clinical judgement and patients to trust machine results. It also covers training gaps, staff turnover and weak change management, with South African testing algorithms, forms and registers updated only after implementation.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-587",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 587",
      "pageStart": 587,
      "pageEnd": 587,
      "target": "../sections/page-587.md",
      "narrative": null,
      "content": "582 \n \n \n \n \nWhen introducing new diagnostics, it is also important to include relevant stakeholders in decision-making \nprocesses and in planning regarding implementation and allow a national policy consensus process. This \ncould involve national and provincial programme managers and health officers, clinicians, and laboratory \nstaff. In South Africa, Xpert had high visibility but its introduction was not inclusive, focused around FIND, \nWHO, NHLS, and the ministry of health, sidelining key national and provincial actors in the TB programme. \nThe lack of inclusion and communication was perpet",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Stresses involving relevant stakeholders and national consensus processes when introducing new diagnostics, citing South Africa where the fast-paced, externally driven Xpert roll-out sidelined national and provincial TB actors. It describes how this lack of inclusion affected implementation. It then states the high-confidence finding that delays, underuse of NAATs and lack of lower-level facilities hamper access, especially for vulnerable groups.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-588",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 588",
      "pageStart": 588,
      "pageEnd": 588,
      "target": "../sections/page-588.md",
      "narrative": null,
      "content": "583 \n \n \n \n \nSeveral studies showed how lengthy diagnostic delays, underutilization of low-complexity automated \nNAAT and lack of TB diagnostic facilities at lower levels where many presumptive patients present, hamper \naccess to prompt and accurate treatment for those that are eligible for testing (Nalugwa, et al., 2020) with \nvulnerable groups and patients with difficult disease patterns (including children (Joshi, et al., 2018; \nMcDowell, et al., 2018; Oliwa, et al., 2020), MDR-TB (Hoang, et al., 2015; Naidoo, et al., 2015)) or patients \nwith limited ability to pay (for fees or transport co",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains that diagnostic delays and underuse of NAATs hit hardest for children, people with drug-resistant TB and those unable to pay fees or transport costs, with private providers sometimes treating rather than testing. It summarizes how health system challenges compound delays and affect access and equity. It warns that these challenges risk undoing the benefits users see in NAATs and may reduce their acceptability.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p589-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 589,
      "pageEnd": 589,
      "target": "../images/img-p589-1.webp",
      "narrative": "draft",
      "content": "Diagram (Figure 1) with 'Critical Aspects Users Value' at the top, listed for patients, health workers, lab technicians and lab managers (e.g. avoid delays, keep cost low, accuracy, biosafety), above a chevron sequence of care steps: Seek Care, Order Test(s), Produce and Transport Sample, Test Runs, Results Reported and Treatment Initiated. Below, 'Challenges to Realizing Those Values' lists barriers at each step, such as stigma and cost concerns, eligibility restrictions, poor sample quality and transport, lack of resources, poor result communication and delays in recalling patients, under the banner 'Compounding Delay & Under Utilization – Undoes User Added Value; Challenges to Access & Equity'.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-589",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 589",
      "pageStart": 589,
      "pageEnd": 589,
      "target": "../sections/page-589.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Do not distribute further. \n \n \nFigure 1 \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n266",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A near-blank page carrying a distribution notice for the WHO Global TB Programme and a figure label.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-590",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 590",
      "pageStart": 590,
      "pageEnd": 590,
      "target": "../sections/page-590.md",
      "narrative": null,
      "content": "Discussion and conclusion \nThis review synthesized qualitative research on user perspectives and experiences with automated NAATs \nof low complexity for detection of tuberculosis and tuberculosis drug resistance. We organized the 18 \nindividual review findings into the following three overarching categories: \n1) critical aspects users value: Patients value reaching diagnostic closure with an accurate diagnosis, \navoiding delays, accessible facilities, keeping cost low. Healthcare professionals similarly value \naspects of accuracy the resulting confidence in TB-NAAT results, rapid turnaround ti",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarizes the review's discussion, grouping its 18 findings into what users value, the challenges to achieving those values and the implications for access and equity. It recaps that patients value accurate, timely, affordable and accessible diagnosis, health professionals value accuracy, speed and drug-resistance information, and laboratory staff value improved working conditions. It calls for better implementation processes, including early inclusive stakeholder engagement, health system strengthening, better ground-level data and management of conflicts of interest.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-591",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 591",
      "pageStart": 591,
      "pageEnd": 591,
      "target": "../sections/page-591.md",
      "narrative": null,
      "content": "of resources. Implementation processes that do not pay attention to these aspects can hamper feasibility, as \nwell as further uptake and impact of diagnostics. \n \n \nProposed evidence for ‘Evidence to Decision’ tables in GRADE \n \nPatient values and preferences: \nPatients in high-burden settings value reaching diagnostic closure (finally knowing what is wrong with \nme) comprising of a diagnosis that is accurate, a diagnosis that is timely and is not delayed (as this \nexacerbates existing financial hardships and emotional and physical suffering and make patients feel \nguilty for infecting others",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Warns that ignoring implementation processes can limit uptake and impact of diagnostics, then proposes evidence for GRADE evidence-to-decision tables. It summarizes patient values and preferences and equity concerns, followed by the values of clinicians and laboratory staff, each with a confidence rating.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-592",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 592",
      "pageStart": 592,
      "pageEnd": 592,
      "target": "../sections/page-592.md",
      "narrative": null,
      "content": "satisfaction and has a symbolic meaning of progress within the TB world. (Low confidence, applicable to \nindex test 1) \nThe identified feasibility challenges with low-complexity automated NAAT utilization and accumulated \ndelays at every step risk undoing the added value/benefits as identified by the users (avoiding delays, \nkeeping cost lost, accurate results, drug resistant information, easing laboratory work), ultimately leading \nto underutilization (High confidence, applicable to index test 1). We can assume that if these values are \nnot met users are less likely to find low-complexity aut",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the proposed evidence-to-decision content, stating that feasibility problems and accumulated delays risk undoing NAAT benefits and leading to underuse, mainly due to health system factors. It reports feedback from guideline group members, who valued the varied perspectives and recommended better diagnostics training for nurses and clinicians.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-593",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 593",
      "pageStart": 593,
      "pageEnd": 593,
      "target": "../sections/page-593.md",
      "narrative": null,
      "content": "• \nThey suggested adding qualitative methods routinely and from the beginning to diagnostic \nevaluation and clinical studies. \n• \nOne member suggested exploring collaborations with costing studies. \n \nAcknowledgements \n \nWe are grateful to Vittoria Lutje, Cochrane Infectious Diseases' Information Specialist, for helping us \nwith the search strategy. We are grateful to Darrel Ronald for helping us with the design of the figure.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Ends the review with suggestions to include qualitative methods routinely in diagnostic studies and to collaborate with costing studies, followed by acknowledgements.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-594",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 594",
      "pageStart": 594,
      "pageEnd": 594,
      "target": "../sections/page-594.md",
      "narrative": null,
      "content": "References \nAlbert, H., Nathavitharana, R. R., Isaacs, C., Pai, M., Denkinger, C. M., & Boehme, C. C. \n(2016). Development, roll-out and impact of Xpert MTB/RIF for tuberculosis: what \nlessons have we learnt and how can we do better? [10.1183/13993003.00543-2016]. \nEuropean Respiratory Journal, 48, 516-525. \nBainomugisa, A., Gilpin, C., Coulter, C., & Marais, B. J. (2020). New Xpert MTB/XDR: added \nvalue and future in the field. European Respiratory Journal, 56(5), 2003616. \nBeisel, U., Umlauf, R., Hutchinson, E., & Chandler, C. I. (2016). The complexities of simple \ntechnologies: re-imagining",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The first page of the reference list for the qualitative review on automated NAATs for TB, citing studies on Xpert MTB/RIF roll-out, diagnostic technologies and TB diagnosis in various countries.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-595",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 595",
      "pageStart": 595,
      "pageEnd": 595,
      "target": "../sections/page-595.md",
      "narrative": null,
      "content": "Hoang, T., Nguyen, N., Dinh, S., Nguyen, H., Cobelens, F., Thwaites, G., et al. (2015). \nChallenges in detection and treatment of multidrug resistant tuberculosis patients in \nVietnam. BMC Public Health, 15(1), 980. \nJaroslawski, S., & Pai, M. (2012). Why are inaccurate tuberculosis serological tests widely used \nin the Indian private healthcare sector? A root-cause analysis. Journal of Epidemiology \nand Global Health, 2, 39-50. \nJoshi, B., Lestari, T., Graham, S. M., Baral, S. C., Verma, S. C., Ghimire, G., et al. (2018). The \nimplementation of Xpert MTB/RIF assay for diagnosis of tuberculosi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list, citing qualitative and mixed-methods studies on TB diagnosis and drug-resistant TB in countries such as Viet Nam, India, Nepal and Kenya.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-596",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 596",
      "pageStart": 596,
      "pageEnd": 596,
      "target": "../sections/page-596.md",
      "narrative": null,
      "content": "Pai, M., Schumacher, S. G., & Abimbola, S. (2018). Surrogate endpoints in global health \nresearch: still searching for killer apps and silver bullets? BMJ Global Health, 3(2). \nPhyo, A. M., Kumar, A. M. V., Soe, K. T., Kyaw, K. W. Y., Thu, A. S., Wai, P. P., et al. (2019). \nContact Investigation of Multidrug-Resistant Tuberculosis Patients: A Mixed-Methods \nStudy from Myanmar. Tropical medicine and infectious disease, 5(1), 3. \nPooran, A., Theron, G., Zijenah, L., Chanda, D., Clowes, P., Mwenge, L., et al. (2019). Point of \ncare Xpert MTB/RIF versus smear microscopy for tuberculosis diagnosis",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Concludes the reference list with studies from Myanmar, southern Africa and Brazil and WHO's 2020 global TB report and consolidated TB guidelines.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-597",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 597",
      "pageStart": 597,
      "pageEnd": 597,
      "target": "../sections/page-597.md",
      "narrative": null,
      "content": "Appendices \n \nAppendix 1: Search strategy \n \nDatabase: Ovid MEDLINE(R) and Epub Ahead of Print, In-Process & Other Non-Indexed Citations, \nDaily and Versions(R) <1946 to September 05, 2020> \nSearch Strategy: \n-------------------------------------------------------------------------------- \n1 \nExtensively Drug-Resistant Tuberculosis/ or Tuberculosis/ or tuberculosis.mp. or Tuberculosis, \nMultidrug-Resistant/ or Tuberculosis, Pulmonary/ or Mycobacterium tuberculosis/ \n2 \n(Tuberculosis or MDR-TB or XDR-TB or tuberculous).ti. or (Tuberculosis or MDR-TB or XDR-TB \nor tuberculous).ab. \n3 \n1 or 2 \n4",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Appendix 1, giving the Ovid MEDLINE search strategy used for the review, combining terms for TB and drug-resistant TB, automated NAAT products such as Xpert and Truenat, and qualitative or mixed methods, limited to 2007 onwards.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-598",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 598",
      "pageStart": 598,
      "pageEnd": 598,
      "target": "../sections/page-598.md",
      "narrative": null,
      "content": "Database: Embase <1996 to 2020 Week 36> \nSearch Strategy: \n-------------------------------------------------------------------------------- \n1 \ntuberculosis/ or tuberculosis.mp. \n2 \ndrug resistant tuberculosis.mp. or drug resistant tuberculosis/ \n3 \nmultidrug resistant tuberculosis.mp. or multidrug resistant tuberculosis/ \n4 \nMDR-TB.mp. \n5 \nXDR-TB.mp. \n6 \nextensively drug resistant tuberculosis/ \n7 \nmycobacterium tuberculosis.mp. or Mycobacterium tuberculosis/ \n8 \n1 or 2 or 3 or 4 or 5 or 6 or 7 \n9 \n(Truenat or Cepheid or Xpert*).mp. \n10 \nGenexpert*.mp. \n11 \ndrug susceptibility test*.mp. \n12",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the search strategy appendix with the Embase search terms for TB, automated NAAT and diagnostic tests, user experience and qualitative methods, limited to 2007 onwards.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-599",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 599",
      "pageStart": 599,
      "pageEnd": 599,
      "target": "../sections/page-599.md",
      "narrative": null,
      "content": "49 \nlimit 48 to yr=\"2007 -Current\" \n \nCinahl (EBSCOHost) Limiters - Published Date: 20070101-20201231 \n# \nQuery \nS22 S8 AND S20 \nS21 S8 AND S20 \nS20 TX systematic review or meta-analysis \nS19 S14 OR S18 \nS18 S8 AND S17 \nS17 S15 OR S16 \nS16 TX focus group* \nS15 TX interview* OR TX ( survey* or questionnaire* ) OR TX ( qualitative research or qualitative \nstudy or qualitative methods or mixed methods ) \nS14 S8 AND S12 \nS13 S8 AND S12 \nS12 S9 OR S10 OR S11 \nS11 TX ( barriers or challenges ) OR TX critical pathway OR TX facilitator* \nS10 TX patient preference* OR TX ( patient satisfaction or patie",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists the final Embase limit and the CINAHL and PsycInfo search strategies for the review, combining TB, NAAT and qualitative or patient experience terms for 2007–2020.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-600",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 600",
      "pageStart": 600,
      "pageEnd": 600,
      "target": "../sections/page-600.md",
      "narrative": null,
      "content": "S15 TX interview* OR TX ( survey* or questionnaire* ) OR TX ( qualitative research or qualitative \nstudy or qualitative methods or mixed methods ) \nS14 S8 AND S12 \nS13 S8 AND S12 \nS12 S9 OR S10 OR S11 \nS11 TX ( barriers or challenges ) OR TX critical pathway OR TX facilitator* \nS10 TX patient preference* OR TX ( patient satisfaction or patients experiences or patients perceptions \nor patients attitudes ) \nS9 TX acceptance of care OR TX health equity OR Health Services Accessibility \nS8 S3 AND S7 \nS7 S4 OR S5 OR S6 \nS6 TX Max MDR-TB OR TI cartridge OR AB cartridge \nS5 TX drug susceptibility tes",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the PsycInfo search strategy for the review, with terms for qualitative methods, patient experience, barriers and automated NAAT products.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-601",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 601",
      "pageStart": 601,
      "pageEnd": 601,
      "target": "../sections/page-601.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Do not distribute further. \n \n \n596\n \n \nAppendix 2: Assessment of Methodological Limitations \n \nStudy ID \nWere steps taken to increase \nrigour in the sampling? \nWere steps taken to increase \nrigour in the data collected? \nWere steps taken to increase \nrigour in the analysis of the \ndata? \nWere the findings of the study \ngrounded in/ supported by the \ndata? \nPlease rate the findings of the \nstudy in terms of their breadth \nand depth. \nMcDowell 2016 \nYes, a fairly thorough attempt was \nmade \nYes, a fairly thorough attempt was \nmade",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Appendix 2, a table assessing the methodological limitations of each included study, rating rigour in sampling, data collection and analysis, how well findings are grounded in the data, and breadth and depth. Several studies are rated as making fairly thorough efforts, while others, such as Cattamanchi 2020, are rated as not reporting or unclear on most criteria. The page carries a notice that the report was prepared for the WHO Global TB Programme and is not for further distribution.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-602",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 602",
      "pageStart": 602,
      "pageEnd": 602,
      "target": "../sections/page-602.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Do not distribute further. \n \n \n597 \n \n \nAppendix 3: CERQual summary of findings \n \nFinding # \nReview  finding \nMethodological  limitations \nCoherence \nRelevance \nAdequacy \nCERQual assessment of \nconfidence in the \nevidence \nExplanation  of  CERQual assessment \nStudies contributing to \nreview finding \nCritical  aspects  users value \n1 \nPatients  in high‐TB  burden  countries value \n1) getting an accurate diagnosis and \nreaching diagnostic closure (finally knowing \nwhat is wrong with me), 2) avoiding \ndiagnostic  delays  as  they ex",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Appendix 3, the CERQual summary-of-findings table, starting with what users value, such as patients' wish for accurate, timely, accessible and affordable diagnosis. For each finding it rates methodological limitations, coherence, relevance and adequacy and gives an overall confidence level with an explanation. It lists the studies contributing to each finding.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-603",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 603",
      "pageStart": 603,
      "pageEnd": 603,
      "target": "../sections/page-603.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Do not distribute further. \n \n \n598 \n \n \n6 \nLaboratory managers appreciate that \nmonitoring of  laboratory work and   training \nis easier than with sputum microscopy and \nthat low-complexity automated NAAT eases \nstaff retention, as it increases staff  \nsatisfaction and has a symbolic meaning of \nprogress within  the TB  world.. \nno/very  minor concerns \nno concerns ‐  good fit of  finding \nwith  primary study \nserious concerns ‐ just one \nsetting (urban, public clinic),  \nstudy early in implementation of \nCB‐NAAT \nmoderate concern",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the CERQual table with the finding that laboratory managers find NAAT eases monitoring, training and staff retention, rated low confidence because it rests on a single study.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-604",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 604",
      "pageStart": 604,
      "pageEnd": 604,
      "target": "../sections/page-604.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Do not distribute further. \n \n \n599 \n \n \nFinding # \nReview  finding \nMethodological  limitations \nCoherence \nRelevance \nAdequacy \nCERQual assessment of \nconfidence in the \nevidence \nExplanation  of  CERQual assessment \nStudies contributing to \nreview finding \nChallenges  to realizing  these values \n7 \nPatients can be reluctant to test for \nTB/MDR‐TB because of stigma related to \nMDR‐TB or related to having interrupted \ntreatment in the past, because of fears of \nside effects, the failure to recognize \nsymptoms, the inability to pro",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the CERQual table with findings on challenges, starting with patients' reluctance to test because of stigma, fear of side effects, sputum problems and travel costs, which is rated high confidence. It also assesses findings on health worker reluctance to test, delays accumulating through health system factors, and poor sample quality and transport causing errors and underuse. Each entry gives ratings for limitations, coherence, relevance and adequacy and lists the contributing studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-605",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 605",
      "pageStart": 605,
      "pageEnd": 605,
      "target": "../sections/page-605.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Do not distribute further. \n \n \n600 \n \n \n \n \n \nFinding # Review  finding \nMethodological  limitations \nCoherence \nRelevance \nAdequacy \nCERQual assessment of \nconfidence in the \nevidence \nEplanation  of  CERQual assessment \nStudies contributing to \nreview finding \n12 \nLow-complexity automated NAAT seems to \ndecrease workload by freeing up time for \nlaboratory staff, but in most settings staff \nmay be hesitant to accept testing with low- \ncomplexity automated NAAT because it \nincreases workload if added onto existing \nlaboratory work",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the CERQual table with findings on laboratory workload, where NAAT can save staff time but may add work if layered on existing tests without staffing changes. It also rates findings on how workflows and patient flows affect use, over-reliance on NAAT results, and implementation weaknesses such as missing guidelines, training and inclusive national decision-making. Most are rated high or moderate confidence, with the contributing studies listed.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-606",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 606",
      "pageStart": 606,
      "pageEnd": 606,
      "target": "../sections/page-606.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Do not distribute further. \n \n \n601 \n \n \nFinding # Review  finding \nMethodological  limitations \nCoherence \nRelevance \nAdequacy \nCERQual assessment of \nconfidence in the \nevidence \nExplanation  of  CERQual assessment \nStudies contributing to \nreview finding \nConcerns  for access/equity \n16 \nStaff and  managers voiced concerns \nregarding sustainability of funding and \nmaintenance, complex conflicts of interest \nbetween donors and implementers and \nconcerns related to the strategic and  \nequitable use of resources, which negatively",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Part of a summary-of-findings table from a qualitative review of TB NAATs, marked as prepared for the WHO Global TB Programme and not for further distribution. It gives GRADE-CERQual assessments for findings on access and equity, such as staff concerns about sustainable funding, maintenance, donor conflicts of interest and equitable use of resources for low-complexity automated NAATs, with contributing studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-607",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 607",
      "pageStart": 607,
      "pageEnd": 607,
      "target": "../sections/page-607.md",
      "narrative": null,
      "content": "Appendix 4: Figure 1: PRISMA flow chart \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n284 \nRecords identified through \nda abas s a hing \n(n =  \n ) \nAdditional records identified \nhrough ot\n sourc s \n(n =   ) \nds af\n dupli a s \nd \n(n = 1308 ) \nds s\nd \n(n = \n) \nds x lud d \n(n =  \n ) \nFull- x a i l s x lud\ni h \nasons \n(n = \n  ) \n  Da a not anal\nd using \nquali a i\n anal\nis \nhods \n  Did not add\nss \nph nom non of int\ns\n  ot a quali a i\n s ud\nsign \n \nvi\n a i l \nFull- x a i l s ass ss d \nf\nligibili \n(n =  \n ) \nS udi s includ d in \nquali a i\n syn\nsis",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A PRISMA flow chart for the qualitative review whose extracted text is badly garbled, with letters and most study counts missing; only the screening stages and 1308 records after de-duplication are legible.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-608",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 608",
      "pageStart": 608,
      "pageEnd": 608,
      "target": "../sections/page-608.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Do not distribute further. \n \n \n603 \n \n \nAppendix 5: Study characteristics table \n \nFirst author \nYear of \npublication \nCountry (income \nclassification) \nGeographical \nsetting \nType of health facility \npublic/private \nBackground \nprevalence \nDiagnostic  technology \nProgrammatic features of the \nintervention (Where and how) \nTarget  population \nTotal number of \nparticipants and types of \nsettings \nResearch  questions/  objectives \nData collection methods \nMwaura \n2020 Kenya (lower \nmiddle income) \nand Eswatini \n(lower middle \nincome",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Begins a table of characteristics of studies included in the qualitative review of TB NAATs, giving country and income level, setting, diagnostic technology, target population, participants, objectives and data collection methods. Entries include a Kenya and Eswatini study on Xpert Ultra and studies such as one on TB screening of pregnant women in India.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-609",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 609",
      "pageStart": 609,
      "pageEnd": 609,
      "target": "../sections/page-609.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Do not distribute further. \n \n \n604 \n \n \nFirst author \nYear of \npublication \nCountry (income \nclassification) \nGeographical \nsetting \nType of health facility \npublic/private \nBackground \nprevalence \nDiagnostic  technology \nProgrammatic features of the \nintervention (Where and how) \nTarget  population \nTotal number of \nparticipants and tpes of \nResearch  questions/  objectives \nData collection methods \nPhyo \n2019 Myanmar (lower \nmiddle income) \nUrban \nCommunity outreach in \ntownships \nOther: NGO led \nHigh burden \nXpert MTB/RIF \nHous",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the study characteristics table with Xpert MTB/RIF studies, including NGO-led household contact investigation for MDR-TB in Myanmar. Other entries describe programme features, participants and methods. The page ends with a mixed-methods time-and-motion study of clinic flow with point-of-care testing in Durban, South Africa.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-610",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 610",
      "pageStart": 610,
      "pageEnd": 610,
      "target": "../sections/page-610.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Do not distribute further. \n \n \n605 \n \n \nFirst author \nYear of \npublication \nCountry (income \nclassification) \nGeographical \nsetting \nType of health facility \npublic/private \nBackground \nprevalence \nDiagnostic  technology \nProgrammatic features of the \nintervention (Where and how) \nTarget  population \nTotal number of \nparticipants and types of \nsettings \nResearch  questions/  objectives \nData collection methods \nRendell \n2017 Mongolia (lower \nmiddle income) \nUrban \nNational TB reference lab, \nprovincial TB clinics, district \nTB cli",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the study characteristics table with further Xpert MTB/RIF studies, including implementation across laboratory and clinic levels in Mongolia. It ends with a mixed-methods study of challenges in the steps from diagnosis to MDR-TB treatment, using focus groups, interviews and document review with 80 health staff.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-611",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 611",
      "pageStart": 611,
      "pageEnd": 611,
      "target": "../sections/page-611.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Do not distribute further. \n \n \n606 \n \n \nFirst author \nYear of \npublication \nCountry (income \nclassification) \nGeographical \nsetting \nType of health facility \npublic/private \nBackground \nprevalence \nDiagnostic  technology \nProgrammatic features of the \nintervention (Where and how) \nTarget  population \nTotal number of \nparticipants and types of \nsettings \nResearch  questions/  objectives \nData collection methods \nColvin \n2015 South Africa \n(upper middle \nincome) \nUrban \nHospital and clinic \nPublic facility \nHigh burden \nXpert MTB/RI",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the study characteristics table, including a South African study of the policy transfer of Xpert and line probe assays with informants from global to district level. It ends with an interview study on why serological TB tests are popular in the private sector.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-612",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 612",
      "pageStart": 612,
      "pageEnd": 612,
      "target": "../sections/page-612.md",
      "narrative": null,
      "content": "Web Annex B.14. User perspectives on nucleic acid amplification tests \nfor tuberculosis and tuberculosis drug resistance: (interviews study) \n \nAuthors: Ricky Janssen, Muthoni Mwaura & Nora Engel \nDepartment of Health, Ethics and Society, Care and Public Health Research Institute (CAPHRI), Maastricht \nUniversity, Maastricht, Netherlands (Kingdom of the) \n \nProduced in preparation for the WHO guideline group meeting on “Nucleic acid amplification tests to detect \ntuberculosis and drug-resistant tuberculosis” on 7-18 December, 2020. \nPlease send comments and feedback to Nora Engel, n.engel@maast",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Opens Web Annex B.14, a 2020 interview study by Maastricht University researchers on user perspectives of NAATs for TB and drug-resistant TB, prepared for a WHO guideline meeting. It argues that user views on value, feasibility and acceptability matter alongside accuracy and that qualitative research can capture them. It describes the study's interviews with clinicians, laboratory staff, programme officers and patient advocates in Moldova, India and South Africa and the participant coding system.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-613",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 613",
      "pageStart": 613,
      "pageEnd": 613,
      "target": "../sections/page-613.md",
      "narrative": null,
      "content": "Interviews were conducted over Zoom, Skype and phone. Topics discussed included: 1. current approach to \ndiagnosing TB, MDR-TB, and XDR-TB including specific challenges; 2. experiences with using molecular \nTB diagnostics and the index tests specifically, including details on steps taken in the diagnostic process; 3. \ndetermining eligibility and treatment initiation as well as challenges and benefits of using the index tests; 4. \noverall usefulness of the index tests; 5. the feasibility of implementing the index tests; 6. their potential impact \non health equity and; 7. how this relates to cur",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists the interview topics, from current diagnostic approaches to feasibility, equity and policy, and notes limitations such as small numbers and remote interviews that prevented triangulation. It outlines the report's structure and introduces the first index test, low-complexity automated NAATs.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-614",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 614",
      "pageStart": 614,
      "pageEnd": 614,
      "target": "../sections/page-614.md",
      "narrative": null,
      "content": "The first class of technologies to be evaluated at the WHO 2020 guideline meeting are low-complexity \nautomated NAATs for the detection of resistance to isoniazid (INH) and second-line anti-TB medicines. One \nsuch test that we investigated in this study is the Xpert MTB/XDR cartridge by Cepheid. This particular \ncartridge detects mutations associated with resistance towards INH, fluoroquinolones (FLQ), second-line \ninjectable drug (SLID) (amikacin, kanamycin, capreomycin) and ethionamide (ETH) (Cepheid, 2020). \n \nBenefits \n \nAlthough only three participants in this study had used the cartridge",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the Xpert MTB/XDR cartridge, which detects resistance to isoniazid, fluoroquinolones, second-line injectables and ethionamide, and notes that few participants had used it. Participants anticipated benefits such as faster drug susceptibility results, earlier detection of isoniazid monoresistance and earlier information to guide treatment regimens.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-615",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 615",
      "pageStart": 615,
      "pageEnd": 615,
      "target": "../sections/page-615.md",
      "narrative": null,
      "content": "was also important in giving clinicians necessary information to develop appropriate treatment regimens as \nsoon as possible (SM-1, IM-2). A laboratory scientist (SL-2) summarized the value of Xpert MTB/XDR as: \n \n“you would get almost universal resistance prediction for a resistant patient, within 6 \nhours. Because you would have RIF, you would have INH, you would have \nfluoroquinolones, you would have aminoglycosides, you would have partially predicted \nethionamide, clinicians have got a lot that they can work with, from a primary sample \nfrom a patient”. \n \nChallenges \n \nWhen it came to exp",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues on Xpert MTB/XDR, quoting a laboratory scientist on near-universal resistance prediction within hours. It reports challenges, chiefly cost and the need for a 10-colour GeneXpert module that may require replacing existing platforms, alongside the broad acceptability of the test.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-616",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 616",
      "pageStart": 616,
      "pageEnd": 616,
      "target": "../sections/page-616.md",
      "narrative": null,
      "content": "staff and clinicians based on its simple user steps, and due to the amount of important information it provides. \nThe new cartridge requires less user training compared to other DST methods (such as LPA and culture), \nmaking it more feasible to implement. However, the cost of the 10-colour system for this cartridge, and the \nvalue of diagnosing MTB over DR TB at primary care, makes it less feasible as a baseline test, and therefore \nmay be more suitable if placed at a district or intermediate level lab. \n \nIndex Test 2: Medium-complexity automated NAATs \nThe second class of technologies to be",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Concludes that Xpert MTB/XDR is acceptable and needs little training but may suit district or intermediate laboratories rather than primary care because of cost. It then introduces medium-complexity automated NAATs for TB, rifampicin and isoniazid resistance, describing their high-throughput automation and benefits such as freeing low-complexity platforms for other areas.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-617",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 617",
      "pageStart": 617,
      "pageEnd": 617,
      "target": "../sections/page-617.md",
      "narrative": null,
      "content": "Potential challenges around medium-complexity automated NAATs include the cost of acquiring the machines \nand provision of good quality samples (ML-1, IL-1, IL-2). Furthermore, the chosen platform has to align with \nthe physical laboratory space. For example, although the Fluorotype platform was valued for its automation, a \nparticipant mentioned concerns regarding the size of the platform. \n“So you just can’t place it anywhere. You can’t just place it on any bench top. But fortunately I work \nin a very nice laboratory, it was funded by USAID, PEPFAR, so I have got a world-class smart, you \nkn",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes challenges of medium-complexity NAATs, including machine costs, sample quality and the physical size of platforms such as Fluorotype, illustrated by a laboratory scientist's quotation. It introduces the third index test, a line probe assay for pyrazinamide resistance (PZA LPA), using the Genoscholar PZA-TB II assay as an example.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-618",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 618",
      "pageStart": 618,
      "pageEnd": 618,
      "target": "../sections/page-618.md",
      "narrative": null,
      "content": "Benefits of PZA LPA \nOnly one participant (SL-2) had experience using the NIPRO test, and this was within a research context in \nSouth Africa. However, laboratory participants did have experience doing line probe assays for first and second \nline DST, and were therefore able to comment on potential benefits and challenges of PZA LPA. Potential \nbenefits of PZA LPA identified include lower cost compared to whole genome sequencing; easier \ninterpretation than whole genome sequencing; faster than culture DST; and the possibility of providing \ninformation on PZA resistance at more than one concent",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that few participants had used the PZA line probe assay but that laboratory staff saw potential benefits, such as lower cost and easier interpretation than genome sequencing and faster results than culture. It then discusses challenges, beginning with the need for skilled sample collection, transport and handling to avoid contamination.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-619",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 619",
      "pageStart": 619,
      "pageEnd": 619,
      "target": "../sections/page-619.md",
      "narrative": null,
      "content": "results (IP-2, IL-2) (also see Finding 10 in report by Engel et al., 2020). Proper sample handling, as well as \nproper temperature maintenance, and use of correct volumes of liquid are crucial in order to obtain accurate \nresults (IL-2). LPAs also require that part of the sample is dedicated to controls, and that there is sufficient \nsample volume to make it worth running controls from a cost perspective (IL-1). The lower the sample load, \nthe more controls need to be run, resulting in a higher overall cost. \nFurthermore, only samples with a high enough DNA load can be run using the LPA, furth",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues on PZA LPA challenges, explaining that accurate results depend on careful handling, controls and a sufficient DNA load, so it is often run only on smear-positive or cultured samples. It describes difficulties in interpreting the test strips and an interpretation app of unclear status.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-620",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 620",
      "pageStart": 620,
      "pageEnd": 620,
      "target": "../sections/page-620.md",
      "narrative": null,
      "content": "The relevance of PZA resistance testing in general seemed to vary between countries. In South Africa, \nparticipants often mentioned that PZA resistance testing is only done for surveillance and research, but is not \nroutinely utilized for clinical settings (SL-2, SM-1, SP-1). \n“nationally we do it as a reference lab for research projects particularly and also for surveillance \nactivities but generally in the routine diagnostic practice we do not test for pyrazinamide, because \nirrespective of its resistance, it’s still used.” (SL-2) \nA South African clinician said that she would need to specif",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that PZA resistance testing is used mainly for surveillance and research rather than routine care in South Africa, for reasons including cost and doubts about results, while relevance varied between countries. It summarizes concerns about the acceptability of PZA LPA, such as performance on smear-negative samples.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-621",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 621",
      "pageStart": 621,
      "pageEnd": 621,
      "target": "../sections/page-621.md",
      "narrative": null,
      "content": "actually detects mutations specific to PZA resistance; and clarification in some settings as to why this specific \nDST drug test is being prioritized, as it is not currently part of routine DST. In regards to feasibility, participant \nconcerns centered around the significant training and laboratory infrastructure required to implement PZA \nLPA, including proper sample handling (please also see section IV.). Feasibility for this test also hinges on the \navailability of an automated interpretation system, as it is difficult to interpret. \n \nPart 2: Findings applicable to all 3 index tests \nSampl",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues concerns about PZA LPA, including whether it detects the relevant mutations, why it is prioritized, and the training, infrastructure and automated interpretation it would need. It then turns to findings applicable to all three tests, covering sample types, collection and transport, especially in rural settings and in Moldova's donor-funded transport system.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-622",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 622",
      "pageStart": 622,
      "pageEnd": 622,
      "target": "../sections/page-622.md",
      "narrative": null,
      "content": "on donor funding for various TB program activities, including sample transport. This raises concern about the \nsustainability of TB program funding and resources, a finding echoed by other qualitative research (Colvin \n2015, deCamargo 2015, Cresswell 2014). As more diagnostics are adopted by programmes, funding towards \nsystems that support these diagnostics (e.g. sample transportation) needs to be sustainable in order to get the \nmost out of them. \nHandling: Sample handling plays an important role in the perceived reliability of LPAs and whole genome \nsequencing (IL-1, IL-2, IL-4, ML-2). If m",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Raises concerns about donor-dependent sample transport and argues for sustainable funding of supporting systems. It discusses how sample handling affects the reliability of LPAs and sequencing and how clinicians interpret complex molecular results, calling for guidelines and algorithms that support communication between clinicians and laboratories.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-623",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 623",
      "pageStart": 623,
      "pageEnd": 623,
      "target": "../sections/page-623.md",
      "narrative": null,
      "content": "such that they can discuss discordant results, and interpret laboratory results in the context of drug availability \nand patient history. Clinicians and laboratory staff also use patient progress to contextualize diagnostic results \nwhen there is uncertainty and use this to make treatment decisions (ML-2), but sometimes a patient is switched \nto another regimen based on test results despite clinical progress (SM-1). We suggest that policy-makers \nfurther incorporate clinical knowledge around patient progress to understand the reliability and clinical \nrelevance of results from new diagnostics",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Discusses how clinicians and laboratories reconcile discordant results with patient progress and urges policy-makers to consider clinical knowledge. It describes social barriers to diagnosis and stresses that access to new diagnostics must accompany access to new treatments.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-624",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 624",
      "pageStart": 624,
      "pageEnd": 624,
      "target": "../sections/page-624.md",
      "narrative": null,
      "content": "Finally, the speed at which WHO guidelines are changing does not match the speed at which many country \nprogrammes are able to implement the guidelines. This translates into differential access to new TB diagnostics \nand treatment at an inter-country level (i.e. between countries that can and cannot quickly keep up with the \nrapidly changing TB diagnostic environment MR-1) as well at an intra-country level (i.e. between patients \nwho can and cannot afford the private health system that is better equipped to quickly adopt new diagnostics \nand policies; IL-1, IR-1). \n \nConclusions \nThe potential",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes that WHO guidelines change faster than many countries can implement them, creating unequal access between and within countries. It concludes that new molecular diagnostics have great potential but their value depends on fit within health systems, and begins a summary of challenges such as sample transport, contamination and clinician–laboratory communication.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-625",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 625",
      "pageStart": 625,
      "pageEnd": 625,
      "target": "../sections/page-625.md",
      "narrative": null,
      "content": "• \nHesitancy of clinicians to make decisions on MDR and XDR TB on their own \n• \nCapacity of decentralized facilities to interpret and disseminate increased volumes of \ncomplex information produced by the tests \n \n6. Equity \n• \nBarriers at program level: cost to acquire new expensive diagnostics as well as difficulties \nkeeping up with WHO guidelines and subsequent training for clinicians and laboratory \npersonnel \n• \nBarriers at patient level: lack of access to information and lack of access to treatments for \nwhich new diagnostics are testing \n• \nLack of diagnostics available for newer treatm",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the summary of challenges, including clinicians' hesitancy over MDR and XDR-TB decisions, information overload at decentralized sites and equity barriers at programme and patient levels. It notes the COVID-19 context and begins recommendations, such as gathering data on the burden of pyrazinamide resistance.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-626",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 626",
      "pageStart": 626,
      "pageEnd": 626,
      "target": "../sections/page-626.md",
      "narrative": null,
      "content": "PZA LPA (and PZA DST in general).  This  information  should  be  monitored  and  taken  into \nconsideration for future guideline development. \n \nAll three index tests: \n4. Sample: \n• \nAlgorithms may need to consider how many samples are required from the patient, where \nthey will be collected, and at which point in the patient journey the sample will be collected. \nThis process should be streamlined where possible. Not doing so, there is risk of losing \nthe patient before a diagnosis is made, or not being able to get additional samples from the \npatient as their conditions improves. \n• \nGuide",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives recommendations for all three index tests on streamlining sample collection and ensuring sufficient sample volumes in diagnostic algorithms. It also suggests content for GRADE evidence-to-decision tables and stresses patients' access to clear information on diagnostics as an equity issue.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-627",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 627",
      "pageStart": 627,
      "pageEnd": 627,
      "target": "../sections/page-627.md",
      "narrative": null,
      "content": "New treatment options need to be matched with new diagnostics: it is important to improve access to \ntreatment based on new diagnostics, it is equally important to improve access to diagnostics for new \ntreatment options. \nThe speed at which WHO guidelines are changing does not match the speed at which many country \nprogrammes are able to implement the guidelines. This translates into differential access to new TB diagnostics \nand treatment at an inter-country level (i.e. between countries that can and cannot quickly keep up with the \nrapidly changing TB diagnostic environment) as well at an i",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarizes the study's points for GRADE evidence-to-decision tables, including matching diagnostics to new treatments and unequal access caused by the speed of guideline change. It covers acceptability factors such as infrastructure, sample quality and clinician–laboratory communication.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-628",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 628",
      "pageStart": 628,
      "pageEnd": 628,
      "target": "../sections/page-628.md",
      "narrative": null,
      "content": "and interpret laboratory results in the context of drug availability, patient history, and patient progress on a \ncurrent drug regimen. \n \nFeasibility for the Xpert MTB/XDR Cartridge (Low-complexity automated NAAT) is challenged by \nthe cost of the 10-colour system for this cartridge, and the value of diagnosing MTB over DR TB at primary \ncare, makes it less feasible as a baseline test, though it would fit at a district or intermediate level laboratory. \nThe Xpert MTB/XDR Cartridge requires less user training compared to other DST methods (such as \nLPA and culture), making it more feasible to",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarizes the feasibility of each index test: Xpert MTB/XDR limited by the cost of the 10-colour system but needing little training, and medium-complexity platforms limited by laboratory space. It ends with references, including qualitative studies and Stop TB Partnership data.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-629",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 629",
      "pageStart": 629,
      "pageEnd": 629,
      "target": "../sections/page-629.md",
      "narrative": null,
      "content": "World Health Organization. (2020). Global Tuberculosis Report 2020. Geneva: World Health Organization.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The final reference on the page, citing the WHO Global Tuberculosis Report 2020.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-630",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 630",
      "pageStart": 630,
      "pageEnd": 630,
      "target": "../sections/page-630.md",
      "narrative": null,
      "content": "Web Annex 4.15. Low complexity automated NAATs: Diagnostic \naccuracy for detection of resistance to isoniazid and second-line anti- \nTB agents. A systematic review \nSamantha Pillay1, Geraint R Davies2, Marty Chaplin3, Margaretha De Vos4, Rob Warren1, \nKaren R Steingart5, Grant Theron1 \n \n1DSI-NRF Centre of Excellence for Biomedical Tuberculosis Research, SAMRC Centre for Tuberculosis \nResearch, Division of Molecular Biology and Human Genetics, Faculty of Medicine and Health Sciences, \nStellenbosch University, Cape Town, South Africa \n2Liverpool, UK \n3Department of Clinical Sciences, Liverpool",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title page of Web Annex 4.15, a systematic review of the diagnostic accuracy of low-complexity automated NAATs for detecting resistance to isoniazid and second-line TB drugs, listing authors from Stellenbosch University, the Liverpool School of Tropical Medicine and FIND.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-631",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 631",
      "pageStart": 631,
      "pageEnd": 631,
      "target": "../sections/page-631.md",
      "narrative": null,
      "content": "EXECUTIVE SUMMARY \n \nXpert® MTB/XDR Assay (Xpert MTB/XDR, Cepheid, Sunnyvale, USA) detects MTBC \n(Mycobacterium tuberculosis complex) DNA and genomic mutations associated with resistance to \nisoniazid, fluoroquinolones, ethionamide, and second-line injectable drugs (amikacin, kanamycin, \ncapreomycin) in a single cartridge. Xpert MTB/XDR is intended for use as a reflex test in clinical \nspecimens (unprocessed sputum or concentrated sputum sediments) already determined to be MTBC- \npositive. The test is included in a class of diagnostic technologies that are cartridge-based and of low \ncomplexit",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The executive summary of a systematic review of the Xpert MTB/XDR cartridge test, which detects tuberculosis and mutations linked to resistance to isoniazid, fluoroquinolones, ethionamide and second-line injectable drugs. It describes the test's proposed role as an initial test for resistance after TB is confirmed, and reports a pooled sensitivity of 94.2% and specificity of 98.0% for isoniazid resistance.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-632",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 632",
      "pageStart": 632,
      "pageEnd": 632,
      "target": "../sections/page-632.md",
      "narrative": null,
      "content": "Results of these studies indicate that in theory, of 1000 people where 50 have isoniazid resistance, 66 \nwould be Xpert MTB/XDR-positive: of these, 19 (29%) would not have isoniazid resistance (false- \npositives) and 934 would be Xpert MTB/XDR-negative: of these, 3 (0%) would have isoniazid \nresistance (false-negatives). \n \nXpert MTB/XDR for fluoroquinolone resistance in people with microbiologically confirmed \npulmonary tuberculosis irrespective of rifampicin resistance, pDST \nXpert MTB/XDR pooled sensitivity and specificity were 93.1% (88.0 to 96.1) and 98.3% (94.5 to \n99.5.) (3 studies, 133",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Illustrates what the isoniazid results would mean per 1000 people tested and gives pooled accuracy for fluoroquinolone, ethionamide and amikacin resistance. It notes that results were similar whether or not people had rifampicin resistance and that sensitivity analyses changed little.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-633",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 633",
      "pageStart": 633,
      "pageEnd": 633,
      "target": "../sections/page-633.md",
      "narrative": null,
      "content": "• \nFor resistance to isoniazid, in people irrespective of rifampicin resistance, Xpert MTB/XDR \nsensitivity was 94.2% against a reference standard of pDST. \n• \nFor resistance to fluoroquinolones, in people irrespective of rifampicin resistance, Xpert \nMTB/XDR sensitivity was 93.1% against a reference standard of pDST. \n• \nFor resistance to ethionamide, in people irrespective of rifampicin resistance, Xpert \nMTB/XDR sensitivity was 56.6% against a reference standard of pDST. \n• \nFor resistance to amikacin, in people irrespective of rifampicin resistance, Xpert MTB/XDR \nsensitivity was 89.1% aga",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists the authors' conclusions on Xpert MTB/XDR sensitivity for each drug, with specificity above 97% in nearly all analyses, and calls for further studies in children, people with HIV, other settings and specimen types.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-634",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 634",
      "pageStart": 634,
      "pageEnd": 634,
      "target": "../sections/page-634.md",
      "narrative": null,
      "content": "BACKGROUND \n \nEarly recognition and improved characterisation of tuberculosis drug resistance is a prerequisite for \nthe rapid delivery of novel regimens to those who could benefit from them. For MDR/rifampicin- \nresistant-tuberculosis, the arrival of novel or repurposed drugs such as bedaquiline, clofazimine, and \nlinezolid has revolutionized the efficacy of longer regimens, dispensing with the need for injectable \ndrugs, and promising to deliver shorter all-oral regimens. Fluoroquinolones have an essential role and \nare also important for protecting second-line drugs like bedaquiline (WHO Co",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives background on the need for early detection of drug-resistant tuberculosis as new all-oral regimens emerge, noting the importance of fluoroquinolones and the gaps in culture-based drug susceptibility testing.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-635",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 635",
      "pageStart": 635,
      "pageEnd": 635,
      "target": "../sections/page-635.md",
      "narrative": null,
      "content": "annually, only 44% were detected and notified (WHO Global tuberculosis report 2020). The WHO \nrecommends that rapid techniques be used as the initial diagnostic tests to detect tuberculosis and \nrifampicin resistance in order to minimize delays in starting appropriate treatment (WHO \nConsolidated Guidelines (Module 3) 2020). The multiplexed nature of these new technologies \ntheoretically permits susceptibility to be detected accurately and comprehensively for a single drug \n(where variants in multiple genes may cause resistance) and to several different drugs, each with their \nown sets of dist",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes that fewer than half of estimated drug-resistant TB cases are detected and explains WHO's support for rapid initial tests and the potential of multiplexed molecular assays. It describes the Xpert MTB/XDR assay and its claimed advantages, such as results in under 90 minutes on existing GeneXpert platforms.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-636",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 636",
      "pageStart": 636,
      "pageEnd": 636,
      "target": "../sections/page-636.md",
      "narrative": null,
      "content": "The limit of detection for Mycobacterium tuberculosis by Xpert MTB/XDR (136 CFU/mL in \nunprocessed sputum) (Cepheid package insert 2020) is similar to that of Xpert MTB/RIF (112.6 \nCFU/mL), but higher than that of Xpert Ultra (15.6 CFU/mL) (Chakravorty 2017). The manufacturer \nstates that “Specimens with “MTB Trace DETECTED” results when tested with the Xpert MTB/RIF \nUltra Assay are expected to be below the limit of detection of the MTB/XDR Assay and are not \nrecommended for testing with the Xpert MTB/XDR Assay,” (Cepheid package insert 2020). As with \nXpert MTB/RIF and Xpert Ultra, Xpert MTB",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Compares the detection limit of Xpert MTB/XDR with other Xpert assays and explains how the test reports TB detection and resistance results for each drug.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-637",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 637",
      "pageStart": 637,
      "pageEnd": 637,
      "target": "../sections/page-637.md",
      "narrative": null,
      "content": "A clinical pathway presents a framework for developing recommendations about the use of a test and \nmay assist in assessing the effect of a new test on management decisions and patient-important \noutcomes (Gopalakrishna 2016). We considered several clinical scenarios in Appendix 3, \n \n \nIn this systematic review, the intended use of Xpert MTB/XDR is for diagnosis of drug resistance. \nThe role of the test would be a replacement test for culture-based phenotypic DST in people \ndiagnosed with tuberculosis irrespective of rifampicin resistance or with detected rifampicin \nresistance. \nThe downstre",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Sets out the clinical pathway for Xpert MTB/XDR as a replacement for culture-based susceptibility testing and describes the consequences of true and false positive and negative results. A footnote explains that the reviewed studies used the test as a reflex test on TB-positive sputum.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-638",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 638",
      "pageStart": 638,
      "pageEnd": 638,
      "target": "../sections/page-638.md",
      "narrative": null,
      "content": "METHODS \n \nTypes of studies \nWe included cross-sectional studies and cohort studies that assessed the diagnostic accuracy of the \nindex test. We included diagnostic accuracy studies in which cases and controls were sampled from a \nsingle source population (referred to as a single gate design). We excluded case-control studies where \ncases and controls were sampled from different populations (referred to as a two-gate design). The \nlatter type of study is prone to bias, particularly when a study enrols participants with severe disease \nand healthy participants without disease (Rutjes 2005). We",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the types of studies included, excluding two-gate case-control designs, along with participants, index test, target conditions and the phenotypic, genotypic and composite reference standards.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-639",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 639",
      "pageStart": 639,
      "pageEnd": 639,
      "target": "../sections/page-639.md",
      "narrative": null,
      "content": "Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). \nCure - we did not identify any studies that reported data for this outcome. \nMortality - we did not identify any studies that reported data for this outcome. \nTime to diagnosis - we did not identify any studies that reported data for this outcome. \nTime to start treatment - we did not identify any studies that reported data for this outcome. \n \nSearch methods \nWe searched the following databases: Ovid MEDLINE (OVID, 1946-present) and Embase (OVID, \n1947-present), for studies evaluating cartridge-ba",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists the review outcomes, noting that no studies reported cure, mortality or time to diagnosis or treatment, and describes the database searches, data extraction, quality assessment and meta-analysis methods.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p640-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 640,
      "pageEnd": 640,
      "target": "../images/img-p640-1.webp",
      "narrative": "draft",
      "content": "Fragment of a colour-coded table with columns 'Drug Class' and 'Result Call': for drug class N/A, the rows read INVALID/ERROR/NO RESULT (white), MTB DETECTED (red) and MTB NOT DETECTED (green). It illustrates how non-determinate Xpert MTB/XDR results relate to MTB DETECTED and MTB NOT DETECTED.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-vfig-p640",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 640,
      "pageEnd": 640,
      "target": "../figures/vfig-p640.png",
      "narrative": "draft",
      "content": "A nested-circle diagram of invalid or inconclusive index test results for Xpert MTB/XDR. Under TB detection, a non-determinate circle contains three result types: Error, “No result” and Invalid. Under drug susceptibility testing (reported only if the index test detects TB), a separate circle shows Indeterminate, for a specific drug, while susceptibilities to other drugs can still be reported.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-640",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 640",
      "pageStart": 640,
      "pageEnd": 640,
      "target": "../sections/page-640.md",
      "narrative": null,
      "content": "For multicentre studies, we anticipated that there would be variability in terms of how laboratory \npractices were carried out between different centres. For this reason, when data were available, we \nperformed meta-analyses at the centre level (i.e. treating each centre as a separate study). \n \nWe excluded MTBC-negative and inconclusive test results from analyses of diagnostic test accuracy. \n \nInconclusive index test results \nThe manufacturer defines two types of inconclusive results, non-determinate and indeterminate. \n \n \nFigure. 1. Overview of different types of inconclusive results for X",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains centre-level meta-analysis for multicentre studies and defines the two kinds of inconclusive Xpert MTB/XDR results, non-determinate and indeterminate, with an explanatory figure.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-641",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 641",
      "pageStart": 641,
      "pageEnd": 641,
      "target": "../sections/page-641.md",
      "narrative": null,
      "content": "MTB NOT DETECTED, we further determined the percentage that were resistant or susceptible \naccording to pDST. \n \nXpert MTB/XDR NON-DETERMINATE \nAmong the specimens initially tested, we determined the percentage of Xpert MTB/XDR NON- \nDETERMINATE results and, of these, the number of ERROR, INVALID, and NO RESULT results. \nWe also determined the percentage of non-determinate results remaining following retesting. \n \nXpert MTB/XDR INDETERMINATE \nAmong specimens reporting Xpert MTB/XDR MTB DETECTED, we determined the percentage that \nwere Xpert MTB/XDR INDETERMINATE (as drug resistance is only eva",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes how missed, non-determinate and indeterminate results were analysed and how the certainty of evidence was assessed with GRADE, including imprecision and publication bias.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-642",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 642",
      "pageStart": 642,
      "pageEnd": 642,
      "target": "../sections/page-642.md",
      "narrative": null,
      "content": "identified several unpublished studies, but no publications. We graded publication \nbias as undetected. \n \n \nRESULTS \n \nResults of the search \nWe identified and screened a total of 1,649 records. Of these, we excluded 1620 for relevance to the \ntopic. We retrieved 29 full text articles, including unpublished reports, and excluded 26 mainly \nbecause they were not rapid, low-complexity cartridge-based tests. We identified three unpublished \nstudies for inclusion in the review, Cepheid 2020, DIAMA 2020, and FIND 2020. Appendix 6 shows \nthe flow of studies in the review. A list of included and exc",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports search results, with three unpublished studies included out of 1,649 records screened, and summarizes the quality assessment and the countries where the studies were done.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p643-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 643,
      "pageEnd": 643,
      "target": "../images/img-p643-1.webp",
      "narrative": "draft",
      "content": "Forest plots of Xpert MTB/XDR sensitivity and specificity for detecting isoniazid resistance on sputum (Figure 3), in six panels: irrespective of rifampicin resistance and with detected rifampicin resistance, each against pDST, gDST and a composite reference standard. Studies are from Cepheid, DIAMA and FIND sites (e.g. China, South Africa, Benin, Cameroon, Rwanda, Moldova, Mumbai, New Delhi). Each row gives a study's TP, FP, FN and TN counts and sensitivity and specificity with 95% CI, plotted as squares with CI lines on 0–1 scales. Most sensitivities are 0.80–1.00 and most specificities 0.93–1.00, with wider intervals in the rifampicin-resistant panels.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p643-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 643,
      "pageEnd": 643,
      "target": "../images/img-p643-2.webp",
      "narrative": "draft",
      "content": "Risk of bias and applicability concerns chart (QUADAS-2) for Xpert MTB/XDR isoniazid-resistance studies (Figure 4), rating Cepheid 2020, DIAMA 2020 and FIND 2020 across patient selection, index test, reference standard and flow and timing. Nearly all ratings are low; Cepheid 2020 is unclear for patient selection and DIAMA 2020 is high for flow and timing. Key: red = high, yellow '?' = unclear, green = low.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-643",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 643",
      "pageStart": 643,
      "pageEnd": 643,
      "target": "../sections/page-643.md",
      "narrative": null,
      "content": "PICO questions \nShould MTB/XDR assay on sputum be used to diagnose isoniazid resistance in patients with \nmicrobiologically confirmed pulmonary TB? \n \n \nFigure 3. Forest plots of Xpert MTB/XDR sensitivity and specificity for detection of isoniazid resistance, by \npopulation and reference standard. Direct refers to testing directly on sputum. \n \n \nFigure 4. Xpert MTB/XDR, isoniazid resistance, risk of bias and applicability concerns.  \n \nShould MTB/XDR assay on sputum be used to diagnose fluoroquinolone resistance in patients with \nmicrobiologically confirmed pulmonary TB?",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists the review questions on using Xpert MTB/XDR to detect isoniazid and fluoroquinolone resistance, with captions for the forest plots and risk-of-bias figures.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p644-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 644,
      "pageEnd": 644,
      "target": "../images/img-p644-1.webp",
      "narrative": "draft",
      "content": "Forest plots of Xpert MTB/XDR sensitivity and specificity for detecting fluoroquinolone resistance on sputum (Figure 5), in six panels: irrespective of rifampicin resistance and with detected rifampicin resistance, each against pDST, gDST and a composite reference standard. Studies are from Cepheid, DIAMA and FIND sites (e.g. China, South Africa, Benin, Cameroon, Rwanda, Moldova, Mumbai, New Delhi). Each row gives a study's TP, FP, FN and TN counts and sensitivity and specificity with 95% CI, plotted as squares with CI lines on 0–1 scales. Estimable sensitivities are mostly 0.83–1.00; specificities are mostly 0.96–1.00, lower for FIND Mumbai (0.84 and 0.71) and FIND New Delhi.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-644",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 644",
      "pageStart": 644,
      "pageEnd": 644,
      "target": "../sections/page-644.md",
      "narrative": null,
      "content": "Figure 5. Forest plots of MTB/XDR sensitivity and specificity for detection of resistance to fluoroquinolones by \npopulation and reference standard. Direct refers to testing directly on sputum.  \n \n \nFigure 6. Xpert MTB/XDR, fluoroquinolone resistance, risk of bias and applicability concerns.  \n \nShould MTB/XDR assay on sputum be used to diagnose ethionamide resistance in patients \nwith microbiologically confirmed pulmonary TB?",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Captions for figures on fluoroquinolone resistance accuracy and the review question on ethionamide resistance.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p645-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 645,
      "pageEnd": 645,
      "target": "../images/img-p645-1.webp",
      "narrative": "draft",
      "content": "Forest plots of Xpert MTB/XDR sensitivity and specificity for detecting ethionamide resistance on sputum (Figure 7), in six panels: irrespective of rifampicin resistance and with detected rifampicin resistance, each against pDST, gDST and a composite reference standard. Studies are from Cepheid, DIAMA and FIND sites (e.g. China, South Africa, Benin, Cameroon, Rwanda, Moldova, Mumbai, New Delhi). Each row gives a study's TP, FP, FN and TN counts and sensitivity and specificity with 95% CI, plotted as squares with CI lines on 0–1 scales. Sensitivity against pDST and the composite standard is low to moderate (about 0.31–0.80), but higher against gDST (0.78–1.00); specificities are 0.86–1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p645-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 645,
      "pageEnd": 645,
      "target": "../images/img-p645-2.webp",
      "narrative": "draft",
      "content": "Risk of bias and applicability concerns chart (QUADAS-2) for Xpert MTB/XDR ethionamide-resistance studies (Figure 8), rating Cepheid 2020 and FIND 2020. FIND 2020 is low in all domains; Cepheid 2020 is unclear for patient selection and high for reference standard and flow and timing, with low applicability concerns. Key: red = high, yellow '?' = unclear, green = low.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-645",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 645",
      "pageStart": 645,
      "pageEnd": 645,
      "target": "../sections/page-645.md",
      "narrative": null,
      "content": "Figure 7. Forest plots of MTB/XDR sensitivity and specificity for detection of resistance to ethionamide by  \npopulation and reference standard. Direct refers to testing directly on sputum. \n \n \n \nFigure 8. Xpert MTB/XDR, ethionamide resistance, risk of bias and applicability concerns. \n \nShould MTB/XDR assay on sputum be used to diagnose amikacin resistance in patients with \nmicrobiologically confirmed pulmonary TB?",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Captions for figures on ethionamide resistance accuracy and the review question on amikacin resistance.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p646-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 646,
      "pageEnd": 646,
      "target": "../images/img-p646-2.webp",
      "narrative": "draft",
      "content": "Forest plots of Xpert MTB/XDR sensitivity and specificity for detecting amikacin resistance on sputum (Figure 9), in six panels: irrespective of rifampicin resistance and with detected rifampicin resistance, each against pDST, gDST and a composite reference standard. Studies are from Cepheid, DIAMA and FIND sites (e.g. China, South Africa, Benin, Cameroon, Rwanda, Moldova, Mumbai, New Delhi). Each row gives a study's TP, FP, FN and TN counts and sensitivity and specificity with 95% CI, plotted as squares with CI lines on 0–1 scales. Sensitivity varies from 0.29 to 1.00, lowest at FIND New Delhi, and some DIAMA estimates are not estimable; specificities are 0.96–1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-646",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 646",
      "pageStart": 646,
      "pageEnd": 646,
      "target": "../sections/page-646.md",
      "narrative": null,
      "content": "Figure 9. Forest plots of MTB/XDR sensitivity and specificity for detection of resistance to amikacin by population \nand reference standard. Direct refers to testing directly on sputum.  \n \n \n \nFigure 10. Xpert MTB/XDR, amikacin resistance, risk of bias and applicability concerns.  \n \nTable 2. Performance of Xpert MTB/XDR for resistance to isoniazid, fluoroquinolones, \nethionamide, and amikacin \nDrug \nReference \nstandard \nNo. studies \n(participants) \nNo. (%) with \ndrug \nresistance \nPooled \nsensitivity % \n(95% CI) \nPooled \nspecificity % \n(95% CI) \nPositive \npredictive \nvalue % (95% \nCI) 1 \nNega",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Captions for figures on amikacin resistance accuracy and the heading of a table summarizing Xpert MTB/XDR performance for each drug.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-647",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 647",
      "pageStart": 647,
      "pageEnd": 647,
      "target": "../sections/page-647.md",
      "narrative": null,
      "content": "Irrespective of rifampicin resistance \nINH \npDST \n3 (1605) \n994 (61.9) \n94.2 \n(89.3 to 97.0) \n98.0 \n(95.2 to 99.2) \n71.3 \n(50.1 to 86.0) \n99.7 \n(99.4 to 99.8) \nINH \ngDST \n2 (999) \n682 (68.3) \n97.3 \n(92.8 to 99.0) \n98.4 \n(95.9 to 99.3) \n75.6 \n(55.4 to 88.6) \n99.9 \n(99.6 to 100.0) \nINH \nComposite 2 (1055) \n768 (72.8) \n93.6 \n(86.5 to 97.1) \n99.7 \n(96.6 to 100.0) \n94.2 \n(58.6 to 99.5) \n99.7 \n(99.3 to 99.8) \nWith detected rifampicin resistance \nINH \npDST \n2 (744) \n684 (91.9) \n97.2 \n(89.7 to 99.3) \n91.5 \n(68.5 to 98.1) \n83.0 \n(51.2 to 95.8) \n99.1 \n(96.6 to 99.8) \nINH \ngDST \n1 (434) \n416 (95.9) \n98.4",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A table of pooled sensitivity, specificity and predictive values of Xpert MTB/XDR for isoniazid, fluoroquinolone, ethionamide and amikacin resistance, by population and reference standard. Footnotes explain the abbreviations and the prevalence assumptions used for predictive values.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-648",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 648",
      "pageStart": 648,
      "pageEnd": 648,
      "target": "../sections/page-648.md",
      "narrative": null,
      "content": "As seen in Table 2, for each drug, Xpert MTB/XDR pooled sensitivity and specificity estimates were \nsimilar in people irrespective of rifampicin resistance and people with detected rifampicin resistance. \nHowever, owing to enrolment criteria in the studies, we note that most participants were rifampicin \nresistant in all analyses. \n \nPICO questions \n1. \nShould MTB/XDR assay on sputum be used to diagnose INH resistance in patients with \nmicrobiologically confirmed pulmonary TB, irrespective of resistance to RIF, MRS? \n2. \nShould MTB/XDR assay on sputum be used to diagnose INH resistance in pati",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes that accuracy was similar regardless of rifampicin resistance and lists the 18 review questions by drug, population and reference standard.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-649",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 649",
      "pageStart": 649,
      "pageEnd": 649,
      "target": "../sections/page-649.md",
      "narrative": null,
      "content": "Table 3. GRADE Certainty of Evidence \nSee Supplement C. GRADE evidence profiles. \n \nPICO \nDrug \nPopulation \nReference \nstandard \nNo. studies \n(participants) \nPooled \nsensitivity \n% (95% CI) \nPooled \nspecificity % \n(95% CI) \nCertainty \nEvidence \nSens \nCertainty \nEvidence \nSpec \nExplanations \n1 \nINH \nIrrespective \nrifampicin \nresistance \npDST \n3 (1605) \n94.2 \n(89.3, 97.0) \n98.0 \n(95.2, 99.2) \nModerate \nModerate \nDowngraded one \nlevel for \nindirectness for \nsensitivity and \nspecificity \n2 \nINH \nIrrespective \nrifampicin \nresistance \nCRS \n2 (1055) \n93.6 \n(86.5, 97.1) \n99.7 \n(96.6, 100.0) \nModerate",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A GRADE table showing the certainty of evidence for Xpert MTB/XDR sensitivity and specificity for each drug and reference standard, with reasons for downgrading.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-650",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 650",
      "pageStart": 650,
      "pageEnd": 650,
      "target": "../sections/page-650.md",
      "narrative": null,
      "content": "level for risk of \nbias (specificity) \n12 \nETO \nWith \ndetected \nrifampicin \nresistance \npDST \n1 (492) \n51.7 \n(33.1, 69.8) \n94.8 \n(84.8, 98.3) \nVery Low Moderate \nDowngraded one \nlevel for risk of \nbias, one level for \ninconsistency, \none level for \nimprecision \n(sensitivity); \ndowngraded one \nlevel for risk of \nbias (specificity) \n13 \nETO \nWith \ndetected \nrifampicin \nresistance \ngDST \n1 (434) \n98.0 \n(74.2, 99.9) \n99.7 \n(83.5, 100.0) \nVery Low Very Low Downgraded two \nlevels for risk of \nbias, one level for \nimprecision \n(sensitivity and \n(specificity) \n14 \nETO \nWith \ndetected \nrifampicin \nresi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continuation of the GRADE certainty table for ethionamide and amikacin resistance, with abbreviations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-651",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 651",
      "pageStart": 651,
      "pageEnd": 651,
      "target": "../sections/page-651.md",
      "narrative": null,
      "content": "Xpert MTB/XDR MTB NOT DETECTED and inconclusive test results \n \nXpert MTB/XDR MTB NOT DETECTED \nHere we summarize results for Xpert MTB/XDR MTB NOT DETECTED and resistant cases \ntherefore missed. Cepheid 2020 was the only study that reported this information. \n \nIsoniazid \n \nOf 530 specimens tested, 512 had pDST results available. Of these 512 specimens with pDST results \navailable, 32 (6.3%) were Xpert MTB/XDR MTB NOT DETECTED. \nBy the pDST reference standard, of these 32 specimens, two (6.3%) were resistant and 30 (93.8%) \nwere susceptible. \n \nFluoroquinolones \n \nOf 530 specimens tested, 453",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports how many specimens were missed as TB not detected by Xpert MTB/XDR and how many were resistant by phenotypic testing, as well as rates of non-determinate results before and after retesting.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-652",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 652",
      "pageStart": 652,
      "pageEnd": 652,
      "target": "../sections/page-652.md",
      "narrative": null,
      "content": "647 \n \n \n \n \n \nIndeterminate test results \nHere we provide a summary of indeterminate results and their pDST status. \n \nIsoniazid \n \nCepheid 2020 \nOf 530 specimens tested, 498 specimens had an Xpert MTB/XDR MTB DETECTED result. Of these \n498 specimens, two (0.4%) had indeterminate results for detection of resistance. \nBy the pDST reference standard, of these two specimens, two (100%) were resistant and zero (0%) \nwere susceptible. \n \nFIND 2020 \nOf 709 specimens tested, 657 had an Xpert MTB/XDR MTB DETECTED result. Of these 657 \nspecimens, two (0.3%) had indeterminate results for detection of r",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarizes indeterminate Xpert MTB/XDR results for each drug in two studies, with their phenotypic resistance status where known.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-653",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 653",
      "pageStart": 653,
      "pageEnd": 653,
      "target": "../sections/page-653.md",
      "narrative": null,
      "content": "648 \n \n \n \n \nSensitivity analyses \n \nTable 4 presents the findings from sensitivity analyses that excluded data from the manufacturer. \nThere are two rows of results presented for each drug. The first row presents the results of the meta- \nanalysis including Cepheid 2020, and the subsequent row, the results of the sensitivity analysis \nexcluding Cepheid 2020 (in bold). \n \nThese sensitivity analyses made little difference to any of the findings. \n \nTable 4. Xpert MTB/XDR accuracy for drug resistance in people irrespective of rifampicin \nresistance, sensitivity analyses \n \nDrug \nReference \nstand",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents sensitivity analyses excluding the manufacturer's study, showing that removing it made little difference to the accuracy estimates.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-654",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 654",
      "pageStart": 654,
      "pageEnd": 654,
      "target": "../sections/page-654.md",
      "narrative": null,
      "content": "649 \n \n \n \n \nDISCUSSION \n \nSummary of main results \n \nThis systematic review summarizes the current literature and included three unpublished studies on \nthe accuracy of Xpert MTB/XDR for detection of resistance to isoniazid, fluoroquinolones, \nethionamide, and amikacin. \n \nXpert MTB/XDR for isoniazid resistance in people with microbiologically confirmed pulmonary \ntuberculosis irrespective of rifampicin resistance, pDST \nXpert MTB/XDR pooled sensitivity and specificity (95% confidence interval) were 94.2% (89.3 to \n97.0) and 98.0% (95.2 to 99.2) (3 studies, 1605 participants, 61.9% with isoni",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Begins the discussion by summarizing the main results for each drug and what they would mean per 1000 people tested.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-655",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 655",
      "pageStart": 655,
      "pageEnd": 655,
      "target": "../sections/page-655.md",
      "narrative": null,
      "content": "650 \n \n \n \n \nXpert MTB/XDR pooled sensitivity and specificity were 89.1% (80.9. to 94.1) and 99.5% (96.9 to \n99.9) (2 studies, 1008 participants, 15.0% with amikacin resistance; high-certainty evidence for \nsensitivity and specificity). \n \nResults of these studies indicate that in theory, of 1000 people where 50 have amikacin resistance, 50 \nwould be Xpert MTB/XDR-positive: of these, 5 (10%) would not have amikacin resistance (false- \npositives) and 950 would be Xpert MTB/XDR-negative: of these, 5 (1%) would have amikacin \nresistance (false-negatives). \n \nAUTHORS' CONCLUSIONS \n \n• \nFor resista",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the summary of amikacin results and sets out the authors' conclusions and research recommendations, such as using comprehensive genotypic reference standards and studying different patient groups.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-656",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 656",
      "pageStart": 656,
      "pageEnd": 656,
      "target": "../sections/page-656.md",
      "narrative": null,
      "content": "651 \n \n \n \n \ndiagnostic accuracy of Xpert MTB/XDR for pulmonary tuberculosis in adults, children, and people \nliving with HIV and in people who are smear negative. \n \nAcknowledgements \n \nWe are grateful to Vittoria Lutje, Cochrane Infectious Diseases' Information Specialist, for help with the search \nstrategy. We are grateful to people at Cepheid and FIND and to all members of the DIAMA consortium for \nproviding data and answering our many questions. \n \nDeclarations of Interest \n \nSP received funding from the World Health Organization Global Tuberculosis Programme, Geneva. \nGRD has no conflict",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Ends the research recommendations and gives the acknowledgements, declarations of interest and the start of the reference list.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-657",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 657",
      "pageStart": 657,
      "pageEnd": 657,
      "target": "../sections/page-657.md",
      "narrative": null,
      "content": "652 \n \n \n \n \nGRADEpro GDT [Computer program]. Version accessed 29 October 2016. Hamilton (ON): McMaster \nUniversity (developed by Evidence Prime), 2015. Available at gradepro.org. \nGopalakrishna 2016 \nGopalakrishna G, Langendam MW, Scholten RJ, Bossuyt PM, Leeflang MM. Defining the clinical pathway in \nCochrane diagnostic test accuracy reviews. BMC Med Res Methodol. 2016;16(1):153. \nHeyckendorf 2017 \nHeyckendorf J, Andres S, Köser CU, Olaru ID, Schön T, Sturegård E, et al. What is resistance? Impact of \nphenotypic versus molecular drug resistance testing on therapy for multi-and extensively dr",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continuation of the reference list, citing methodological sources and WHO tuberculosis guidance and reports.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-658",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 658",
      "pageStart": 658,
      "pageEnd": 658,
      "target": "../sections/page-658.md",
      "narrative": null,
      "content": "653 \n \n \n \n \nWHO Critical concentrations 2018 \nWorld Health Organization. Technical Report on critical concentrations for drug susceptibility testing of \nmedicines used in the treatment of drug-resistant tuberculosis. (WHO/CDS/TB/2018.5). License: CC BY-NC- \nSA 3.0 IGO. who.int/tb/publications/2018/WHO_technical_report_concentrations_TB_drug_susceptibility/en/ \n(accessed 16 July 2020). \nWHO Definitions and reporting 2020 \nWorld Health Organization. Definitions and reporting framework for tuberculosis – 2013 revision (updated \nDecember 2014 and January 2020). who.int/tb/publications/definitions",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The final references, citing WHO technical reports on critical concentrations and on tuberculosis definitions and reporting.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-659",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 659",
      "pageStart": 659,
      "pageEnd": 659,
      "target": "../sections/page-659.md",
      "narrative": null,
      "content": "654 \n \n \n \n \nAPPENDICES \n \nAppendix 1. Glossary of terms \n \nAmplification \nAmplification is replication of a DNA fragment to generate copies. Both the original and the \nnewly synthesized copies can be described as the amplicons. \nCodon \nA codon is a sequence of three DNA or RNA bases that corresponds to a specific amino acid \nor a signal to start or stop transcription or translation. The DNA in coding regions of the \ngenome is read in groups of three bases (A, G, C, T). \nCritical concentration \nThe critical concentration of an anti-tuberculous agent has been adopted and modified from \ninternat",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The start of a glossary of terms used in the Xpert MTB/XDR review, defining terms such as amplification, codon, critical concentration and gene targets of the assay.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-660",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 660",
      "pageStart": 660,
      "pageEnd": 660,
      "target": "../sections/page-660.md",
      "narrative": null,
      "content": "655 \n \n \n \n \nHeteroresistance is defined as resistance to certain antibiotics in a subset of a larger \nmicrobial population that is generally considered to be susceptible to these antibiotics \naccording to traditional phenotypic drug susceptibility testing. \nIndeterminate test result \nAn indeterminate Xpert MTB/XDR test result is one that indicates that resistance to a given \ndrug could not definitively be detected based on the test’s algorithm. \ninhA promoter \nGene target included in the Xpert MTB/XDR test to detect MTB and resistance to isoniazid \nand ethionamide. Mutations in the inhA promo",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continuation of the glossary, defining heteroresistance, indeterminate results, the inhA promoter and other gene targets, presumptive tuberculosis and resistance-determining regions.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-661",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 661",
      "pageStart": 661,
      "pageEnd": 661,
      "target": "../sections/page-661.md",
      "narrative": null,
      "content": "656 \n \n \n \n \nGene target included in the Xpert MTB/XDR test to detect mutations that confer resistance \nto second line injectable drugs, amikacin, kanamycin, and capreomycin. \nSanger sequencing \nTechnique for DNA sequencing based upon the selective incorporation of chain-terminating \ndideoxynucleotides by DNA polymerase during in vitro DNA replication, also known as ‘the \nchain termination method’. \nTargeted gene sequencing \nThe process for detecting predetermined mutations in DNA or genomic regions. \nWhole genome sequencing (WGS) \nThe process of determining the complete genome sequence for a",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The end of the glossary, defining the rrs gene target, Sanger sequencing, targeted gene sequencing and whole genome sequencing.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p662-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 662,
      "pageEnd": 662,
      "target": "../images/img-p662-1.webp",
      "narrative": "draft",
      "content": "Photo of an Xpert MTB/XDR cartridge beside a colour-coded table of possible result calls by drug class (N/A, isoniazid, fluoroquinolone, amikacin, kanamycin, capreomycin, ethionamide). Results such as 'Resistance DETECTED' and 'Low Resistance DETECTED' are shaded red, 'NOT DETECTED' green and 'INDETERMINATE' yellow.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-662",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 662",
      "pageStart": 662,
      "pageEnd": 662,
      "target": "../sections/page-662.md",
      "narrative": null,
      "content": "657 \n \n \n \n \nAppendix 2. Possible test results for each target in the Xpert MTB/XDR assay \n \n \nPositive results for the Xpert MTB/XDR assay can be MTB DETECTED and all resistance targets are NOT \nDETECTED, or MTB DETECTED and one or more of the resistance targets is DETECTED, or MTB \nDETECTED and/or one or more of the following resistance targets is INDETERMINATE. Copyright © [2020] \n[Cepheid Inc]: reproduced with permission.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "An appendix outlining the possible results for each target of the Xpert MTB/XDR assay, reproduced with permission from the manufacturer.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p663-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 663,
      "pageEnd": 663,
      "target": "../images/img-p663-1.webp",
      "narrative": "draft",
      "content": "Clinical pathway diagram with four flowcharts, Scenarios A to D, under an arrow labelled 'Increasing likelihood of resistance'. A: index test to diagnose TB and do DST; B: DST in patients newly diagnosed with TB by another test; C: DST in newly diagnosed patients whose RIF susceptibility is already known; D: DST in patients on treatment. Each branches from the index test into TB-negative/TB-positive, no resistance or resistance to at least one drug, and ends in treatment.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-663",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 663",
      "pageStart": 663,
      "pageEnd": 663,
      "target": "../sections/page-663.md",
      "narrative": null,
      "content": "658 \n \n \n \n \nAppendix 3. Figure. Clinical pathway \n \n \n \nThe index test may be used in the following scenarios. \nA. Index test used to diagnose tuberculosis and detect drug resistance. \nB. Index test used to detect drug resistance in patients newly diagnosed with tuberculosis by another test where \nrifampicin susceptibility is unknown. Proposed role of Xpert MTB/XDR would be an initial test for resistance to \nisoniazid and second-line drugs (replacement for LPAs and culture-based DST as initial tests). \nC. Index test used to detect drug resistance in patients newly diagnosed with tuberculosis",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "An appendix describing the clinical pathway scenarios in which Xpert MTB/XDR could be used and how results would inform treatment regimens.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-664",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 664",
      "pageStart": 664,
      "pageEnd": 664,
      "target": "../sections/page-664.md",
      "narrative": null,
      "content": "659 \n \n \n \n \n \nDatabase: Ovid MEDLINE(R) and Epub Ahead of Print, In-Process & Other Non-Indexed Citations, \nDaily and Versions(R) <1946 to present> \n-------------------------------------------------------------------------------- \n \n1 Extensively Drug-Resistant Tuberculosis/ or Tuberculosis, Multidrug-Resistant/ or Tuberculosis, \nPulmonary/ \n \n2 (tuberculosis adj3 (lung or pulmonary)).mp. \n3 (tuberculosis adj3 respiratory).mp. \n4 (isoniazid resistance or isoniazid resistant).mp. \n \n5 ((Fluoroquinolone adj3 resistance) or (Fluoroquinolone adj3 resistant)).mp. \n6 (Second-line injectable drug ad",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The MEDLINE search strategy used for the Xpert MTB/XDR review, combining terms for drug-resistant tuberculosis and cartridge-based tests from 2015 onward.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-665",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 665",
      "pageStart": 665,
      "pageEnd": 665,
      "target": "../sections/page-665.md",
      "narrative": null,
      "content": "660 \n \n \n \n \nAppendix 5. QUADAS-2 \n \nDomain 1: Patient selection \nDetection of tuberculosis \nRisk of bias: Could the selection of patients have introduced bias? \n \nSignalling question 1: Was a consecutive or random sample of patients enrolled? \nWe answered yes if the study enrolled a consecutive or random sample of eligible patients; no if the study \nselected patients by convenience; and unclear if the study did not report the manner of patient selection or was \nnot clearly reported. \nSignalling question 2: Was a case-control design avoided? \nWe answered yes if the study enrolled patients with",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Sets out how the QUADAS-2 tool was applied for patient selection, explaining how each signalling question was answered and how applicability concerns were judged.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-666",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 666",
      "pageStart": 666,
      "pageEnd": 666,
      "target": "../sections/page-666.md",
      "narrative": null,
      "content": "661 \n \n \n \n \nRisk of bias: could the conduct or interpretation of the index test have introduced bias? \n \nSignalling question 1: were the index test results interpreted without knowledge of the results of the reference \nstandard? \nWe answered this question yes for all studies where results are automatically generated and the user is provided \nwith printable test results. Thus, there is no room for subjective interpretation of test results. For those assays, \nwhich require user interpretation, we answered yes if the reader of the assay was blinded to results of reference \ntests. We answered no",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the QUADAS-2 criteria for the index test and reference standard domains, explaining how blinding and the type of culture were judged.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-667",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 667",
      "pageStart": 667,
      "pageEnd": 667,
      "target": "../sections/page-667.md",
      "narrative": null,
      "content": "662 \n \n \n \n \nDetection of drug resistance \nRisk of bias: could the reference standard, its conduct or its interpretation have introduced bias? \n \nSignalling question 1: is the reference standard likely to correctly classify the target condition? \nWe answered these questions for each target condition separately by reference standard as follows. \nDrug \npDST \ngDST, targeted \nsequencing \nComposite (pDST \nand gDST, targeted \nsequencing) \ngDST, whole \ngenome \nsequencing) \nComposite (pDST \nand gDST, whole \ngenome sequencing) \nIsoniazid \nYes \nUnclear if few loci \nare investigated, \nand yes, if all \nre",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains how the reliability of each reference standard was judged for detecting resistance to each drug, with a table of required gene loci, and introduces the question of testing on the same material.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-668",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 668",
      "pageStart": 668,
      "pageEnd": 668,
      "target": "../sections/page-668.md",
      "narrative": null,
      "content": "663 \n \n \n \n \nresistant strains present in the original specimen (which would have been used for the index test if direct testing \nwas done) and cause discrepant results. We think this is an important question as it addresses heteroresistance, \nwhich often explains discordance between genotypic and phenotypic results. \nFor direct testing of a clinical specimen by the index test: we answered yes if the reference test was done \ndirectly on the same clinical specimen; no if the reference standard was done on a culture isolate; and unclear if \nwe could not tell. For indirect testing of a culture is",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains how bias from testing different material types and heteroresistance was assessed and gives a table of critical concentrations and breakpoints for drugs used against drug-resistant tuberculosis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-669",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 669",
      "pageStart": 669,
      "pageEnd": 669,
      "target": "../sections/page-669.md",
      "narrative": null,
      "content": "664 \n \n \n \n \nGatifloxacin \n0.5 \n- \n- \n0.25 \nSecond-line \nAmikacin \n30.0 \n2.0 \n- \n1.0 \nOther \nEthionamide \n40.0 \n5.0 \n10.0 \n5.0 \nAbbreviations: CB: critical breakpoint; CC: critical concentration",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The end of the table of critical concentrations, covering gatifloxacin, amikacin and ethionamide.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p670-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 670,
      "pageEnd": 670,
      "target": "../images/img-p670-1.webp",
      "narrative": "draft",
      "content": "Flow diagram of studies in the review: 1647 records from database searching and 3 from other sources; 1649 after duplicates removed and screened; 1620 excluded; 29 full-text articles (datasets) assessed for eligibility, of which 26 were excluded (not a cartridge-based test for isoniazid or second-line drug resistance 17, index test is BD MAX 3, prototype test 2, not a diagnostic accuracy study 2, extrapulmonary specimen 1, not commercially available 1). Three studies were included in both qualitative and quantitative synthesis (meta-analysis).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-670",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 670",
      "pageStart": 670,
      "pageEnd": 670,
      "target": "../sections/page-670.md",
      "narrative": null,
      "content": "665 \n \n \n \n \nAppendix 6. Flow of studies in the review",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A heading for the appendix on the flow of studies through the review.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-671",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 671",
      "pageStart": 671,
      "pageEnd": 671,
      "target": "../sections/page-671.md",
      "narrative": null,
      "content": "666 \n \n \n \n \nAppendix 7. List of studies \n  \nIncluded studies  \n \nCepheid 2020 \nClinical evaluation of the Xpert® MTBXDR assay \nReport R244C2 Xpert MTB/XDR Rev 1.0 \nSponsor: Cepheid, Sunnyvale, USA \n \nDIAMA 2020 \nDIAgnostics for Multidrug Resistant Tuberculosis in Africa \nClinicalTrials.gov Identifier: NCT03303963 \nSponsor: Dissou Affolabi, Kigali, Rwanda \n \nFIND 2020 \nAnalytical Performance and Clinical Diagnostic Accuracy of the Xpert MTB/XDR Assay for TB and Expanded \nResistance Detection, September 2020 \nClinicalTrials.gov Identifier: NCT03728725 \nSponsor: Foundation for Innovative New Dia",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists the studies included in the review of the Xpert MTB/XDR assay, such as Cepheid, DIAMA and FIND evaluations, then begins a list of excluded studies with the reason for each exclusion. The exclusions include studies that did not test a cartridge-based assay for isoniazid or second-line drug resistance and studies that were not diagnostic accuracy studies. Entries give full bibliographic details.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-672",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 672",
      "pageStart": 672,
      "pageEnd": 672,
      "target": "../sections/page-672.md",
      "narrative": null,
      "content": "667 \n \n \n \n \nTuberculosis and Respiratory Diseases. 2015;38(9):680-5. (Not a cartridge-based test for isoniazid or second-line drug \nresistance) \n14. \nKim S, Kim Y, Chang Y, Hirgo WK, Chang CL, Shim T-S, et al. Comparison of Quantamatrix multiplexed assay \nplatform and GenoType MTBDR assay using smear-positive sputum specimens from patients with multidrug- \nresistant/extensively drug-resistant tuberculosis in South Korea. Frontiers in Microbiology. 2019;10:1075. (Not a cartridge- \nbased test for ISONIAZID or second-line drug resistance) \n15. \nKlotoe BJ, Molina-Moya B, Gomes HM, Gomgnimbou MK,",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the list of excluded studies for the Xpert MTB/XDR review, giving each citation and its reason for exclusion, such as being a prototype test or a different assay.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-673",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 673",
      "pageStart": 673,
      "pageEnd": 673,
      "target": "../sections/page-673.md",
      "narrative": null,
      "content": "668 \n \n \n \n \nAppendix 8. Table. Characteristics of included studies \n \nStudy year \nCountries/centres \n(High MDR \nBurden) \nStudy \ndesign \nNumber of \npatients (% \ndetected RR) \nAge of \nenrolment \nPLHIV \nReference \nstandard \nfor drug \nresistance \nLoci included in \ngDST reference \nstandard \nCepheid \n2020 \nChina (yes) \nSouth Africa \n(yes) \ncross- \nsectional, \nretrospective1 \n530 (47.9%) \n≥ 15 \nyears2 \nNR \npDST \ngDST \ncomposite \nkatG, inhA \npromoter, oxyR- \nahpC intergenic \nregion, fabG1, \ngyrA, gyrB, rrs, eis \npromoter \nDIAMA \n2020 \nBenin (no) \nCameroon (no) \nRwanda (no) \ncross- \nsectional, \nprospe",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A table of characteristics of the studies included in the Xpert MTB/XDR review, giving countries, design, number of patients, age, HIV status and the reference standards and gene loci used.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p674-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 674,
      "pageEnd": 674,
      "target": "../images/img-p674-1.webp",
      "narrative": "draft",
      "content": "Product photo of the Sanity 2.0 all-in-one benchtop instrument used for MeltPro XDR-TB testing (Figure S1): a white cabinet with a touchscreen at the top and a front opening, bearing the Zeesan name.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-674",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 674",
      "pageStart": 674,
      "pageEnd": 674,
      "target": "../sections/page-674.md",
      "narrative": null,
      "content": "669 \n \n \n \n \nSUPPLEMENTARY INFORMATION \n \nSupplement A. MeltPro® XDR-TB \n \nBackground \nOn 30 October 2020, we were notified by the WHO Global Tuberculosis Programme about a clinical \nstudy conducted in China evaluating MeltPro® XDR-TB (Xiamen Zeesan Biotech Co., Ltd., China), \na low complexity test for resistance to anti-tuberculosis drugs. The WHO provided us with a report \nsummarizing the clinical study. We corresponded with study authors for additional information and \nclarifications. \n \nMeltPro XDR-TB is a commercially available, low complexity test for detection of mutations \nassociated w",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces a supplement on MeltPro XDR-TB, a low-complexity test for TB drug resistance evaluated in a Chinese clinical study reported to the reviewers by WHO, and describes the study's setting and participants.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-675",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 675",
      "pageStart": 675,
      "pageEnd": 675,
      "target": "../sections/page-675.md",
      "narrative": null,
      "content": "670 \n \n \n \n \nSputum samples were selected from patients who had submitted specimens for culture and subsequent \ndrug-susceptibility testing during the routine work of the clinical laboratory of the hospital. All \nselected samples were first tested by the MeltPro®MTB Test Kit (Xiamen Zeesan Biotech Co., Ltd.) \nand if found to be TB positive, the sample was eligible for inclusion. The authors stated, “In other \nwords, samples selected for enrolment were by convenience.” Samples were included if they were \ntuberculosis positive, had DST results, and at least 2 mL were available after the laborato",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes how sputum samples were selected for the MeltPro XDR-TB study and how discordant results with phenotypic testing were investigated by sequencing, starting with isoniazid.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p676-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 676,
      "pageEnd": 676,
      "target": "../images/img-p676-1.webp",
      "narrative": "draft",
      "content": "Forest plots of MeltPro XDR-TB sensitivity and specificity on sputum against pDST (Figure S2), from a single study (Zeesan 2020), in six panels for isoniazid, fluoroquinolone and amikacin resistance, each in rifampicin-susceptible and rifampicin-resistant patients. Each row gives a study's TP, FP, FN and TN counts and sensitivity and specificity with 95% CI, plotted as squares with CI lines on 0–1 scales. Sensitivities range from 0.79 to 0.91 and specificities from 0.86 to 0.98.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-676",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 676",
      "pageStart": 676,
      "pageEnd": 676,
      "target": "../sections/page-676.md",
      "narrative": null,
      "content": "671 \n \n \n \n \nisoniazid sensitive by MeltPro XDR-TB. Sequencing results showed that none of them showed any \nmutation in the coverage area of the probes. \nFluoroquinolones: There were 20 samples whose DST was fluoroquinolone sensitive but detected as \nfluoroquinolone resistant by MeltPro XDR-TB. Sequencing results showed that all of them had \nmutations in gyrA. There were 10 samples whose DST was fluoroquinolone resistant but detected as \nfluoroquinolone sensitive by MeltPro XDR-TB. Sequencing results showed that two samples had \nD94G heterogenic mutation, while the remaining eight samples show",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports sequencing findings for MeltPro XDR-TB results that disagreed with phenotypic testing for isoniazid and fluoroquinolones, with forest plots of sensitivity and specificity by rifampicin resistance status.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p677-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 677,
      "pageEnd": 677,
      "target": "../images/img-p677-1.webp",
      "narrative": "draft",
      "content": "Forest plots of Xpert MTB/XDR sensitivity and specificity for diagnosing pulmonary TB on sputum against culture, in three panels: adults (6 studies), children under 15 (no data) and people with HIV (3 studies). Each row gives a study's TP, FP, FN and TN counts and sensitivity and specificity with 95% CI, plotted as squares with CI lines on 0–1 scales. Sensitivity is high (0.86–0.99), while specificity in adults varies widely (0.00–1.00) with very wide intervals.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-677",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 677",
      "pageStart": 677,
      "pageEnd": 677,
      "target": "../sections/page-677.md",
      "narrative": null,
      "content": "672 \n \n \n \n \nSupplement B. Should Xpert MTB/XDR assay on sputum be used to diagnose PTB in people \nwith signs and symptoms of TB? \n \n \n \nFigure. Forest plots of Xpert MTB/XDR sensitivity and specificity for the diagnosis of pulmonary tuberculosis \nby population, culture reference standard. Direct refers to testing directly on sputum",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A supplement showing forest plots of Xpert MTB/XDR accuracy for diagnosing pulmonary TB by population against culture.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-678",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 678",
      "pageStart": 678,
      "pageEnd": 678,
      "target": "../sections/page-678.md",
      "narrative": null,
      "content": "673 \n \n \n \n \nSupplement C. GRADE evidence profiles \n \n1. Question: Should MTB/XDR assay on sputum be used to diagnose INH resistance in patients with microbiologically confirmed pulmonary TB, \nirrespective of resistance to RIF, pDST? \nSensitivity \n0.94 (95% CI: 0.89 to 0.97) \nSpecificity \n0.98 (95% CI: 0.95 to 0.99) \nOutco\nm e \n№ of \nstudie \ns (№ \nof \npatien\nt s) \nStudy \ndesign \nFactors that may decrease certainty of evidence \nEffect per 1,000 patients \ntested \nTest \naccurac\ny CoE \nRisk \nof \nbias \nIndirect\nn ess \nInconsist\ne ncy \nImprecis\ni on \nPublicati \non bias \npre-test \nprobabil \nity of \n2",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A GRADE evidence profile for Xpert MTB/XDR detection of isoniazid resistance against phenotypic testing, with pooled sensitivity of 0.94 and specificity of 0.98, effects per 1000 patients at different prevalences, and certainty judgements.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-679",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 679",
      "pageStart": 679,
      "pageEnd": 679,
      "target": "../sections/page-679.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Please do not distribute further. \n \n \n674\n \n \n \nWeb Annex B.15. Drug concentrations used in culture-based drug susceptibility testing for included studies on \nsecond-line line probe assays diagnostic accuracy7 \nTable 1. Ofloxacin, levofloxacin, and moxifloxacin, drug concentrations used in culture-based drug susceptibility testing in relation to the WHO- \nrecommended critical concentrations \nStudy \nReference standard \nConcentration used \n(ug/ml) \nMet WHO-recommended \ncritical concentration \nComments \nAjbani 2012 \nMGIT 960 \nOfloxac",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Begins Web Annex B.15, a table of fluoroquinolone drug concentrations used in culture-based susceptibility testing in studies of second-line line probe assays, compared with WHO-recommended critical concentrations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-680",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 680",
      "pageStart": 680,
      "pageEnd": 680,
      "target": "../sections/page-680.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Please do not distribute further. \n \n \n675\n \n \nHillemann 2009 \nMGIT 960 and LJ \nOfloxacin: 2.0 for both \nmedia \nYes for MGIT; no for LJ \nHuang 2011 \nMGIT 960 and Middlebrook 7H11 \nOfloxacin: 2.0 \nYes for both media \nIgnatyeva 2012 \nMGIT 960 \nOfloxacin: 2.0 \nYes \nJin 2013 \nLJ and BacT/ALERT 3D \nOfloxacin: 5.0 (LJ); 50 \n(BacT/ALERT 3D) \nNo for LJ; Not applicable \nBacT/ALERT 3D \nNo WHO- recommended concentration \nspecified for BacT/ALERT 3D media \nKambli 2015a \nMGIT 960 \nOfloxacin: 2.0 \nYes \nMoxifloxacin: 0.25 \nNo \nKambli 2015b \nMGIT",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the table of ofloxacin, levofloxacin and moxifloxacin concentrations used by each study, by culture medium, and whether they matched WHO critical concentrations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-681",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 681",
      "pageStart": 681,
      "pageEnd": 681,
      "target": "../sections/page-681.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Please do not distribute further. \n \n \n676\n \n \nTomasicchio \n2016 \nMGIT 960 \nOfloxacin: 2.0 \nYes \nTukvadze 2014 \nLJ (proportion method) \nOfloxacin: 2.0 \nNo \nvan Ingen 2010 \nMiddlebrook 7H10 (agar \nproportion) \nMoxifloxacin: 1.0 \nNo \nZivanovic 2012 \nMGIT 960 and LJ agar proportion \nOfloxacin: 2.0 for both \nmedia \nYes for MGIT; no for LJ \nLJ-Löwenstein-Jensen; MGIT –Mycobacterial growth indicator tube \nReference: WHO, Updated interim critical concentrations for first-line and second-line DST (as of May 2012) \nhttp://www.stoptb.org/wg",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the fluoroquinolone concentration table and begins a table of amikacin, kanamycin and capreomycin concentrations used in included studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-682",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 682",
      "pageStart": 682,
      "pageEnd": 682,
      "target": "../sections/page-682.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Please do not distribute further. \n \n \n677\n \n \nCapreomycin: unknown \nNot applicable \nFan 2011 \nMGIT 960 \nAmikacin: 1.0 \nYes \nFerro 2013 \nMiddlebrook 7H10 (agar \nproportion) \nAmikacin: 5.0 \nNo \nKanamycin: 5.0 \nYes \nFIND 2016 \nMGIT 960 \nAmikacin: 1.0 \nYes \nKanamycin: 2.5 \nYes \nCapreomycin: 2.5 \nYes \nHillemann 2009 \nMGIT 960 and LJ (agar \nproportion) \nAmikacin: 1.0 for MGIT \nand 40.0 for LJ \nYes for MGIT; no for LJ \nCapreomycin: 2.5 for \nMGIT and 40.0 for LJ \nYes \nYes for both types of media \nHuang 2011 \nMiddlebrook 7H11 and MGIT 960",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the table of second-line injectable drug concentrations used in culture-based susceptibility testing, by study and medium, with agreement with WHO recommendations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-683",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 683",
      "pageStart": 683,
      "pageEnd": 683,
      "target": "../sections/page-683.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Please do not distribute further. \n \n \n678\n \n \nMiotto 2012 \nMGIT 960 \nAmikacin: 1.0 \nYes \nKanamycin: 5.0 \nNo \nCapreomycin: 2.5 \nYes \nNICD 2015 \nMGIT 960 \nAmikacin: 1.0 \nYes \nKanamycin: 2.5 \nYes \nCapreomycin: 2.5 \nYes \nSaid 2012 \nMiddlebrook 7H11 \nKanamycin: 5.0 \nNo \nCapreomycin: 10.0 \nNo \nSimons 2015 \nMGIT 960 and Middlebrook 7H10 \n(agar dilution) \nAmikacin: 1.0 for MGIT \nand 5.0 for Middlebrook \n7H10 \nYes for MGIT; no for 7H10 \nCapreomycin: 2.5 for \nMGIT and 10.0 for \nMiddlebrook 7H10 \nYes for MGIT; no for 7H10 \nTagliani 2015 \nMGI",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Further rows of the amikacin, kanamycin and capreomycin concentration table for included studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-684",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 684",
      "pageStart": 684,
      "pageEnd": 684,
      "target": "../sections/page-684.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Please do not distribute further. \n \n \n679\n \n \nCapreomycin: 2.5 for \nMGIT and 40.0 for LJ \nYes for both types of media \nLJ-Löwenstein-Jensen; MGIT –Mycobacterial growth indicator tube \nReference: WHO, Updated interim critical concentrations for first-line and second-line DST (as of May 2012) \nhttp://www.stoptb.org/wg/gli/assets/documents/Updated%20critical%20concentration%20table_1st%20and%202nd%20line%20drugs.pdf",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The end of the injectable drug concentration table, with abbreviations and a reference to WHO's 2012 interim critical concentrations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-685",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 685",
      "pageStart": 685,
      "pageEnd": 685,
      "target": "../sections/page-685.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme. Please do not distribute further. \n \n \n680",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A near-empty page carrying only the distribution notice and a page number.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-686",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 686",
      "pageStart": 686,
      "pageEnd": 686,
      "target": "../sections/page-686.md",
      "narrative": null,
      "content": "Web Annex D.17. High complexity hybridization-based NAATs: \nDiagnostic accuracy for detection of resistance to pyrazinamide. A \nsystematic review \n \n \n \nDr. Michael G Whitfield1,2 and Prof. Dr. Annelies van Rie1 \n \n \n1Family Medicine and Population Health (FAMPOP), Faculty of Medicine and Health Sciences, \nUniversity of Antwerp, Antwerp, Belgium; 2Institute of Health Informatics Research, University \nCollege London, London, United Kingdom. \n \nCorresponding author: \nMichael G Whitfield, PhD \nInstitute of Health Informatics Research \n222 Euston Road \nUniversity College London \nLondon \nNW1 2DA \nU",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title page of Web Annex D.17, a systematic review of the accuracy of high-complexity hybridisation-based NAATs for detecting pyrazinamide resistance, with authors and affiliations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-687",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 687",
      "pageStart": 687,
      "pageEnd": 687,
      "target": "../sections/page-687.md",
      "narrative": null,
      "content": "BACKGROUND \nMycobacterium tuberculosis (MTB) remains in the top 10 of causes for death globally, with an \nestimated 10 million cases and 1.2 million deaths globally in 2019.1 Of the 10 million cases, an \nestimated 465,000 were incident cases of multidrug resistant (MDR)/ rifampicin resistant (RR) \ntuberculosis, a marker for MDR-tuberculosis (TB).1 \nThere have been great improvements in the detection of MTB and resistance to rifampicin (RIF) since \nthe recommendation by WHO of the Xpert MTB/RIF and Xpert ULTRA MTB/RIF assays (Cepheid, \nSunnyvale, USA) for routine diagnosis of MTB.1-5 The other",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives background on the TB and MDR/RR-TB burden in 2019 and on the need to detect pyrazinamide resistance, and explains that the GenoScholar PZA-TB line probe assay lacks specific mutant probes because resistance mutations are spread across the pncA gene.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-688",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 688",
      "pageStart": 688,
      "pageEnd": 688,
      "target": "../sections/page-688.md",
      "narrative": null,
      "content": "GenoScholar PZA-TB assay targets a 700bp fragment that covers the entire pncA gene and promoter \nregion up to nucleotide -18 of the wild type H37Rv reference strain. \nDNA extracted from clinical specimens or cultures is amplified with primers by the polymerase chain \nreaction (PCR). Amplified DNA than hybridize to complementary oligonucleotides probes that are \nbound on a membrane-based strip. After allowing for hybridization of complimentary amplicons, \nalkaline phosphatase-labelled streptavidin is added to bind to any hybrids formed in the previous step. \nThe enzymatic reaction with the subs",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes how the GenoScholar PZA-TB assay amplifies and hybridises the pncA gene region, and sets out the review's eligibility criteria requiring comparison with phenotypic or sequencing-based reference standards.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-689",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 689",
      "pageStart": 689,
      "pageEnd": 689,
      "target": "../sections/page-689.md",
      "narrative": null,
      "content": "684 \n \n \n \n \nfor inclusion is they reported data comparing the index test to an acceptable reference standard (defined \nbelow) from which we could extract true positive (TP), true negative (TN), false negative (FN) and \nfalse positive (FP) values. \nWe only included data on Mycobacterium tuberculosis and excluded any data on nontuberculous \nmycobacteria. We included results obtained from sputum samples or MTB cultures retrieved from \npeople diagnosed pulmonary tuberculosis, independent of age, and HIV status. Samples were included \nirrespective of knowledge of rifampicin resistance (rifampicin",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the inclusion criteria, limited to M. tuberculosis data from sputum or cultures, and describes the literature search and contact with the manufacturer for unpublished data.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-690",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 690",
      "pageStart": 690,
      "pageEnd": 690,
      "target": "../sections/page-690.md",
      "narrative": null,
      "content": "685 \n \n \n \nTwo review authors independently assessed studies for eligibility. We resolved disagreements by \ndiscussion between the two reviewer authors. We illustrate the study selection process in a PRISMA \ndiagram. \n \nData extraction \nTwo review authors independently extracted data from the included studies using a standardized form. \nData was extracted using a standardized form and included first author name; publication year, \npublication title, PZA LPA result, phenotypic PZA DST result, genotypic pncA result, smear status, \nresistance profile, treatment information and any estimates of ac",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes study selection and data extraction by two reviewers, quality assessment and the approach to GRADE certainty, including projected test results at given resistance prevalences.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p691-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 691,
      "pageEnd": 691,
      "target": "../images/img-p691-1.webp",
      "narrative": "draft",
      "content": "PRISMA flow diagram (Figure 1): records from PubMed (n = 17), Web of Science (n = 38) and Embase (n = 72); 31 duplicate articles removed, giving 96 unique articles; 89 excluded after screening title and abstract, leaving 7 screened articles; 2 unpublished trials/studies added, for 9 articles/trials included.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-691",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 691",
      "pageStart": 691,
      "pageEnd": 691,
      "target": "../sections/page-691.md",
      "narrative": null,
      "content": "686 \n \n \n \njudgements on imprecision from these calculations. Prevalence’s were selected based on published \nprevalence’s of PZA resistance, which are mainly determined by the MTB drug resistance profile: 8% \nprevalence PZA-R for any person with TB; 50% for MDR/RR-TB and 90% for extensively drug \nresistant (XDR)-TB.8-10, 22-24 \nPublication bias: we rated publication bias as undetected (not serious) because of \nthe \ncomprehensiveness of the literature search and following extensive outreach to tuberculosis researchers \nto identify studies. \n \nRESULTS \nIdentification of eligible studies \nThe ini",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives the pyrazinamide resistance prevalences used for GRADE judgements and the rating of publication bias, then summarises the search results with a PRISMA diagram and exclusion reasons.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-vfig-p692",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 692,
      "pageEnd": 692,
      "target": "../figures/vfig-p692.png",
      "narrative": "draft",
      "content": "A page render showing the end of a table of excluded article types (total 89), a text section on quality assessment, and Table 1, a QUADAS-2 table for nine studies (Ando, Sekiguchi, Mitarai, Rienthong, Driesen, Willby, Matsumoto, ITM trial and Burhan trial). Each risk-of-bias and applicability cell holds a green plus (low), red minus (high) or yellow question mark (unclear); the Burhan trial is rated low in every domain.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-692",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 692",
      "pageStart": 692,
      "pageEnd": 692,
      "target": "../sections/page-692.md",
      "narrative": null,
      "content": "687 \n \nRISK OF BIAS \nAPPLICABILITY \nCONCERNS \n \n \nReview article \n20 \nGeneral article about TB (DR-TB) \n9 \nArticle studying NTMs \n3 \nCommentary or editorials \n3 \nTotal \n89 \n \nQuality assessment of studies \nThe results of the assessment of methodical quality using the QUADAS-2 tool is shown in Table 1 and \nFigure 2. For the patient selection domain, we judge 44% of studies to have high risk of bias because \nthey selected specific isolates based on PZA resistance instead of a sample representative for the target \npopulation. For the index domain, we judged 44% studies to have a high risk of bias",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents the QUADAS-2 quality assessment of included studies, noting that many had high risk of bias in patient selection because isolates were chosen by resistance status, with a per-study table.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-693",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 693",
      "pageStart": 693,
      "pageEnd": 693,
      "target": "../sections/page-693.md",
      "narrative": null,
      "content": "688\n \n \n \n+ Low \n- High \n? Unclear \n \n \nFigure 2: Overview of QUADAS-2 assessment results. \n \n \nPatient Selection \n \nIndex Test \n \n \nReference Standard \n \n \nFlow & Timing \nLow \nHigh \nUnclear \n \n \n \n \n \n0% \n20% \n40% \n60% \n80% \n100% \n \nProportion of studies with high, low or unclear \nRISK of BIAS \n \n \nPatient Selection \n \n \nIndex Test \n \n \nReference Standard \n \nLow \nHigh \nUnclear \n \n0% \n20% \n40% \n60% \n80% \n100% \n \nProportion of studies with high, low or unclear \nCONCERNS regarding APPLICABILITY \n \nPICO questions \nPICO 1: What is the overall diagnostic accuracy of PZA LPA assay for detection of P",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A chart of QUADAS-2 risk of bias and applicability results, followed by the PICO and GRADE questions on using the PZA line probe assay on sputum.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-694",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 694",
      "pageStart": 694,
      "pageEnd": 694,
      "target": "../sections/page-694.md",
      "narrative": null,
      "content": "No pooled estimates – too few studies to inform on certainty.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A one-line note that no pooled estimates were possible because too few studies were available.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p695-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 695,
      "pageEnd": 695,
      "target": "../images/img-p695-1.webp",
      "narrative": "draft",
      "content": "Forest plot of Nipro PZA LPA sensitivity and specificity for pyrazinamide (PZA) resistance against phenotypic DST, with two studies: Mitarai (sensitivity 1.00 [0.40, 1.00], specificity 1.00) and Sekiguchi (sensitivity not estimable, specificity 1.00). Each row gives a study's TP, FP, FN and TN counts and sensitivity and specificity with 95% CI, plotted as squares with CI lines on 0–1 scales.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p695-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 695,
      "pageEnd": 695,
      "target": "../images/img-p695-2.webp",
      "narrative": "draft",
      "content": "Forest plot of Nipro PZA LPA sensitivity and specificity for PZA resistance against genotypic DST, with one study (Driesen: TP 14, FN 0, TN 7; sensitivity 1.00 [0.77, 1.00], specificity 1.00 [0.59, 1.00]). Each row gives a study's TP, FP, FN and TN counts and sensitivity and specificity with 95% CI, plotted as squares with CI lines on 0–1 scales.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-695",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 695",
      "pageStart": 695,
      "pageEnd": 695,
      "target": "../sections/page-695.md",
      "narrative": null,
      "content": "690 \n \n \n \n \n \nFigure 3: Forest plots of Nipro PZA LPA sensitivity and specificity for detection of PZA resistance \nagainst a reference standard of phenotypic drug susceptibility PZA assay (microbiological reference \nstandard). The squares represent the sensitivity and specificity of one study, the black line its confidence \ninterval. TP: true-positive; FP: false-positive; FN: false-negative; TN: true-negative. \n \n43) Should PZA LPA assay on sputum be used to diagnose PZA resistance in patients with \nmicrobiologically confirmed PTB, irrespective of resistance to RIF, gDST? \nNo studies availabl",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Forest plots of Nipro PZA line probe assay accuracy against phenotypic testing, with further PICO questions on sputum testing against other reference standards.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-696",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 696",
      "pageStart": 696,
      "pageEnd": 696,
      "target": "../sections/page-696.md",
      "narrative": null,
      "content": "691 \n \n \n \nNo studies available to inform on decision. \n \nTable 2: PICO 1 summary: diagnostic accuracy of PZA LPA for detection of PZA resistance in sputum \nfrom pulmonary TB patients with and without RR-TB. \n \n \n \n \n \n \n \n \n \n \n \n \n \n \nAbbreviations: CI: confidence interval; pDST: phenotypic drug susceptibility test; gDST: genomic drug \nsusceptibility test; CRS: composite reference standard; NA: not applicable. \n \nTable 3: GRADE certainty of evidence for sub PICO questions. \n \nPICO \nsub \nquestio \nn \nPopulatio \nn \nReferen \nce \nStandar \nd \nStudies \n(participant \ns) \nPooled \nSensitivit \ny (95%",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "States that no studies were available for some questions and presents summary and GRADE tables for PZA line probe assay accuracy on sputum, mostly with no data.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p697-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 697,
      "pageEnd": 697,
      "target": "../images/img-p697-1.webp",
      "narrative": "draft",
      "content": "Table of risk of bias (patient selection, index test, reference standard, flow and timing) and applicability concerns (patient selection, index test, reference standard) for seven Nipro PZA LPA studies: Ando, Sekiguchi, Mitarai, Rienthong, Driesen, Matsumoto and Burhan trial, using red, yellow and green markers. Several studies are rated high risk for patient selection and the index test. Key: red = high, yellow '?' = unclear, green = low.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p697-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 697,
      "pageEnd": 697,
      "target": "../images/img-p697-2.webp",
      "narrative": "draft",
      "content": "Forest plot of Nipro PZA LPA sensitivity and specificity for PZA resistance in seven studies: Ando, Burhan trial, Driesen, Matsumoto, Mitarai, Rienthong and Sekiguchi. Each row gives a study's TP, FP, FN and TN counts and sensitivity and specificity with 95% CI, plotted as squares with CI lines on 0–1 scales. Sensitivity ranges from 0.36 (Rienthong) and 0.39 (Burhan trial) to 1.00; specificity is 0.93–1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-697",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 697",
      "pageStart": 697,
      "pageEnd": 697,
      "target": "../sections/page-697.md",
      "narrative": null,
      "content": "692 \n \n \n \nAbbreviations: CI: confidence interval; pDST: phenotypic drug susceptibility test; gDST: genomic drug \nsusceptibility test; CRS: composite reference standard; NA: not applicable. \n \n48) Should PZA LPA assay on isolates be used to diagnose PZA resistance in patients with \nmicrobiologically confirmed PTB, irrespective of resistance to RIF, pDST? \nPZA LPA pooled sensitivity (95% CI) was 81.2% (75.4 to 85.8) and specificity (95% CI) was 97.8% \n(96.5 to 98.6), (7 studies, 964 participants; very low certainty evidence for sensitivity and low certainty \nevidence for specificity).",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports PZA line probe assay accuracy on culture isolates, with pooled sensitivity of 81.2% against phenotypic testing and 96.4% sensitivity with 100% specificity against genomic testing.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p698-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 698,
      "pageEnd": 698,
      "target": "../images/img-p698-1.webp",
      "narrative": "draft",
      "content": "A QUADAS-2 style table of risk of bias (patient selection, index test, reference standard, flow and timing) and applicability concerns (patient selection, index test, reference standard) for six studies: Ando, Sekiguchi, Mitarai, Rienthong, Driesen and ITM trial. Each cell holds a coloured marker: green for low, red for high and yellow for unclear concern.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p698-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 698,
      "pageEnd": 698,
      "target": "../images/img-p698-2.webp",
      "narrative": "draft",
      "content": "A forest plot listing six studies (Ando, Driesen, ITM trial, Mitarai, Rienthong, Sekiguchi) with TP, FP, FN and TN counts and sensitivity and specificity with 95% CI, plotted on 0–1 scales. Sensitivity ranges from 0.79 (Ando) to 1.00, and specificity is 1.00 in all studies where it is estimable (not estimable for Ando).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p698-3",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 698,
      "pageEnd": 698,
      "target": "../images/img-p698-3.webp",
      "narrative": "draft",
      "content": "A forest plot listing seven studies (Ando, Driesen, ITM trial, Mitarai, Rienthong, Sekiguchi, Willby) with TP, FP, FN and TN counts and sensitivity and specificity with 95% CI on 0–1 scales. Sensitivity ranges from 0.93 (Willby) to 1.00, and specificity from 0.86 (Driesen) to 1.00.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-698",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 698",
      "pageStart": 698,
      "pageEnd": 698,
      "target": "../sections/page-698.md",
      "narrative": null,
      "content": "693 \n \n \n \n \n \n \n \nFigure 7: Forest plots of Nipro PZA LPA sensitivity and specificity for detection of PZA resistance \nagainst a reference standard of genomic drug susceptibility PZA assay (microbiological reference \nstandard). The squares represent the sensitivity and specificity of one study, the black line its confidence \ninterval. TP: true-positive; FP: false-positive; FN: false-negative; TN: true-negative. \n \n \n \n \nFigure 8. Risk of bias and applicability concerns for gDST as reference standard. \n \n \n50) Should PZA LPA assay on isolates be used to diagnose PZA resistance in patients with",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Forest plots and risk of bias charts for the PZA line probe assay on isolates against genomic and composite reference standards.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p699-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 699,
      "pageEnd": 699,
      "target": "../images/img-p699-1.webp",
      "narrative": "draft",
      "content": "A QUADAS-2 style table of risk of bias (patient selection, index test, reference standard, flow and timing) and applicability concerns (patient selection, index test, reference standard) for seven studies: Ando, Sekiguchi, Mitarai, Rienthong, Driesen, Willby and ITM trial. Cells hold green (low), red (high) or yellow (unclear) markers.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-699",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 699",
      "pageStart": 699,
      "pageEnd": 699,
      "target": "../sections/page-699.md",
      "narrative": null,
      "content": "694 \n \n \n \nrepresent the sensitivity and specificity of one study, the black line its confidence interval. TP: true- \npositive; FP: false-positive; FN: false-negative; TN: true-negative. \n \n \n \n \nFigure 10. Risk of bias and applicability concerns for CRS as reference standard. \n \n \n51) Should PZA LPA assay on isolates be used to diagnose PZA resistance in patients with \nmicrobiologically confirmed PTB, with detected resistance to RIF, pDST? \nNo studies available to inform on decision. \n \n \n52) Should PZA LPA assay on isolates be used to diagnose PZA resistance in patients with \nmicrobiological",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the figure captions and notes that no studies were available for isolates from people with rifampicin resistance.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-700",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 700",
      "pageStart": 700,
      "pageEnd": 700,
      "target": "../sections/page-700.md",
      "narrative": null,
      "content": "695\n \n \n \nTable 4: PICO 2 summary: diagnostic accuracy of PZA LPA for detection of PZA resistance in isolates \nfrom pulmonary TB patients with and without RIF resistant TB. \n \n \n \n \n \n \n \n \n \n \n \nAbbreviations: CI: confidence interval; pDST: phenotypic drug susceptibility test; gDST: genomic drug \nsusceptibility test; CRS: composite reference standard; NA: not applicable. \n \nTable 5: GRADE certainty of evidence for sub PICO questions. \n \nPICO \nsub \nquesti \non \nPopulati \non \nRefere \nnce \nStanda \nrd \nStudies \n(participa \nnts) \nPooled \nSensiti\nv ity \n(95% \nCI) \nPooled \nSpecifi\nc ity \n(95% \nCI) \nC",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summary and GRADE certainty tables for PZA line probe assay accuracy on isolates by reference standard, including composite standard results of 95.7% sensitivity and 98.6% specificity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-701",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 701",
      "pageStart": 701,
      "pageEnd": 701,
      "target": "../sections/page-701.md",
      "narrative": null,
      "content": "696 \n \n \n \nDowngraded one \nlevel for risk of \nbias and one level \nfor \ninconsistency(speci \nficity) \n45 \nWith \ndetected \nrifampici \nn \nresistanc \ne \npDST \n0 (0) \nNA \nNA \nNA \nNA \n46 \nWith \ndetected \nrifampici \nn \nresistanc \ne \ngDST \n0 (0) \nNA \nNA \nNA \nNA \n47 \nWith \ndetected \nrifampici \nn \nresistanc \ne \nCRS \n0 (0) \nNA \nNA \nNA \nNA \nAbbreviations: CI: confidence interval; pDST: phenotypic drug susceptibility test; gDST: genomic drug \nsusceptibility test; CRS: composite reference standard; NA: not applicable.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A fragment of a GRADE table noting downgrading for risk of bias and inconsistency, with empty rows for rifampicin-resistant subgroups.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-702",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 702",
      "pageStart": 702,
      "pageEnd": 702,
      "target": "../sections/page-702.md",
      "narrative": null,
      "content": "697 \n \n \n \nSUMMARY OF MAIN RESULTS \n1) Should PZA LPA assay on sputum be used to diagnose PZA resistance in patients with \nmicrobiologically confirmed PTB, irrespective of resistance to RIF, pDST? \nToo few studies to address this question. \n \n2) Should PZA LPA assay on sputum be used to diagnose PZA resistance in patients with \nmicrobiologically confirmed PTB, irrespective of resistance to RIF, gDST? \nToo few studies to address this question. \n \n3) Should PZA LPA assay on sputum be used to diagnose PZA resistance in patients with \nmicrobiologically confirmed PTB, irrespective of resistance to",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarises the main results question by question, stating that most sputum-based questions could not be answered owing to too few studies, with estimates given only for isolates.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-703",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 703",
      "pageStart": 703,
      "pageEnd": 703,
      "target": "../sections/page-703.md",
      "narrative": null,
      "content": "698 \n \nAUTHORS CONCLUSIONS \nThe Nipro Genoscholar PZA TB assay appears to perform accurately to detect the presence of a wide \nrange of different pncA variants. Majority of the studies conducted the assay using stored culture \nisolates; thus, we are unable to reach any conclusions regarding the use of the assay in clinical \nspecimens. None of the studies stratified the results by resistance status, smear status or treatment \noutcome and so we are unable to reach any conclusions in this regard. \n \nIMPLICATIONS FOR RESEARCH \nThis review suggests that the Nipro Genoscholar PZA TB assay is accurat",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Concludes that the Nipro GenoScholar PZA-TB assay appears accurate in detecting diverse pncA variants but evidence is mostly from stored isolates, and suggests using it as a reflex test after rifampicin resistance is detected.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-704",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 704",
      "pageStart": 704,
      "pageEnd": 704,
      "target": "../sections/page-704.md",
      "narrative": null,
      "content": "699 \n \n \n \nACKNOWLEDGEMENTS \nThe authors would like to thank all members of the international TORCH consortium for their valuable \ncomments. \nWe thank Erlina Burhan from Persahabatan Hospital in Indonesia for sharing data from the ongoing \ntrial. \nWe thank Leen Rigouts and Bouke de Jong from Institute of Tropical Medicine (ITM) in Antwerp, \nBelgium for sharing data from their finalized trial. \n \nCONTRIBUTIONS OF AUTHORS \nBoth MGW and AVR were involved in all processes of this report. \n \n \nDECLARATIONS OF INTEREST \nMGW received funding from the World Health Organization Global Tuberculosis Prog",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Acknowledgements, author contributions and declarations of interest for the pyrazinamide review, noting WHO Global TB Programme funding.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-705",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 705",
      "pageStart": 705,
      "pageEnd": 705,
      "target": "../sections/page-705.md",
      "narrative": null,
      "content": "700 \n \n \n \nREFERENCES \n1. \nWorld Health Organization. Global Tuberculosis Report 2020. \nhttps://apps.who.int/iris/bitstream/handle/10665/336069/9789240013131-eng.pdf?ua=1. \n2. \nWorld Health Organization. Xpert MTB/RIF implementation manual Technical and \noperational ‘how-to’: practical considerations 2014. \nhttp://apps.who.int/iris/bitstream/handle/10665/112469/9789241506700_eng.pdf?sequence=1 \n. \n3. \nWorld Health Organization. Rapid Implementation of the Xpert MTB/RIF diagnostic \ntest Technical and Operational ‘How-to’Practical considerations 2011. \nhttp://apps.who.int/iris/bitstream/handle/1",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The first part of the reference list for the pyrazinamide resistance review, citing WHO TB reports and studies on pncA and pyrazinamide testing.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-706",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 706",
      "pageStart": 706,
      "pageEnd": 706,
      "target": "../sections/page-706.md",
      "narrative": null,
      "content": "701 \n \n \n \n18. \nSchünemann HJ, Mustafa R, Brozek J et al. GRADE Guidelines: 16. GRADE \nevidence to decision frameworks for tests in clinical practice and public health. J Clin \nEpidemiol 2016; 76: 89-98. \n19. \nSchünemann HJ, Oxman AD, Brozek J et al. Grading quality of evidence and strength \nof recommendations for diagnostic tests and strategies. BMJ 2008; 336: 1106-10. \n20. \nSchünemann HJ, Mustafa RA, Brozek J et al. GRADE guidelines: 21 part 1. Study \ndesign, risk of bias, and indirectness in rating the certainty across a body of evidence for test \naccuracy. J Clin Epidemiol 2020; 122: 129-4",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list for the pyrazinamide review, citing GRADE methodology papers, PZA line probe assay studies and the PRISMA statement.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-707",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 707",
      "pageStart": 707,
      "pageEnd": 707,
      "target": "../sections/page-707.md",
      "narrative": null,
      "content": "702 \n \n \n \nAPPENDICES \nAppendix 1. QUADAS-2 tool \nDomain 1: Patient Selection \nRisk of Bias: Could the selection of patients/specimens have introduced bias? \nSignalling question 1: Was a consecutive or random sample of patients or specimens enrolled? \n‘yes’ when the study enrolled a consecutive or random sample; \n‘no’ for all other studies; \n‘unclear’ when the study did not report on patient selection. \nSignalling question 2: Was a case-control design avoided? \n‘yes’ when a prospective or cross-sectional design was used; \n‘no’ when a case-control design was used; \n‘unclear’ for all other study",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "An appendix setting out the QUADAS-2 signalling questions and how they were answered for patient selection and index test domains in the pyrazinamide review.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-708",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 708",
      "pageStart": 708,
      "pageEnd": 708,
      "target": "../sections/page-708.md",
      "narrative": null,
      "content": "703 \n \n \n \nApplicability: Are there concerns that the index test, its conduct, or its interpretation differ from the \nreview question? \nThere are a limited number of articles available for this review question, and the index test comes from \na single manufacturer. If the study used multiple or blinded readers for the interpretation of the PZA \nLPA we deemed the applicability as ‘low concern’. Where a study used a single, unblinded reader for \nthe interpretation of the PZA LPA we deemed the applicability as ‘high concern’ and for all other studies \nwe deemed applicability as ‘unclear’. \n \nDomai",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the QUADAS-2 guidance on applicability, reference standard and flow and timing as applied in the review.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-709",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 709",
      "pageStart": 709,
      "pageEnd": 709,
      "target": "../sections/page-709.md",
      "narrative": null,
      "content": "704 \n \n \n \n‘yes’ when the same reference standard was used; \n‘no’ when the samples were processed in different laboratories; \n‘unclear’ not applicable. \nSignaling question 3: Were all patients included in the analysis? \n‘yes’ when all patients/isolates were included in the analysis; \n‘no’ when not all patients/isolates were included in the analysis; \n‘unclear’ not applicable.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Final QUADAS-2 signalling questions on use of the same reference standard and inclusion of all patients in analysis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-710",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 710",
      "pageStart": 710,
      "pageEnd": 710,
      "target": "../sections/page-710.md",
      "narrative": null,
      "content": "Web Annex B.16. Review of the diagnostic accuracy of \ntargeted next-generation sequencing technologies for \ndetection of drug resistance among people diagnosed \nwith TB \n \n \n \n \n \n \nPhu Phan Trieu1, Eli Harriss2, Timothy M Walker1,3 \n \n1 Oxford University Clinical Research Unit, Ho Chi Minh City, Vietnam \n2 Bodleian Health Care Libraries, University of Oxford, Oxford, UK \n3 Nuffield Department of Medicine, University of Oxford, Oxford, UK \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \nCorrespondence:  \nTimothy M Walker \ntwalker@oucru.org \ntimothy.walker@ndm.ox.ac.uk",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title page of Web Annex B.16, a review of targeted next-generation sequencing for detecting TB drug resistance, with authors and affiliations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-711",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 711",
      "pageStart": 711,
      "pageEnd": 711,
      "target": "../sections/page-711.md",
      "narrative": null,
      "content": "Acknowledgements \n \nWe would like to thanks the following for support and advice in the generation of this \ndocument: Cecily Miller, Nazir Ismail, Mathieu Bastard, Matteo Zignol, Mikashmi Kohli, \nSamuel Schumacher, Tamara Kredo and Karen Steingart.  \nWe would also like to thank all the contributors of data, all of whom we corresponded \nwith regularly to ensure data accuracy. In no particular order: Mohammad Khaja Mafij \nUddin, Faridath Massou, Leen Rigouts, Conor Meehan, Corinne Merle, Stefan \nNiemann, Silke Feuerriegel, Isabelle Bonnet, Tim Rodwell, Rebecca Coleman, Marva \nSeifert, Anita Sure",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Acknowledgements for the targeted sequencing review, thanking advisers and data contributors and listing funders such as WHO, FIND, the Gates Foundation and Wellcome.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-712",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 712",
      "pageStart": 712,
      "pageEnd": 712,
      "target": "../sections/page-712.md",
      "narrative": null,
      "content": "Executive summary \n \n \nMolecular diagnostic tests for drug susceptibility testing (DST) of Mycobacterium tuberculosis strains \nare now widely understood to be the best hope for attaining the W O’s goal of universal access to \ndrug susceptibility testing. WHO recommended rapid diagnostic (WRD) tests detect resistance to \nonly a few select drugs due to the limited number of mutations that can be probed. Some national \nTB programmes have instituted whole-genome sequencing as a routine molecular test, but this \napproach remains culture dependent. Targeted next-generation sequencing (tNGS) technolo",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The opening of the executive summary on targeted next-generation sequencing for TB drug susceptibility testing, explaining that current rapid tests detect resistance to few drugs. It describes the review's objective and the use of individual-level data from several studies. It outlines the modelling approach, including covariables such as Xpert semi-quantitative results and rifampicin resistance.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-713",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 713",
      "pageStart": 713,
      "pageEnd": 713,
      "target": "../sections/page-713.md",
      "narrative": null,
      "content": "resistance result; whether a sample had been sequenced on one or on two tNGS platform (all the \nsamples from FIND were sequenced using both Deeplex® Myc-TB and NanoTB®). Data on HIV test \nresults were available only from FIND, where they were generated as part of the study, and were \nonly included in models addressing the specific sub-analysis looking at diagnostic accuracy in \npeople living with HIV infection. Study sites, including different countries within a single study, were \nincluded as a random effect in the models.  \n \nThere were two PICO (population, intervention, comparator and outc",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the description of methods, including the use of HIV data from FIND and how study sites were handled, and reports rifampicin resistance results. It notes that phenotypic testing misses borderline resistance, so a composite reference standard is preferred for rifampicin.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-714",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 714",
      "pageStart": 714,
      "pageEnd": 714,
      "target": "../sections/page-714.md",
      "narrative": null,
      "content": "Isoniazid – pDST as reference standard.   \nPooled sensitivity and specificity (95% confidence interval) were 96% (93-99) and 97% (95-99) \nrespectively, based on up to 12 studies and up to 1440 samples and a 74% prevalence of isoniazid \nresistance. The indeterminate rate was 15%. Quality of evidence was judged to be moderate for \nboth sensitivity and specificity. \n \nResults indicate that, in theory, of 1000 patients among whom 100 have isoniazid resistance, 96 \nwould be detected (true positive) by tNGS and 4 would not be detected by tNGS (false negative); \nwhereas 873 without isoniazid resistan",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports pooled accuracy of targeted sequencing for isoniazid resistance, at 96% sensitivity and 97% specificity, and for levofloxacin, moxifloxacin and pyrazinamide. Each result gives the number of studies, indeterminate rates, certainty of evidence and expected results per 1000 patients.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-715",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 715",
      "pageStart": 715,
      "pageEnd": 715,
      "target": "../sections/page-715.md",
      "narrative": null,
      "content": "Ethambutol – composite as reference standard.   \nPooled sensitivity and specificity (95% confidence interval) were 96% (94-99) and 99% (98-100) \nrespectively, based on up to 4 studies and up to 432 samples and a 62% prevalence of ethambutol \nresistance. The indeterminate rate was 16%. Quality of evidence was judged to be low for both \nsensitivity and specificity. \n \nResults indicate that, in theory, of 1000 patients among whom 30 have ethambutol resistance, 29 \nwould be detected (true positive) by tNGS and 1 would not be detected by tNGS (false negative); \nwhereas 960 without ethambutol resist",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports pooled targeted sequencing accuracy for ethambutol against composite and phenotypic standards, and repeats results for isoniazid and fluoroquinolones. Each result gives the certainty of evidence and the expected true and false results per 1000 patients.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-716",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 716",
      "pageStart": 716,
      "pageEnd": 716,
      "target": "../sections/page-716.md",
      "narrative": null,
      "content": "would be reported as moxifloxacin resistant by tNGS (false positive). \n \nPyrazinamide – composite as reference standard.   \nPooled sensitivity and specificity (95% confidence interval) were 90% (87-93) and 99% (97-100) \nrespectively, based on 3 studies and up to 346 samples and a 56% prevalence of pyrazinamide \nresistance. The indeterminate rate was 18%. Quality of evidence was judged to be high for both \nsensitivity and specificity. \n \nResults indicate that, in theory, of 1000 patients among whom 500 have pyrazinamide resistance, \n450 would be detected (true positive) by tNGS and 50 would not",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports targeted sequencing accuracy for pyrazinamide, with 90% sensitivity and 99% specificity against a composite standard. It then covers bedaquiline, linezolid and clofazimine against phenotypic testing, giving expected results per 1000 patients.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-717",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 717",
      "pageStart": 717,
      "pageEnd": 717,
      "target": "../sections/page-717.md",
      "narrative": null,
      "content": "Amikacin – pDST as reference standard.   \nPooled sensitivity and specificity (95% confidence interval) were 87% (75-100) and 99% (98-100) \nrespectively, based on up to 8 studies and up to 1003 samples and a 10% prevalence of amikacin \nresistance. The indeterminate rate was 18%. Quality of evidence was judged to be very low for \nsensitivity and moderate for specificity. \n \nResults indicate that, in theory, of 1000 patients among whom 100 have amikacin resistance, 87 \nwould be detected (true positive) by tNGS and 13 would not be detected by tNGS (false negative); \nwhereas 891 without amikacin re",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports targeted sequencing accuracy for amikacin, ethambutol and streptomycin, and summarises that indeterminate rates were mostly around 10-20%. It leads into the authors' conclusions.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-718",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 718",
      "pageStart": 718,
      "pageEnd": 718,
      "target": "../sections/page-718.md",
      "narrative": null,
      "content": "The impact of tNGS platforms will depend on where they are placed. In this review the quality of \nevidence supporting their use as an initial test of drug resistance is lower than that supporting their \nuse as a reflex test for patients already diagnosed with rifampicin resistance. This is largely because \nthe datasets were strongly enriched for rifampicin resistance. Although the prevalence of resistance \nshould not unduly impact sensitivity or specificity, the results for PICO 1 should still be interpreted \nwith caution. Whether tNGS is placed as an initial test of drug resistance will also",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Concludes that the impact of targeted sequencing depends on where it is placed, with stronger evidence for reflex testing after rifampicin resistance is found than as an initial test, and calls for studies on extrapulmonary specimens.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-719",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 719",
      "pageStart": 719,
      "pageEnd": 719,
      "target": "../sections/page-719.md",
      "narrative": null,
      "content": "Background  \n \nRapid, accurate detection of drug resistance for patients with tuberculosis is an important component \nwithin the process of ensuring optimal treatment outcomes both at an individual patient and \npopulation level. While empirical treatment with first line anti-tuberculosis drugs (isoniazid, \nrifampicin, ethambutol and pyrazinamide) would still result in successful outcomes for most patients \nin the world, this approach has become increasingly hazardous in some countries over the past \ndecades.(1) Offering patients treatment regimens containing too few efficacious drugs risks \ntr",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Sets out the background to a WHO-commissioned review of targeted next-generation sequencing (tNGS) for detecting drug resistance in tuberculosis, explaining that empirical treatment risks failure and amplified resistance where resistance is common. It notes that phenotypic drug susceptibility testing remains out of reach for most patients and that molecular tests such as Xpert MTB/RIF, line-probe assays and Xpert XDR cover only some drugs. It argues that new regimens such as BPaLM have outpaced companion diagnostics and that sequencing directly from samples could help protect these drugs.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-720",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 720",
      "pageStart": 720,
      "pageEnd": 720,
      "target": "../sections/page-720.md",
      "narrative": null,
      "content": "Clinical pathway \n \nThe analyses presented here first assess the diagnostic accuracy of tNGS assays for DST for \nisoniazid, rifampicin, ethambutol, pyrazinamide and fluoroquinolones (levofloxacin and moxifloxacin) \namong all patients diagnosed with TB and then assess the diagnostic accuracy for all covered drugs \n(except rifampicin) for patients whose strains have already been identified as resistant to rifampicin \nby Xpert MTB/RIF® or Xpert Ultra®. Further sub-group analyses examine the performance by semi-\nquantitative Xpert MTB/RIF® or Xpert Ultra®  esult (‘ e y lo ’ o  ‘lo ’ co pa e  to ‘",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the clinical pathway assessed, first testing tNGS for first-line drugs and fluoroquinolones in all TB patients and then for other drugs in rifampicin-resistant patients, with sub-analyses by Xpert bacterial load and HIV status. It explains the four possible test outcomes, true and false positives and negatives, and their consequences for patient treatment. It also notes that Deeplex Myc-TB can report uncharacterized mutations, which the review counts with failures as indeterminate results; some quoted terms on the page have missing letters in the extracted text.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-721",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 721",
      "pageStart": 721,
      "pageEnd": 721,
      "target": "../sections/page-721.md",
      "narrative": null,
      "content": "PICOs \n \nGenerically stated, this review explored two PICOs. Each is applied to a specific list of drugs: \n \n1. Should tNGS as the initial test be used to diagnose drug resistance in patients with \nbacteriologically confirmed pulmonary TB disease? \n \nApplies to \n• \nRifampicin, using a composite reference standard of pDST and WGS and MTB/RIF \nXpert® or Xpert Ultra® \n• \nRifampicin, using pDST as the reference standard \n• \nIsoniazid, using pDST as the reference standard \n• \nLevofloxacin, using pDST as the reference standard \n• \nMoxifloxacin, using pDST as the reference standard \n• \nPyrazinamide,",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Sets out the review's two questions: whether tNGS should be the first test for drug resistance in bacteriologically confirmed pulmonary TB, and whether it should be used for patients with rifampicin-resistant TB, listing the drugs and reference standards for each. It states the review objective and notes sub-analyses for people with HIV and low bacterial load as a proxy for children.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-722",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 722",
      "pageStart": 722,
      "pageEnd": 722,
      "target": "../sections/page-722.md",
      "narrative": null,
      "content": "Methods \n \nTypes of studies and participants \n \nData were included from both published and unpublished, prospective, observational clinical studies \nof tNGS platform diagnostic accuracy. We included all studies where tNGS had been performed \ndirectly from clinical samples and excluded studies where tNGS had been performed exclusively on \ncultured isolates. We included all samples as long as data on tNGS result and reference had been \nfinalized and were not subject to change. All studies were required to have comparator pDST data \nas a reference, and for rifampicin, ethambutol and pyrazinamide",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the methods, including prospective clinical studies testing tNGS directly on samples, phenotypic DST as the reference and composite references with whole genome sequencing for rifampicin, ethambutol and pyrazinamide. It lists the drugs assessed and gives the database search strategy, run in 2022 and repeated in 2023, with independent screening by two reviewers.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-723",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 723",
      "pageStart": 723,
      "pageEnd": 723,
      "target": "../sections/page-723.md",
      "narrative": null,
      "content": "they have, or know of, unpublished data that could be contributed. \n \nInclusion criteria were set as follows, and all needed to be met:  \n- \nStudies using ‘ esi n loc e ’ and ‘ a  et  ea y’ tNGS platforms \n- \nStudies assessing samples from patients diagnosed with bacteriologically confirmed \ntuberculosis disease by culture, WHO recommended molecular assay or lateral flow \nlipoarabinomannan (LF-LAM) \n- \nStudies using a gold standard comparator of phenotypic drug susceptibility testing (pDST) \nor a composite of culture based whole genome sequencing genomic DST (gDST) for \nrifampicin, ethambutol",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists inclusion criteria, such as specific tNGS platforms, bacteriologically confirmed TB, an accepted reference standard and available individual patient data, and explains the decision to run only an individual patient data meta-analysis. It then begins describing the QUADAS-2 quality assessment, starting with patient selection.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-724",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 724",
      "pageStart": 724,
      "pageEnd": 724,
      "target": "../sections/page-724.md",
      "narrative": null,
      "content": "A. Risk of bias: could the conduct or interpretation of the index test have introduced bias? \ni. \nSignalling question 1: was the index test result interpreted without knowledge of the \nreference test result? \nWe scored ‘ i    is ’ if the operator of the index test was unblinded to the result of \nthe reference. Although the output of the index test is automated, there is still some \nroom for interpretation by the operator, such as when the call is based on low \nfrequency resistance alleles. We also observed that some operators initially wrongly \ninterpreted the meaning of Deeplex® Myc-T ’s ‘non",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the QUADAS-2 assessment for the index test, reference standard and flow and timing, noting that some operators misread Deeplex Myc-TB's uncharacterized variant calls. It then introduces GRADE grading of the certainty of evidence.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-725",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 725",
      "pageStart": 725,
      "pageEnd": 725,
      "target": "../sections/page-725.md",
      "narrative": null,
      "content": "- \nLow (downgraded 2 levels) \n- \nVery low (downgraded by >2 levels) \nbased on \n- \nRisk of bias \n- \nIndirectness \n- \nInconsistency \n- \nImprecision \n- \nPublication bias \n \nWe did not downgrade the risk of bias where QUADAS-2 scored ‘ i    is ’ for PICO 1 because of \nstudies enriching for rifampicin resistant samples. Instead, we downgraded one for indirectness \nwhere this was the case. We did not downgrade for risk of bias because the prevalence of resistance \nshould not have an undue impact on sensitivity and specificity, unlike for positive and negative \npredictive values. Where a non WHO-reco",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains the GRADE rules applied, such as downgrading for indirectness when studies were enriched for rifampicin resistance and for imprecision when confidence intervals were wide. It describes the statistical analysis, including minimum sample numbers per platform and drug and rules for excluding weaker platforms to support a class-based recommendation.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-726",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 726",
      "pageStart": 726,
      "pageEnd": 726,
      "target": "../sections/page-726.md",
      "narrative": null,
      "content": "setting minimal performance characteristics which individual platforms within class will have to meet \nin the future. \n \nWe performed and individual patient data (IPD) meta-analysis instead of a classical meta-analysis. \nThere were a number of reasons for this: \n- \nThe studies identified in the literature were generally too small to contribute to a classical \nmeta-analysis \n- \nAn IPD meta-analysis allowed for co-variables to be included in the model \n- \nA large proportion of the data came from FIND where the same samples were sequenced \non different platforms. The IPD approach allowed us to co",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains why an individual patient data meta-analysis was used, citing small studies, the ability to include covariates and repeated testing of the same samples on several platforms. It describes building mixed-effects models for sensitivity, specificity and predictive values with covariates such as Xpert rifampicin result, bacterial load and HIV status.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-727",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 727",
      "pageStart": 727,
      "pageEnd": 727,
      "target": "../sections/page-727.md",
      "narrative": null,
      "content": "single result, the FIND data included results on two MTB/RIF Xpert® or Xpert Ultra® results. The first \nwas from the screening test for entry into the study and the second was repeated as part of the \nstudy on the same sample that was also processed for the index test. We used the repeat test results \nfrom FIND to inform the composite reference standard for rifampicin, and the semi-quantitative \nresults. However, we used the result of the screening test for the co-variable in the multi-variable \nmodel. We judged that this result best represented how tNGS might be positioned in the real world",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains that, for FIND data, the repeat Xpert result informed the reference standard while the screening result was used as a model covariate, reflecting the likely real-world use of tNGS as a follow-on test after Xpert.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-728",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 728",
      "pageStart": 728,
      "pageEnd": 728,
      "target": "../sections/page-728.md",
      "narrative": null,
      "content": "Results \n \nResults of the literature search and call for data \n \nWe identified 876 articles from the literature review. Nine full texts were reviewed after 867 articles \nwere excluded on the basis of title alone. Three met inclusion criteria. An additional published article \nwas identified after contacting experts in the field.  \n \nSources of published data: \n \nThe four papers identified from the published literature were: \n \n1. Feuerriegel et. al. “ api  genomic first – and second-line drug resistance prediction from clinical \nMycobacterium tuberculosis specimens using Deeplex-MycT ”  Europea",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports the search results, with 876 articles screened and four published studies included from Germany, India and France, describing their settings and sample processing. It introduces unpublished data, the largest being a FIND head-to-head study of Deeplex Myc-TB and NanoTB in South Africa, India and Georgia.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-729",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 729",
      "pageStart": 729,
      "pageEnd": 729,
      "target": "../sections/page-729.md",
      "narrative": null,
      "content": "frozen until a sufficient number of samples for a tNGS sequencing run had been collected.  \n \nThe “  A nostics for Multidrug Resistant Tuberculosis in A  ica” (DIAMA) study is on-going and set \nin Rwanda, Cameroon, Ethiopia, Guinea, Mali, and Benin. It is has enriched for patients with \nrifampicin resistance, aiming to recruit approximately an equal number of patients with and without \nrifampicin resistance. It is assessing the diagnostic accuracy of Deeplex® Myc-TB.  \n \nThe icddr,b in Bangladesh contributed data from an on-going study of Deeplex® Myc-TB in \nBangladesh. Samples have been colle",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes further unpublished data sources, including the DIAMA study in six African countries, studies in Bangladesh, South Africa and Italy, and notes which sources provided Xpert rifampicin and semi-quantitative results.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p730-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 730,
      "pageEnd": 730,
      "target": "../images/img-p730-1.webp",
      "narrative": "draft",
      "content": "A flow chart of study selection: 1754 records identified through database searching; 878 duplicates removed; 876 records screened by titles, of which 867 were excluded; 9 full-text articles (datasets) assessed for eligibility, of which 6 did not meet inclusion criteria; 3 studies included from the literature review. Together with 1 published study obtained by corresponding with experts and 6 unpublished studies from a WHO call for data, 10 studies were included in total.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-730",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 730",
      "pageStart": 730,
      "pageEnd": 730,
      "target": "../sections/page-730.md",
      "narrative": null,
      "content": "Flow chart 1: Flow of studies in the review \n \n \n \n \n \n \n \n \n \n \nIdentifying which drugs and platforms to include in the class-based assessment \n \nThere were sufficient samples among the pooled data to address each PICO. No drugs were \ntherefore dropped from the analysis.  \n \nFor each drug, between 0 and 2 platforms were dropped from the analysis based on either the \noverall number of resistant or susceptible samples available for that platform and drug, or based on \nperformance characteristics that were insufficiently close to the diagnostic test accuracy of the best \nperforming platform. Tab",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents the flow chart of studies and explains that no drugs were dropped, while up to two platforms per drug were excluded for too few samples or performance well below the best platform, as shown in Table 1.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-731",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 731",
      "pageStart": 731,
      "pageEnd": 731,
      "target": "../sections/page-731.md",
      "narrative": null,
      "content": "Table 1: Results of selection process of platforms for meta-analysis  \n \n \n =” esistant”  y  e e ence stan a  ”; S=”Suscepti le  y  e e ence stan a  ”  Sensiti ity an  speci icity \nindicate a percentage. \nPlatform\nDrug\nR\nS\nSensitivity\nSpecificity\nDeeplex\nIsoniazid\n1056\n498\n97.0\n95.6\nNanoTB\nIsoniazid\n505\n144\n95.4\n98.6\nTBSeq\nIsoniazid\n229\n221\n90.4\n79.6\nDeeplex\nRifampicin\n1083\n548\n99.1\n89.1\nNanoTB\nRifampicin\n497\n152\n97.4\n82.2\nTBSeq\nRifampicin\n285\n165\n94.4\n66.7\nDeeplex\nRifampicin_composite\n915\n290\n95.9\n96.9\nNanoTB\nRifampicin_composite\n534\n115\n95.7\n100.0\nTBSeq\nRifampicin_composite\n78\n0\n100.0\nn/a\nDe",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents Table 1, giving resistant and susceptible sample numbers and raw sensitivity and specificity of the Deeplex, NanoTB and TBSeq platforms for each drug. It shows high sensitivity for most drugs but low values for some platforms with bedaquiline, clofazimine and linezolid.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-vfig-p732",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 732,
      "pageEnd": 732,
      "target": "../figures/vfig-p732.png",
      "narrative": "draft",
      "content": "A colour-coded table of the platforms and drugs included in the IPD meta-analysis, giving sample numbers for Deeplex, NanoTB and TBSeq across 16 drug entries (from isoniazid to clofazimine). Deeplex has data for all drugs except ethambutol (e.g. 1631 for rifampicin), NanoTB for eight entries (617 to 680 samples), and TBSeq only for ethambutol_composite (95) and ethambutol (450).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-732",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 732",
      "pageStart": 732,
      "pageEnd": 732,
      "target": "../sections/page-732.md",
      "narrative": null,
      "content": "After exclusion of the platforms highlighted in red in table 1, data were available for the following \ndrugs from the following platforms (table 2) for inclusion in the IPD meta-analysis. The colours in \ntable 2 are to differentiate the platforms and have no other significance. The number of samples per \nstudy are plotted in figure 1. \n \nTable 2: Platforms and drugs included in the IPD meta-analysis \n \n \n \n \nFigure 1: Sources of data identified \n \n \n \n \n \nDrug\nDeeplex NanoTB TBSeq\nIsoniazid\n1554\n649\nRifampicin_composite\n1205\n649\nRifampicin\n1631\nEthambutol_composite\n672\n95\nEthambutol\n450\nPyrazi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists, in Table 2, the number of samples from each platform included in the meta-analysis for each drug, with a figure on data sources.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-733",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 733",
      "pageStart": 733,
      "pageEnd": 733,
      "target": "../sections/page-733.md",
      "narrative": null,
      "content": "The results of the IPD meta-analysis showed a sensitivity was over 85% for all drugs with the \nexception of Bedaquiline, Linezolid and Clofazimine for which sensitivity was 68%, 69%, and 70%, \nrespectively. Specificity was over 90% for all drugs with the exception of Streptomycin (75%) and \nRIF (81%) where only the pDST was used as the reference standard (table 3). \n \n \nTable 3: Summary of the results from the IPD meta-analysis addressing each PICO \n \n \n \n \n \n \nThe findings of the sub-analyses by HIV test result and semi-quantitative MTB/RIF Xpert are in table \n4. This shows pooled sensitivity",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarises the meta-analysis results, with sensitivity above 85% for all drugs except bedaquiline, linezolid and clofazimine (around 68–70%) and specificity above 90% except for streptomycin and for rifampicin against phenotypic DST alone. Table 3 gives pooled sensitivity, specificity and predictive values for each drug under both review questions.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-734",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 734",
      "pageStart": 734,
      "pageEnd": 734,
      "target": "../sections/page-734.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme.  \n729 \n \nTable 4: Results of sub-analyses on semi-quantitative results and for patients living with HIV \n \n \nDrug\nSub-group\nTP\nFN\nSensitivity\nTN\nFP\nSpecificity\nIndeterminate (%)\nIsoniazid\nSemi-quantitative Xpert result: medium or high\n1239\n47\n96 (93 - 99)\n437\n10\n98 (96 - 99)\n147 (8%)\nIsoniazid\nSemi-quantitative Xpert result: low or very low\n149\n6\n97 (94 - 100)\n66\n5\n93 (87 - 100)\n136 (38%)\nIsoniazid\nHIV test result: Positive\n70\n18\n86 (73 - 98)\n48\n1\n*96 (89 - 100)\n59 (30%)\nIsoniazid\nHIV test result: Negative\n725\n21\n95 (91 - 100)\n208",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents Table 4 with sub-analysis results by Xpert bacterial load and HIV status for isoniazid, rifampicin, ethambutol and pyrazinamide, showing true and false results, sensitivity, specificity and indeterminate rates. Indeterminate results were much more common in low-load and HIV-positive samples.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-735",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 735",
      "pageStart": 735,
      "pageEnd": 735,
      "target": "../sections/page-735.md",
      "narrative": null,
      "content": "This report has been prepared for the WHO Global TB Programme.  \n730 \n \n \nPyrazinamide\nSemi-quantitative Xpert result: medium or high\n337\n55\n86 (80 - 91)\n329\n19\n94 (92 - 97)\n47 (6%)\nPyrazinamide\nSemi-quantitative Xpert result: low or very low\n27\n6\n82 (69 - 95)\n30\n5\n89 (79 - 99)\n71 (51%)\nPyrazinamide\nHIV test result: Positive\n21\n3\n89 (77 - 100)\n47\n2\n97 (93 - 100)\n25 (26%)\nPyrazinamide\nHIV test result: Negative\n224\n26\n89 (86 - 93)\n196\n13\n93 (90 - 97)\n75 (14%)\nLevofloxacin\nSemi-quantitative Xpert result: medium or high\n573\n25\n94 (88 - 100)\n775\n27\n96 (94 - 99)\n41 (3%)\nLevofloxacin\nSemi-quantitativ",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues Table 4 for pyrazinamide, levofloxacin, moxifloxacin, amikacin, streptomycin and the newer drugs, showing accuracy by bacterial load and HIV status and higher indeterminate rates in low-load samples. Footnotes flag where models could not control for certain covariates. For linezolid, very few resistant samples were available in the sub-groups.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-736",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 736",
      "pageStart": 736,
      "pageEnd": 736,
      "target": "../sections/page-736.md",
      "narrative": null,
      "content": "Detailed results addressing each PICO in turn are reported below. For each PICO, the \nnumber of TP, FP, TN and FN samples are shown that were included in the model, \nseparated by study. Accompanying forest plots are depicted. For each PICO there are also \nbar charts showing the proportion of studies, and proportion of samples across all studies, \nt at  e e assesse  to  e at ‘lo   is ’  ‘unclea   is ’  o  ‘ i    is ’ o   ias  o  any o  t e \nQUADAS2 domains. Similarly, a second histogram is included showing the proportion of \nstu ies assesse  to  e ‘lo   is ’  ‘unclea   is ’  o  ‘ i    is ’  o",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains how the detailed results for each question are presented and reports that, for rifampicin against a composite reference, nine studies gave pooled sensitivity of 93% and specificity of 96%. Evidence was rated moderate for sensitivity and low for specificity after downgrading for indirectness and imprecision.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-737",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 737",
      "pageStart": 737,
      "pageEnd": 737,
      "target": "../sections/page-737.md",
      "narrative": null,
      "content": "QUADAS-2 assessment: \n \n \nGRADE assessment: \n \n \n \nPICO 1.2: Should tNGS as the initial test be used to diagnose drug resistance \nto rifampin (RIF) (pDST) in patients with bacteriologically confirmed \npulmonary TB disease? \n \nThirteen studies with 961 samples were included in the model for sensitivity, and 12 studies \nwith 425 samples for specificity. Prevalence of resistance to rifampicin (pDST) among \nsamples included in the model was 69% (95% CI 67-72%). The proportion of samples with \nno t GS  esult  epo te  (t e ‘in ete  inate  ate’)  as 9 9% (95% C  8 4-11.6).",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces the analysis of tNGS for rifampicin resistance against phenotypic DST as the reference, with sample numbers, resistance prevalence and an indeterminate rate of about 10%.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-738",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 738",
      "pageStart": 738,
      "pageEnd": 738,
      "target": "../sections/page-738.md",
      "narrative": null,
      "content": "Results from multivariable, mixed-effects model: \n \nPooled sensitivity was 99% (95% confidence interval (CI) 97%-100%) \n \nPooled specificity was 81% (95% confidence interval (CI) 69%-93%) \n \nNegative predictive value was 98% (95% confidence interval (CI) 96%-99%) \n \nPositive predictive value was 94% (95% confidence interval (CI) 92%-96%) \n \n \nGRADE assessment: \nWe downgraded one for indirectness for both sensitivity and specificity as all studies were \nenriched for rifampicin resistance, leading to applicability concerns, but not for risk of bias \nas sensitivity and specificity should not be u",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports pooled sensitivity of 99% and specificity of 81% for rifampicin against phenotypic DST, with certainty rated moderate for sensitivity and very low for specificity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-739",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 739",
      "pageStart": 739,
      "pageEnd": 739,
      "target": "../sections/page-739.md",
      "narrative": null,
      "content": "GRADE table: \n \n \n \n \nPICO 1.3: Should tNGS as the initial test be used to diagnose drug resistance \nto isoniazid (INH) (pDST) in patients with bacteriologically confirmed \npulmonary TB disease? \n \nTwelve studies with 1440 samples were included in the model for sensitivity, and 12 studies \nwith 517 samples for specificity. Prevalence of resistance to isoniazid among samples \nincluded in the model was 74% (95% CI 72-76%). The proportion of samples with no tNGS \n esult  epo te  (t e ‘in ete  inate  ate’)  as 14 6% (95% C  13 0-16.2).  \n \n \nResults from multivariable, mixed-effects model: \n \nPool",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that tNGS as the initial test for isoniazid resistance had pooled sensitivity of 96% and specificity of 97%, based on twelve studies. Certainty was rated moderate for both after downgrading for indirectness.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-740",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 740",
      "pageStart": 740,
      "pageEnd": 740,
      "target": "../sections/page-740.md",
      "narrative": null,
      "content": "QUADAS-2 assessment: \n \n \n \nGRADE table: \n \n \n \n \nPICO 1.4: Should tNGS as the initial test be used to diagnose drug resistance \nto levofloxacin (LFX) (pDST) in patients with bacteriologically confirmed \npulmonary TB disease? \n \nSix studies with 654 samples were included in the model for sensitivity, and 7 studies with \n913 samples for specificity. Prevalence of resistance to levofloxacin among samples \nincluded in the model was 42% (95% CI 39-44%). The proportion of samples with no tNGS \n esult  epo te  (t e ‘in ete  inate  ate’)  as 9 2% (95% C  7 8-10.7).",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces the analysis of tNGS for levofloxacin resistance as the initial test, with sample numbers, a 42% resistance prevalence and an indeterminate rate of about 9%.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-741",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 741",
      "pageStart": 741,
      "pageEnd": 741,
      "target": "../sections/page-741.md",
      "narrative": null,
      "content": "Results from multivariable, mixed-effects model: \n \nPooled sensitivity was 94% (95% confidence interval (CI) 88%-100%) \n \nPooled specificity was 96% (95% confidence interval (CI) 93%-99%) \n \nNegative predictive value was 97% (95% confidence interval (CI) 96%-98%) \n \nPositive predictive value was 93% (95% confidence interval (CI) 88%-98%) \n \n \nGRADE assessment: \nWe downgraded one for indirectness for both sensitivity and specificity as all studies were \nenriched for rifampicin resistance, leading to applicability concerns, but not for risk of bias \nas sensitivity and specificity should not be u",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports pooled sensitivity of 94% and specificity of 96% for levofloxacin, with certainty rated low for sensitivity and moderate for specificity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-742",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 742",
      "pageStart": 742,
      "pageEnd": 742,
      "target": "../sections/page-742.md",
      "narrative": null,
      "content": "PICO 1.5: Should tNGS as the initial test be used to diagnose drug resistance \nto moxifloxacin (MFX) (pDST) in patients with bacteriologically confirmed \npulmonary TB disease? \n \nSix studies with 652 samples were included in the model for sensitivity, and 8 studies with \n921 samples for specificity. Prevalence of resistance to moxifloxacin among samples \nincluded in the model was 41% (95% CI 39-44%). The proportion of samples with no tNGS \n esult  epo te  (t e ‘in ete  inate  ate’)  as 9 3% (95% C  7 9-10.9).  \n \n \nResults from multivariable, mixed-effects model: \n \nPooled sensitivity was 96%",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that tNGS as the initial test for moxifloxacin resistance had pooled sensitivity of 96% and specificity of 96%, with certainty rated moderate for both.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-743",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 743",
      "pageStart": 743,
      "pageEnd": 743,
      "target": "../sections/page-743.md",
      "narrative": null,
      "content": "QUADAS-2 assessment: \n \n \nGRADE table: \n \n \n \nPICO 1.6: Should tNGS as the initial test be used to diagnose drug resistance \nto pyrazinamide (PZA) (composite) in patients with bacteriologically \nconfirmed pulmonary TB disease? \n \nThree studies with 346 samples were included in the model for sensitivity, and 3 studies with \n269 samples for specificity. Prevalence of resistance to pyrazinamide (composite) among \nsamples included in the model was 56% (95% CI 52-60%). The proportion of samples with \nno t GS  esult  epo te  (t e ‘in ete  inate  ate’)  as 17 6% (95% C  14 6-20.8).  \n \n \nResults from",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces the analysis of tNGS for pyrazinamide resistance against a composite reference, reporting pooled sensitivity of 88%.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-744",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 744",
      "pageStart": 744,
      "pageEnd": 744,
      "target": "../sections/page-744.md",
      "narrative": null,
      "content": "Pooled specificity was 99% (95% confidence interval (CI) 97%-100%) \n \nNegative predictive value was 88% (95% confidence interval (CI) 81%-95%) \n \nPositive predictive value was 99% (95% confidence interval (CI) 98%-100%) \n \n \nGRADE assessment: \nWe downgraded one for indirectness for both sensitivity and specificity as all studies were \nenriched for rifampicin resistance, leading to applicability concerns, but not for risk of bias \nas sensitivity and specificity should not be unduly impacted by a change in the prevalence \nof resistance. Quality of evidence was thereby assessed to be moderate for",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports pooled specificity of 99% for pyrazinamide against the composite reference, with certainty rated moderate for both sensitivity and specificity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-745",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 745",
      "pageStart": 745,
      "pageEnd": 745,
      "target": "../sections/page-745.md",
      "narrative": null,
      "content": "PICO 1.7: Should tNGS as the initial test be used to diagnose drug resistance \nto pyrazinamide (PZA) (pDST) in patients with bacteriologically confirmed \npulmonary TB disease? \n \nSix studies with 425 samples were included in the model for sensitivity, and 6 studies with \n379 samples for specificity. Prevalence of resistance to pyrazinamide (pDST) among \nsamples included in the model was 52% (95% CI 49-56%). The proportion of samples with \nno t GS  esult  epo te  (t e ‘in ete  inate  ate’)  as 14 7% (95% C  12 3-17.3).  \n \n \nResults from multivariable, mixed-effects model: \n \nPooled sensitivity",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that tNGS for pyrazinamide resistance against phenotypic DST had pooled sensitivity of 85% and specificity of 94%. Certainty was rated low for sensitivity, because a Bangladesh study was an outlier, and moderate for specificity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-746",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 746",
      "pageStart": 746,
      "pageEnd": 746,
      "target": "../sections/page-746.md",
      "narrative": null,
      "content": "GRADE table: \n \n \n \nPICO 1.8: Should tNGS as the initial test be used to diagnose drug resistance \nto ethambutol (EMB) (composite) in patients with bacteriologically \nconfirmed pulmonary TB disease? \n \nFour studies with 432 samples were included in the model for sensitivity, and 4 studies with \n268 samples for specificity. Prevalence of resistance to ethambutol (composite) among \nsamples included in the model was 62% (95% CI 58-65%). The proportion of samples with \nno t GS  esult  epo te  (t e ‘in ete  inate  ate’)  as 16 3% (95% C  13 5-19.2).  \n \n \nResults from multivariable, mixed-effects m",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that tNGS for ethambutol resistance against a composite reference had pooled sensitivity of 96% and very high specificity and positive predictive value. It notes the model could not control for rifampicin resistance because of collinearity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-747",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 747",
      "pageStart": 747,
      "pageEnd": 747,
      "target": "../sections/page-747.md",
      "narrative": null,
      "content": "The model for specificity did not control for rifampicin resistance as this variable was \ncollinear in the original model. The quality assessment of the evidence was not downgraded \nas a result. \n \nWe downgraded one for risk of bias for both sensitivity and specificity as difference samples \nwere used for the index (tNGS) and  reference tests. We also downgraded one for \nindirectness for both sensitivity and specificity as all studies were enriched for rifampicin \nresistance, leading to applicability concerns, but not for risk of bias as sensitivity and \nspecificity should not be unduly impact",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains the GRADE judgement for ethambutol against the composite reference, rated low for both sensitivity and specificity because different samples were used for index and reference tests and studies were enriched for rifampicin resistance.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-748",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 748",
      "pageStart": 748,
      "pageEnd": 748,
      "target": "../sections/page-748.md",
      "narrative": null,
      "content": "PICO 1.9: Should tNGS as the initial test be used to diagnose drug resistance \nto ethambutol (EMB) (pDST) in patients with bacteriologically confirmed \npulmonary TB disease? \n \nOne study with 100 samples was included in the model for sensitivity, and one study with \n334 samples for specificity. Prevalence of resistance to ethambutol (pDST) among samples \nincluded in the model was 23% (95% CI 19-27%). The proportion of samples with no tNGS \n esult  epo te  (t e ‘in ete  inate  ate’)  as 0% (95% C  0-1).  \n \n \nResults from multivariable, mixed-effects model: \n \nPooled sensitivity was 88% (95% co",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that tNGS for ethambutol resistance against phenotypic DST, based on a single study, had pooled sensitivity of 88%. Certainty was rated very low for both sensitivity and specificity because of risk of bias, enrichment and reliance on one study.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-749",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 749",
      "pageStart": 749,
      "pageEnd": 749,
      "target": "../sections/page-749.md",
      "narrative": null,
      "content": "Data by study: \n \n \n \nQUADAS-2 assessment: \n \n \n \nGRADE table: \n \n \n \nPICO 2 - TNGS as test for drug resistance in patients with bac-\nconfirmed rif-resistant TB (RR-TB) for INH, LFX, MFX, PZA, BDQ, \nLZD, CFZ, AMK, EMB, ETO, PTO, STR \n \nPICO 2.1: Should tNGS be used to diagnose drug resistance to isoniazid \n(INH) (pDST) in patients with bacteriologically confirmed rifampicin-\nresistant pulmonary TB disease? \n \nTwelve studies with 1440 samples were included in the model for sensitivity, and 12 studies \nwith 517 samples for specificity. Prevalence of resistance to isoniazid among samples",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces the second review question on tNGS for drug resistance in rifampicin-resistant TB, beginning with isoniazid in twelve studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-750",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 750",
      "pageStart": 750,
      "pageEnd": 750,
      "target": "../sections/page-750.md",
      "narrative": null,
      "content": "included in the model was 74% (95% CI 72-76%). The proportion of samples with no tNGS \n esult  epo te  (t e ‘in ete  inate  ate’)  as 14 6% (95% C  13 0-16.2).  \n \nResults from multivariable, mixed-effects model: \n \nPooled sensitivity was 96% (95% confidence interval (CI) 94%-99%) \n \nPooled specificity was 96% (95% confidence interval (CI) 92%-100%) \n \nNegative predictive value was 76% (95% confidence interval (CI) 67%-84%) \n \nPositive predictive value was 100% (95% confidence interval (CI) 99%-100%) \n \n \nGRADE assessment: \nWe did not downgrade for any categories. Quality of evidence was there",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports pooled sensitivity and specificity of 96% for isoniazid in rifampicin-resistant TB, with high certainty for both.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-751",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 751",
      "pageStart": 751,
      "pageEnd": 751,
      "target": "../sections/page-751.md",
      "narrative": null,
      "content": "PICO 2.2: Should tNGS be used to diagnose drug resistance to levofloxacin \n(LFX) (pDST) in patients with bacteriologically confirmed rifampicin-\nresistant pulmonary TB disease? \n \nSix studies with 654 samples were included in the model for sensitivity, and 7 studies with \n913 samples for specificity. Prevalence of resistance to levofloxacin among samples \nincluded in the model was 42% (95% CI 39-44%). The proportion of samples with no tNGS \n esult  epo te  (t e ‘in ete  inate  ate’)  as 9 2% (95% C  7 8-10.7).  \n \n \nResults from multivariable, mixed-effects model: \n \nPooled sensitivity was 96%",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that tNGS for levofloxacin resistance in rifampicin-resistant TB had high pooled sensitivity and specificity. Certainty was rated moderate for sensitivity, because one FIND study was an outlier, and high for specificity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-752",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 752",
      "pageStart": 752,
      "pageEnd": 752,
      "target": "../sections/page-752.md",
      "narrative": null,
      "content": "GRADE table: \n \n \n \n \nPICO 2.3: Should tNGS be used to diagnose drug resistance to moxifloxacin \n(MFX) (pDST) in patients with bacteriologically confirmed rifampicin-\nresistant pulmonary TB disease? \n \nSix studies with 652 samples were included in the model for sensitivity, and 8 studies with \n921 samples for specificity. Prevalence of resistance to moxifloxacin among samples \nincluded in the model was 41% (95% CI 39-44%). The proportion of samples with no tNGS \n esult  epo te  (t e ‘in ete  inate  ate’)  as 9 3% (95% C  7 9-10.9).",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces the analysis of tNGS for moxifloxacin resistance in rifampicin-resistant TB, with sample numbers, resistance prevalence and indeterminate rate.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-753",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 753",
      "pageStart": 753,
      "pageEnd": 753,
      "target": "../sections/page-753.md",
      "narrative": null,
      "content": "Results from multivariable, mixed-effects model: \n \nPooled sensitivity was 97% (95% confidence interval (CI) 94%-100%) \n \nPooled specificity was 95% (95% confidence interval (CI) 91%-99%) \n \nNegative predictive value was 97% (95% confidence interval (CI) 96%-99%) \n \nPositive predictive value was 95% (95% confidence interval (CI) 91%-99%) \n \n \nGRADE assessment: \nWe did not downgrade for any categories. Quality of evidence was thereby assessed to be \nhigh for both sensitivity and specificity. \n \n \nData by study: \n \n \n \n \n \nQUADAS-2 assessment: \n \n \n \nGRADE table:",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports pooled sensitivity of 97% and specificity of 95% for moxifloxacin in rifampicin-resistant TB, with high certainty for both.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-754",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 754",
      "pageStart": 754,
      "pageEnd": 754,
      "target": "../sections/page-754.md",
      "narrative": null,
      "content": "PICO 2.4: Should tNGS be used to diagnose drug resistance to pyrazinamide \n(PZA) (composite) in patients with bacteriologically confirmed rifampicin-\nresistant pulmonary TB disease? \n \nThree studies with 346 samples were included in the model for sensitivity, and 3 studies with \n269 samples for specificity. Prevalence of resistance to pyrazinamide (composite) among \nsamples included in the model was 56% (95% CI 52-60%). The proportion of samples with \nno t GS  esult  epo te  (t e ‘in ete  inate  ate’)  as 17 6% (95% C  14 6-20.8).  \n \n \nResults from multivariable, mixed-effects model: \n \nPoole",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that tNGS for pyrazinamide resistance against a composite reference in rifampicin-resistant TB had high accuracy, with specificity of 99% and high certainty for both measures.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-755",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 755",
      "pageStart": 755,
      "pageEnd": 755,
      "target": "../sections/page-755.md",
      "narrative": null,
      "content": "GRADE table: \n \n \n \n \nPICO 2.5: Should tNGS be used to diagnose drug resistance to pyrazinamide \n(PZA) (pDST) in patients with bacteriologically confirmed rifampicin-\nresistant pulmonary TB disease? \n \nSix studies with 425 samples were included in the model for sensitivity, and 6 studies with \n379 samples for specificity. Prevalence of resistance to pyrazinamide (pDST) among \nsamples included in the model was 53% (95% CI 49-56%). The proportion of samples with \nno t GS  esult  epo te  (t e ‘in ete  inate  ate’)  as 14 7% (95% C  12 3-17.3).  \n \n \nResults from multivariable, mixed-effects model",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that tNGS for pyrazinamide resistance against phenotypic DST in rifampicin-resistant TB had pooled sensitivity and specificity of 90% each.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-756",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 756",
      "pageStart": 756,
      "pageEnd": 756,
      "target": "../sections/page-756.md",
      "narrative": null,
      "content": "We downgraded on for inconsistency for sensitivity as the study from Bangladesh was an \noutlier (see forest plot below). Quality of evidence was thereby assessed to be moderate for \nsensitivity and high for specificity. \n \nData by study: \n \n \n \n \nQUADAS-2 assessment: \n \n \n \n \nGRADE table: \n \n \n \nPICO 2.6: Should tNGS be used to diagnose drug resistance to bedaquiline \n(BDQ) (pDST) in patients with bacteriologically confirmed rifampicin-\nresistant pulmonary TB disease?",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Rates the pyrazinamide evidence moderate for sensitivity and high for specificity, and introduces the bedaquiline analysis for rifampicin-resistant TB.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-757",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 757",
      "pageStart": 757,
      "pageEnd": 757,
      "target": "../sections/page-757.md",
      "narrative": null,
      "content": "Three studies with 31 samples were included in the model for sensitivity, and 4 studies with \n519 samples for specificity. Prevalence of resistance to bedaquiline among samples \nincluded in the model was 6% (95% CI 4-8%). The proportion of samples with no tNGS \n esult  epo te  (t e ‘in ete  inate  ate’)  as 16 7% (95% C  13 7-20.1).  \n \n \nResults from multivariable, mixed-effects model: \n \nPooled sensitivity was 68% (95% confidence interval (CI) 43%-93%) \nModel not controlled for semi-quantitative result from MTB/RIF Xpert as this variable was collinear. \n \nPooled specificity was 97% (95% conf",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that tNGS for bedaquiline resistance, based on only 31 resistant samples, had pooled sensitivity of 68% and specificity of 97%. Certainty was rated low for sensitivity, because of few resistant samples and very wide confidence intervals, and high for specificity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-758",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 758",
      "pageStart": 758,
      "pageEnd": 758,
      "target": "../sections/page-758.md",
      "narrative": null,
      "content": "GRADE table: \n \n \n \nPICO 2.7: Should tNGS be used to diagnose drug resistance to linezolid (LZD) \n(pDST) in patients with bacteriologically confirmed rifampicin-resistant \npulmonary TB disease? \n \nFour studies with 31 samples were included in the model for sensitivity, and 6 studies with \n1093 samples for specificity. Prevalence of resistance to linezolid among samples included \nin the model was 3% (95% CI 2-4%). The proportion of samples with no tNGS result reported \n(t e ‘in ete  inate  ate’)  as 15 1% (95% C  13 1-17.3).  \n \n \nResults from multivariable, mixed-effects model: \n \nPooled sensi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that tNGS for linezolid resistance, with only 31 resistant samples and 3% resistance prevalence, had pooled sensitivity of 69% and specificity of 100%. It notes model restrictions due to collinear variables.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-759",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 759",
      "pageStart": 759,
      "pageEnd": 759,
      "target": "../sections/page-759.md",
      "narrative": null,
      "content": "Positive predictive value was 93% (95% confidence interval (CI) 84%-100%) \nModel not controlled for semi-quantitative result from MTB/RIF Xpert as this variable was collinear, and only includes samples \nthat were resistant to rifampicin by MTB/RIF Xpert as this variable was also collinear in the original model. \n \n \nGRADE assessment \nThe model for specificity was restricted to isolates that were resistant to rifampicin by \nMTB/RIF Xpert and was not controlled for semi-quantitative Xpert result as both results \n(semi-quantitative and rifampicin resistance were collinear in the original model).",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives the positive predictive value and GRADE judgement for linezolid, rated low for sensitivity because of imprecision and high for specificity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-760",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 760",
      "pageStart": 760,
      "pageEnd": 760,
      "target": "../sections/page-760.md",
      "narrative": null,
      "content": "PICO 2.8: Should tNGS be used to diagnose drug resistance to clofazimine \n(CFZ) (pDST) in patients with bacteriologically confirmed rifampicin-\nresistant pulmonary TB disease? \n \nFour studies with 36 samples were included in the model for sensitivity, and 6 studies with \n789 samples for specificity. Prevalence of resistance to clofazimine among samples \nincluded in the model was 3% (95% CI 2-4%). The proportion of samples with no tNGS \n esult  epo te  (t e ‘in ete  inate  ate’)  as 11 6% (95% C  9 5-14.0).  \n \n \nResults from multivariable, mixed-effects model: \n \nPooled sensitivity was 70% (95",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that tNGS for clofazimine resistance had pooled sensitivity of 70% and specificity of 96%, based on very few resistant samples. Certainty was rated low for sensitivity because of outlying studies and wide confidence intervals, and high for specificity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-761",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 761",
      "pageStart": 761,
      "pageEnd": 761,
      "target": "../sections/page-761.md",
      "narrative": null,
      "content": "QUADAS-2 assessment: \n \n \n \nGrade table: \n \n \n \n \nPICO 2.9: Should tNGS be used to diagnose drug resistance to amikacin \n(AMK) (pDST) in patients with bacteriologically confirmed rifampicin-\nresistant pulmonary TB disease? \n \nFive studies with 115 samples were included in the model for sensitivity, and 8 studies with \n1003 samples for specificity. Prevalence of resistance to amikacin among samples included \nin the model was 10% (95% CI 9-12%). The proportion of samples with no tNGS result \n epo te  (t e ‘in ete  inate  ate’)  as 17 8% (95% C  15 6-20.2).  \n \n \nResults from multivariable, mixed",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces the analysis of tNGS for amikacin resistance in rifampicin-resistant TB, with sample numbers, a 10% resistance prevalence and an indeterminate rate of about 18%.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-762",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 762",
      "pageStart": 762,
      "pageEnd": 762,
      "target": "../sections/page-762.md",
      "narrative": null,
      "content": "Pooled sensitivity was 87% (95% confidence interval (CI) 75%-100%) \nModel restricted to samples that tested resistant to rifampicin by MTB/RIF Xpert as this variable was collinear in the original \nmodel. \n \nPooled specificity was 99% (95% confidence interval (CI) 98%-100%) \nModel restricted to samples that tested resistant to rifampicin by MTB/RIF Xpert as this variable was collinear in the original \nmodel. \n \nNegative predictive value was 98% (95% confidence interval (CI) 96%-100%) \nModel not controlled for rifampicin resistance as determined by MTB/RIF Xpert as this variable was collinear. M",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports pooled sensitivity of 87% and specificity of 99% for amikacin. Certainty was rated very low for sensitivity and moderate for specificity after downgrading for bias, inconsistency and imprecision.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-763",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 763",
      "pageStart": 763,
      "pageEnd": 763,
      "target": "../sections/page-763.md",
      "narrative": null,
      "content": "GRADE table: \n \n \n \n \nPICO 2.10: Should tNGS be used to diagnose drug resistance to ethambutol \n(EMB) (composite) in patients with bacteriologically confirmed rifampicin-\nresistant pulmonary TB disease? \n \nFour studies with 431 samples were included in the model for sensitivity, and 4 studies with \n123 samples for specificity. Prevalence of resistance to ethambutol (composite) among \nsamples included in the model was 78% (95% CI 74-81%). The proportion of samples with \nno t GS  esult  epo te  (t e ‘in ete  inate  ate’)  as 20 6% (95% C  17 3-24.2).  \n \n \nResults from multivariable, mixed-effec",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces the analysis of tNGS for ethambutol resistance against a composite reference in rifampicin-resistant TB and reports specificity of 98% and high predictive values.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-764",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 764",
      "pageStart": 764,
      "pageEnd": 764,
      "target": "../sections/page-764.md",
      "narrative": null,
      "content": "Model not controlled for rifampicin resistance as determined by MTB/RIF Xpert as this variable was collinear. Model therefore \nrun only on isolates that tested resistant to rifampicin by Xpert. \n \nGRADE assessment: \nThe model for specificity was restricted to samples that were resistant to rifampicin by \nMTB/RIF Xpert as this variable was collinear in the original model. The model does control \nfor semi-quantitative Xpert result. The quality assessment of the evidence was not \ndowngraded as a result. \n \nWe downgraded one for bias for both sensitivity and specificity as different samples were",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives the GRADE judgement for ethambutol against the composite reference in rifampicin-resistant TB, rated moderate for both measures because different samples were used for index and reference tests.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-765",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 765",
      "pageStart": 765,
      "pageEnd": 765,
      "target": "../sections/page-765.md",
      "narrative": null,
      "content": "PICO 2.11: Should tNGS be used to diagnose drug resistance to ethambutol \n(EMB) (pDST) in patients with bacteriologically confirmed rifampicin-\nresistant pulmonary TB disease? \n \nOne study with 89 samples was included in the model for sensitivity, and 1 study with 213 \nsamples for specificity. Prevalence of resistance to ethambutol (pDST) among samples \nincluded in the model was 29% (95% CI 24-35%). The proportion of samples with no tNGS \n esult  epo te  (t e ‘in ete  inate  ate’)  as 0% (95% C  0-1). \n \n \nResults from multivariable, mixed-effects model: \n \nPooled sensitivity was 91% (95% conf",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that tNGS for ethambutol resistance against phenotypic DST in rifampicin-resistant TB, from a single study, had pooled sensitivity of 91% and specificity of 92%. Certainty was rated low for both because of risk of bias and data from only one country.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-766",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 766",
      "pageStart": 766,
      "pageEnd": 766,
      "target": "../sections/page-766.md",
      "narrative": null,
      "content": "GRADE table: \n \n \n \nPICO 2.12: Should tNGS be used to diagnose drug resistance to \nstreptomycin (STR) (pDST) in patients with bacteriologically confirmed \nrifampicin-resistant pulmonary TB disease? \n \nFive studies with 493 samples were included in the model for sensitivity, and 5 studies with \n250 samples for specificity. Prevalence of resistance to streptomycin among samples \nincluded in the model was 66% (95% CI 63-70%). The proportion of samples with no tNGS \n esult  epo te  (t e ‘in ete  inate  ate’)  as 18 8% (95% C  16 1-21.8). \n \nResults from multivariable, mixed-effects model: \nPooled",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that tNGS for streptomycin resistance in rifampicin-resistant TB had pooled sensitivity of 98% and specificity of 75%. Specificity was downgraded for an outlying study and wide confidence intervals.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-767",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 767",
      "pageStart": 767,
      "pageEnd": 767,
      "target": "../sections/page-767.md",
      "narrative": null,
      "content": "was thereby assessed to be high for sensitivity but low for specificity. \n \nData by study: \n \n \n \nQUADAS-2 assessment: \n \n \nGRADE table: \n \n \n \nAuthor’s conclusions \n \nRegarding positioning tNGS as the first test for drug resistance for patients (PICO 1):  \n \nPooled sensitivity to first line drugs was 88%-96%, with the highest estimate for isoniazid \nand the lowest for pyrazinamide. This is similar to the results obtained from WGS \nsequencing from culture,(10) a routine diagnostic approach already adopted by a number of \ncountries. Sensitivity for the flouroquinolones was 94%-96%, which is com",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Begins the authors' conclusions, noting that, as an initial test, tNGS sensitivity for first-line drugs was 88–96% and for fluoroquinolones 94–96%, comparable to whole genome sequencing from culture.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-768",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 768",
      "pageStart": 768,
      "pageEnd": 768,
      "target": "../sections/page-768.md",
      "narrative": null,
      "content": "Pooled specificity to first line drugs was 96%-99%, with the highest estimate for ethambutol \nand the lowest for rifampicin. This is again comparable to the results obtained from WGS \nsequencing from culture. Specificity for the flouroquinolones was 96%, indicating minimal \nloss in specificity for any gain in sensitivity through the detection of low frequency resistance \nalleles. \n \nAny recommendation to position tNGS as the first test of drug resistance needs to consider \na wide range of factors, including the added value of diagnosing isoniazid mono-resistance \nwhich is not detected by MTB/R",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the conclusions, reporting specificity of 96–99% for first-line drugs and 96% for fluoroquinolones and noting that positioning tNGS first would need head-to-head comparisons, cost and operational data. For rifampicin-resistant TB, sensitivity was around 68–70% for bedaquiline, clofazimine and linezolid, which the authors see as encouraging given the lack of other molecular tests, but they flag indeterminate rates of 10–20% as a concern.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-769",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 769",
      "pageStart": 769,
      "pageEnd": 769,
      "target": "../sections/page-769.md",
      "narrative": null,
      "content": "References \n \n1. Global Tuberculosis Report 2022. Geneva: World Health Organization; Licence: CC \nBY-NC-SA 3.0 IGO  \n2. Karmakar M, Ragonnet R, Ascher DB, Trauer JM, Denholm JT. Estimating \ntuberculosis drug resistance amplification rates in high-burden settings. BMC Infect \nDis. 2022 Jan 24;22(1):82.  \n3. The End TB Strategy. Geneva: World Health Organization; WHO/HTM/TB/2015.19. \nAvailable from: https://www.who.int/publications/i/item/WHO-HTM-TB-2015.19 \n(Accessed 6 April 2023) \n4. WHO consolidated guidelines on tuberculosis. Module 3: diagnosis - rapid diagnostics \nfor tuberculosis detectio",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The reference list for the tNGS accuracy review, citing WHO TB reports, strategies and guidelines and studies on resistance amplification and genomic sequencing for TB.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-770",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 770",
      "pageStart": 770,
      "pageEnd": 770,
      "target": "../sections/page-770.md",
      "narrative": null,
      "content": "14. Brankin AE, Fowler PW. Inclusion of minor alleles improves catalogue-based \nprediction of    \n      fluoroquinolone resistance in Mycobacterium tuberculosis. JAC Antimicrob Resist 2023 \nApr   \n      4;5(2) \n \n15. Catalogue of mutations in Mycobacterium tuberculosis complex and their association \nwith drug resistance. Geneva: World Health Organization; 2021. Licence: CC BY-NC-SA \n3.0 IGO",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Ends the reference list with a study on minor alleles and fluoroquinolone resistance and WHO's catalogue of TB resistance mutations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-771",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 771",
      "pageStart": 771,
      "pageEnd": 771,
      "target": "../sections/page-771.md",
      "narrative": null,
      "content": "Web Annex B.17. A Systematic and Scoping Review of the \nEconomic Evidence around Targeted Next Generation \nSequencing (tNGS) for the Diagnosis of Drug-Resistant \nTuberculosis (DR-TB) \n \n \n \n \n \n \n \n \nPrepared for WHO Global TB Programme, Guideline Development Group \n \n \n \n \n \n \n \n \nPatrick Galange1, Suvesh Shrestha2, Israa Zareef2, Alice Zwerling2 \n1 Dalla Lana School of Public Health, University of Toronto, Toronto, Canada \n2 School of Epidemiology & Public Health, University of Ottawa, Ottawa, Canada \n \n \nCorresponding Author: \nAlice Zwerling \nazwerlin@uottawa.ca \n \n \n \n \n \nEXECUTIVE SUMMARY",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title page of Web Annex B.17, a systematic and scoping review of the economic evidence on targeted next-generation sequencing for diagnosing drug-resistant TB, prepared for the WHO Global TB Programme by authors from the universities of Toronto and Ottawa.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-772",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 772",
      "pageStart": 772,
      "pageEnd": 772,
      "target": "../sections/page-772.md",
      "narrative": null,
      "content": "Background: Tuberculosis (TB) remains one of the leading causes of morbidity and mortality \nglobally. Our review aims to summarize the current economic evidence around using targeted Next \nGeneration Sequencing (tNGS) for the diagnosis of DR-TB. \nMethods: We conducted a systematic review on the economic evaluation of using either tNGS or \nWGS to diagnose DR-TB. The following three databases were used: PubMed, EMBASE and \nSCOPUS. The search was run on October 30, 2022. We did not restrict by year of publication; age \ngroup; country; income-level; comparator group; HIV status, or other comorbidi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarises the economic review, which searched three databases in 2022 and included ten studies on tNGS and whole genome sequencing for drug-resistant TB, finding tNGS costs of about US$70 per sample and wider ranges for whole genome sequencing. It identifies sequencing as the costliest step, reagents and consumables as the main cost components and drivers such as sequencer choice, coverage depth and batching, and calls for more cost-effectiveness research.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-773",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 773",
      "pageStart": 773,
      "pageEnd": 773,
      "target": "../sections/page-773.md",
      "narrative": null,
      "content": "1. INTRODUCTION  \n \nTuberculosis remains one of the leading causes of morbidity and mortality globally (1). In 2021, the \nWorld Health Organization (WHO) estimated approximately 10.6 million new cases of TB \noccurred, and over 1.6 million deaths, the latter being an increase in mortality compared to trends \nseen between 2005 and 2019 (2, 3). DR-TB was first recognized in the late 1940s and has become \nan increasing public health concern (2). DR-TB can be further classified as Rifampin-resistant TB \n(RR-TB), Isoniazid-resistant TB (IR-TB), Multidrug-resistant TB (MDR-TB), Pre-extensively \nDrug-",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces the economic review, citing WHO estimates of 10.6 million TB cases and 1.6 million deaths in 2021 and defining forms of drug-resistant TB from rifampicin-resistant to extensively drug-resistant. It notes that resistance is concentrated in some high-burden countries and regions. It then explains that conventional phenotypic testing is slow and needs high-level biosafety laboratories and skilled staff.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-774",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 774",
      "pageStart": 774,
      "pageEnd": 774,
      "target": "../sections/page-774.md",
      "narrative": null,
      "content": "delays in diagnosis and treatment, negatively impacting patient outcomes (17, 18).  \n \nAlternative second-line diagnostic approaches for DR include Line Probe Assays (LPAs), which \navoid many of the steps in conventional pDST (19). This is another type of probe-based molecular \ntest. In 2008, the WHO approved the use of LPAs for the diagnosis of MDR-TB, and there are \ncurrently ten different types commercially available (19, 20). Most LPAs can only diagnose \nrifampin resistance via the rpoB gene, and/or isoniazid resistance via the katG gene; but newer \nmodalities can also detect resistance in",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes line-probe assays as faster molecular alternatives with limits for pre-XDR, XDR and extrapulmonary TB, and explains next-generation sequencing approaches, including whole genome sequencing, tNGS and metagenomics, and their workflow steps. It then begins the methods, noting the review followed PRISMA guidelines with a librarian-supported search strategy.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-775",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 775",
      "pageStart": 775,
      "pageEnd": 775,
      "target": "../sections/page-775.md",
      "narrative": null,
      "content": "DR-TB); concept 2 (economic data, such as cost or cost-effectiveness); and concept 3 (sequencing \ntechnologies, such as tNGS).  \n \nWe used the following three electronic databases to perform the search as of October 30, 2022: \nPubMed, EMBASE, and SCOPUS for new studies, not restricted by year of publication or \nlanguage. Additionally, we screened the reference list of review articles and key identified articles \nfor additional studies.   \n \nFor the systematic review, we included studies that looked at participants with a confirmed or \nsuspected diagnosis of pulmonary or non-pulmonary DR-TB via",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the search and eligibility methods of systematic and scoping reviews on the costs of targeted next-generation sequencing (tNGS) and whole-genome sequencing (WGS) for diagnosing drug-resistant TB. It lists the databases searched in October 2022, the inclusion of studies on confirmed or suspected DR-TB, and the cost outcomes of interest such as cost per sample, component costs, costs per step and ICERs.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-776",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 776",
      "pageStart": 776,
      "pageEnd": 776,
      "target": "../sections/page-776.md",
      "narrative": null,
      "content": "2.4. Statistical Analysis  \nDue to the limited number of studies available and the high degree of methodological heterogeneity \nbetween studies, a meta-analysis was deemed not appropriate. Instead, a narrative summary review \nwas performed. \n \nAll cost data was inflated to 2021 in local currency and converted to 2021 USD using the relevant \nInternational Monetary Fund (IMF) exchange rates and inflation index (36). We contacted study \nauthors for missing or unclear data.  \n \n2.5. Assessing the Quality of Evidence  \nWe used a modified version of the Consolidated Health Economic Evaluation Report",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains that heterogeneity ruled out meta-analysis, so a narrative synthesis was done with costs converted to 2021 US dollars, and that study quality was rated with a modified CHEERS checklist. It then describes which cost measures the included tNGS and WGS studies reported and notes that few costed the phenotypic comparator.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-777",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 777",
      "pageStart": 777,
      "pageEnd": 777,
      "target": "../sections/page-777.md",
      "narrative": null,
      "content": "3.1.1. Economic Evidence Around tNGS \n3.1.1.1. Summary of Findings \nSix studies reported on costs for tNGS approaches (Table 1). Three of the studies were performed in \nthe US, one collected specimen from both Ethiopia and Hong Kong, one in the UK, and one in \nMoldova with year of cost valuation ranging from 2018-2021 (Table 1). Of the tNGS studies, two \nused on-site testing, and four required sample transport to a central lab for processing (Table 1). We \nidentified two studies that assessed diagnostic cost per sample for tNGS, and one that assessed this \nfor pDST as a diagnostic comparator.",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarizes six studies reporting tNGS costs for drug-resistant TB, their settings and whether testing was on site or centralized. It reports the limited data on cost per sample and the wide range in sequencer prices, from about $1 000 for a Nanopore MinION to over $200 000 for an Illumina NextSeq.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-778",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 778",
      "pageStart": 778,
      "pageEnd": 778,
      "target": "../sections/page-778.md",
      "narrative": null,
      "content": "MinION. The cost of reagents and consumables also varied by identity of sequencing manufacturer, \nwith the lowest being Illumina MiSeq100, and the highest Nanopore MinION. However, this was \nbased on limited data (n=1). The per sample cost for reagents and consumables was found to be as \nlow as $6.20 for MiSeq and $15.32 for iSeq100 (Table 7B). For this determination, samples had \n15,000-fold coverage, which account for 99% of the genome, and a 30x-fold depth of coverage. For \nChan et al. (2020), with Nanopore MinION, reagents and consumables were as high as $67.02 per \nsample per 24-plex work",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports per-sample reagent and consumable costs for different tNGS platforms, which varied widely by manufacturer and workflow. It then presents a five-year budget impact analysis comparing tNGS, WGS and phenotypic testing scenarios, with start-up costs in the first year driven by equipment.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-779",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 779",
      "pageStart": 779,
      "pageEnd": 779,
      "target": "../sections/page-779.md",
      "narrative": null,
      "content": "$74,507.00) scenarios had additional projected start-up costs in year 1 due to capital expenditures \nand equipment costs (39). \n \nThe results of the economic evidence of diagnosing DR-TB with WGS are summarized in \nAppendix 8.1.  \n \n3.1.3. Quality of Health Economic Studies \nOur quality of health economic studies extraction form was used to evaluate the quality of included \ntNGS (Supplementary Tables 1A and 1B) and WGS (Supplementary Tables 2A and 2B) \nmanuscripts. This contained 17 questions, and was adapted from a modified CHEERS, as described \nearlier. Quality scores varied from 9/17 to 14/",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes the extra first-year start-up costs in the budget scenarios and points to an appendix for WGS results. It reports modified CHEERS quality scores for the tNGS and WGS studies, then begins the scoping review of sequencing costs for other diseases, giving cost per sample ranges and reagent costs.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-780",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 780",
      "pageStart": 780,
      "pageEnd": 780,
      "target": "../sections/page-780.md",
      "narrative": null,
      "content": "$104.61, both for HIV. The remaining studies were between $33.49 to $62.72 for reagent and \nconsumables. Only the manuscript by Gachogo et al. (2020) assessed component costs per sample \nfor other components in addition to reagents and consumables (50). These ranged from $2.44 for \nmaintenance per sample, to $105.56 for equipment per sample. The paper by Merel et al. (2001) \nwas the only one that assessed total capital costs per sequencer, but this was not incorporated into \nthe per sample cost (52). These ranged in cost from $127,132.98 for the capillary electrophoresis \n(CEQ 2000) in-house p",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the scoping review findings on sequencing costs for diseases such as HIV, covering component costs, sequencer capital costs and costs per step. It lists factors that influence costs, such as laboratory efficiency, staff, technology and volume discounts, and begins the quality assessment.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-781",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 781",
      "pageStart": 781,
      "pageEnd": 781,
      "target": "../sections/page-781.md",
      "narrative": null,
      "content": "from a modified CHEERS, as described earlier. Quality scores varied from 9/21 to 19/21 with an \naverage quality score of 12/21 among. The summary of this assessment is displayed graphically on \nFigure 4. \n \n \n4. DISCUSSION  \n4.1. Interpretation of Systematic Review Findings \nTo the best of our knowledge, this systematic review is the first of its kind to look at the economic \nevidence around using tNGS for the diagnosis of DR-TB. tNGS is a relatively new technology, and \ndespite broad inclusion criteria, only ten suitable articles were found. There were no manuscripts \nwith direct comparisons",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives quality scores for the scoping review studies and opens the discussion, describing the systematic review as the first on tNGS economics for drug-resistant TB and noting the scarcity of data. It discusses why per-sample costs vary, including platform choice, coverage depth and early-run inefficiencies reported in one study.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-782",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 782",
      "pageStart": 782,
      "pageEnd": 782,
      "target": "../sections/page-782.md",
      "narrative": null,
      "content": "Lastly, lower sample volumes and less batching per run are likely to be associated with a greater \ncosts per sample. This includes batching across TB samples or multiplexing with different \norganisms. While not explicitly mentioned in any of the papers, this could likely impact costs by \nspreading out certain overhead (e.g., labor costs) or equipment costs (e.g., sequencer flow cell) \namong more samples. This latter impact was discussed with the authors of one of the included \nmanuscripts (41).  \n \n4.2. Interpretation of Scoping Review Findings \nSimilar to our systematic review and to the best",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Adds that low sample volumes and limited batching raise per-sample sequencing costs. It then interprets the scoping review, highlighting a modelling study in which molecular testing plus WGS was the most cost-effective TB diagnostic strategy owing to greater sensitivity and faster diagnosis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p783-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 783,
      "pageEnd": 783,
      "target": "../images/img-p783-1.webp",
      "narrative": "draft",
      "content": "A PRISMA flow diagram for the systematic review: 100 studies imported for screening, 21 duplicates removed; 79 studies screened, 68 irrelevant; 11 full-text studies assessed for eligibility, 1 excluded for wrong intervention; 0 studies ongoing and 0 awaiting classification; 10 studies included.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-783",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 783",
      "pageStart": 783,
      "pageEnd": 783,
      "target": "../sections/page-783.md",
      "narrative": null,
      "content": "5. CONCLUSION \n \nGiven the growing interest in tNGS as a diagnostic modality, there are a number of additional data \npoints that would be helpful to characterize. First, further studies are needed to determine additional \neconomic markers, such as cost-effectiveness analyses or ICERs. Second, it will be prudent to \ndevelop a better understanding of the various factors that may impact cost, such as the impact of \nbatching and multiplexing, as well as country-dependent factors. Lastly, it would be helpful to have \nmore data points that include all relevant costs, such as overhead, total lab equi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Concludes that more research is needed on cost-effectiveness, cost drivers such as batching and multiplexing, and full costing including overheads before tNGS and WGS economics for drug-resistant TB are well understood. It introduces the PRISMA flow diagram for the systematic review.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p784-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 784,
      "pageEnd": 784,
      "target": "../images/img-p784-1.webp",
      "narrative": "draft",
      "content": "A PRISMA flow diagram for the scoping review: 1054 studies imported for screening, 246 duplicates removed; 808 studies screened, 790 irrelevant; 18 full-text studies assessed for eligibility, 5 excluded for wrong outcomes; 0 studies ongoing and 0 awaiting classification; 13 studies included.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-784",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 784",
      "pageStart": 784,
      "pageEnd": 784,
      "target": "../sections/page-784.md",
      "narrative": null,
      "content": "Figure 2. PRISMA Flow Diagram for the Scoping Review",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A figure caption for the PRISMA flow diagram of the scoping review.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-785",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 785",
      "pageStart": 785,
      "pageEnd": 785,
      "target": "../sections/page-785.md",
      "narrative": null,
      "content": "_(no text on this page)_",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A blank page with no extracted text.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-786",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 786",
      "pageStart": 786,
      "pageEnd": 786,
      "target": "../sections/page-786.md",
      "narrative": null,
      "content": "781 \n \nTable 1. Characteristics of Included tNGS Studies for the Systematic Review \n \nManuscript \nCountry \nsetting \nYear of Cost \nValuation \nCurrency \nStudy Population  \nDiagnostic \nStrategies (tNGS or \nWGS) \nReference \nDiagnostic \nStrategies (pDST \nor mWRDs)  \nAnalysis \nPerspective \n(Healthcare or \nSocietal) \nType of \nEconomic \nEvaluation  \nPrimary Outcome \nSecondary Outcome(s) \nSource of Costing  \nColman RE et al \n(2019) \nUSA \n2018 \nUSD \nTB isolates \ntNGS (iSeq100 and \nmiSeq) and WGS \n(iSeq100 and MiSeq) \nNone \nHealthcare \nSome cost data  \nCost per sample  \nCapital costs of \nsequencing \ninst",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A table describing the included tNGS studies in the systematic review, giving country, cost year, currency, population, sequencing strategy, comparator, perspective, type of evaluation and cost sources.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-787",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 787",
      "pageStart": 787,
      "pageEnd": 787,
      "target": "../sections/page-787.md",
      "narrative": null,
      "content": "782 \n \n \nTable 2A. Characteristics of Included WGS Studies for the Systematic Review \n \nManuscript \nCountry Setting \nYear of Cost Valuation \nCurrency \nStudy Population \nDiagnostic Strategies (tNGS or WGS) \nReference Diagnostic Strategies (pDST \nor mWRDs) \nCirillo DM et al (2016) \nItaly \n2016 \nEUR \nConfirmed MDR/RR-TB cases in general population \nWGS  \nStandard testing \nPankhurst LJ et al (2016) \nEurope (UK, Ireland, \nGermany and France) and \nNorth America (Canada) \n2014 \nGBP \nConfirmed MDR/RR-TB cases in general population \nWGS  \nRoutine MTBC diagnostic workflows \nColman RE et al (2019) \nUSA",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Tables describing the included WGS studies in the systematic review, with their settings, cost years, populations, comparators, perspectives, outcomes and costing sources.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-788",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 788",
      "pageStart": 788,
      "pageEnd": 788,
      "target": "../sections/page-788.md",
      "narrative": null,
      "content": "783 \n \n \nTable 3A. Characteristics of Included Studies for the Scoping Review \n \n \nManuscript \nCommunicable disease \nCountry \nsetting \nLaboratory \nclassification \n(peripheral, \nintermediate or \ncentral) \nYear of \ncost \nvaluation \nCurrency \nStudy population  \nDiagnostic strategies \n(tNGS or WGS) \nReference \ndiagnostic \nstrategies  \nAnalysis \nperspective \n(healthcare or \nsocietal) \nType of \neconomic \nevaluation  \nType of \nmodel if \napplicable \nWTP \nthreshold if \napplicable \nPrimary outcome \nSecondary \noutcome(s) \nSource of costing  \nAlleweldt et al \n(2021) \nAvian influenza; \ninfluenza A&B; food-",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A table describing the studies in the scoping review on sequencing costs for other diseases such as influenza, food-borne pathogens, HIV and malaria, with settings, laboratory levels, strategies and outcomes.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-789",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 789",
      "pageStart": 789,
      "pageEnd": 789,
      "target": "../sections/page-789.md",
      "narrative": null,
      "content": "784 \n \n \nTable 3B. Characteristics of Included Studies for the Scoping Review \n \n \n \n \nManuscript \nCommunicable disease \nCountry \nsetting \nLaboratory \nclassification \n(peripheral, \nintermediate or \ncentral) \nYear of \ncost \nvaluation \nCurrency \nStudy population  \nDiagnostic strategies \n(tNGS or WGS) \nReference \ndiagnostic \nstrategies  \nAnalysis \nperspective \n(health care or \nsocietal) \nType of \neconomic \nevaluation \nType of \nmodel if \napplicable \nWTP \nthreshold if \napplicable \nPrimary outcome \nSecondary \noutcome(s) \nSource of costing  \nMerel et al \n(2001) \nHIV \nFrance \nCentral  \nNot stated \nUSD",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the table of scoping review studies, covering HIV and hepatitis sequencing studies including a cost-effectiveness model of genotypic resistance testing.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-790",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 790",
      "pageStart": 790,
      "pageEnd": 790,
      "target": "../sections/page-790.md",
      "narrative": null,
      "content": "785 \n \nTable 4. Methodologies of Included tNGS Studies for the Systematic Review \n \nManuscript \nDepth for costing \nCoverage for costing \nVarying Sample Numbers for \nCosting \nDifferent Scenarios Modelled \nfor Costing \nKey Scenarios/Variables Explored in \nSensitivity Analyses (PSA, One Way, \nor Two Way)  \nTime to Completion \nColman RE et al (2019) \nYes \nNo \nYes \nNo \nNo \nMiSeq 24 hours; iSeq100 17.5 \nhours for sequencing \nChan WS et al (2020) \nNo \nNo \nNo \nNo \nNo \nMiSeq 48.05 hours; MinION 8.68 \nhours; pDST 69.5 days for end-to-\nend solution \nTafess K et al (2020) \nNo \nNo \nNo \nNo \nNo \nMiSeq 38 hou",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Tables of methods used in the included tNGS and WGS studies, showing whether costing considered depth, coverage, sample numbers, scenarios and sensitivity analyses, plus turnaround times.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-791",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 791",
      "pageStart": 791,
      "pageEnd": 791,
      "target": "../sections/page-791.md",
      "narrative": null,
      "content": "786 \n \n1. The paper by Rowlinson MC and Musser KA (2020) is a report \n \nTable 6. Methodologies of Included Studies for the Scoping Review \n \nManuscript \nDepth for costing \nCoverage for costing \nVarying sample \nnumbers for \ncosting \nDifferent scenarios \nmodelled for costing \nKey scenarios/variables explored in \nsensitivity analyses (PSA, one way, or two \nway)  \nDid authors include \ninformation on feedback \nand/or knowledge \ntranslation? \nTime to completion \nAlleweldt et al (2021) \nNo \nNo \nNo \nNo \nNo \nNo \nN/A \nChaturbhuj et al (2014) \nNo \nNo \nNo \nNo \nNo \nNo \nN/A \nDimitrov et al (2017) \nNo \nNo \nN",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A table of methods used in the scoping review studies, showing whether costing considered depth, coverage, sample numbers, scenarios, sensitivity analyses and feedback, along with turnaround times.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-792",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 792",
      "pageStart": 792,
      "pageEnd": 792,
      "target": "../sections/page-792.md",
      "narrative": null,
      "content": "787 \n \nTable 7A. Output from Included tNGS Studies for the Systematic Review: Diagnostic Cost per Sample \n \nManuscript \nDiagnostic Cost per Sample1,2,3 \nDiagnostic Intervention (tNGS) \nIllumina MiSeq (15,000x coverage) \nNanopore MinION \nTafess K et al (2020) \n$69.64 \n  \n$73.47 \n  \n1. Values are written in USD 2021 unless otherwise specified \n2. Value looks at Illumina MiSeq and for lower and upper values, respectively, both at 15,000x coverage \n3. Only studies with published data for each outcome were included \n \nTable 7B. Output from Included tNGS Studies for the Systematic Review: Component",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Tables of tNGS outputs from the systematic review, giving diagnostic cost per sample for MiSeq and MinION and component costs for instruments and reagents, with notes on how costs were estimated.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-023",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 7A. Output from Included tNGS Studies for the Systematic Review: Diagnostic Cost per Sample",
      "pageStart": 792,
      "pageEnd": 792,
      "target": "../tables/table-7a.csv",
      "narrative": "draft",
      "content": "Diagnostic cost per sample of targeted next-generation sequencing (tNGS) reported by an included study, for Illumina MiSeq and Nanopore MinION platforms (2021 USD).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/table-024",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 7B. Output from Included tNGS Studies for the Systematic Review: Component Costs",
      "pageStart": 792,
      "pageEnd": 792,
      "target": "../tables/table-7b.csv",
      "narrative": "draft",
      "content": "Component costs of targeted next-generation sequencing (tNGS) reported by four included studies: sequencing instruments (Illumina MiSeq, Illumina NextSeq, ONT MinION), reagents and consumables (2021 USD).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-793",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 793",
      "pageStart": 793,
      "pageEnd": 793,
      "target": "../sections/page-793.md",
      "narrative": null,
      "content": "788 \n \nTable 7C. Output from Included tNGS Studies for the Systematic Review: Costs per step \n \nManuscript \nCosts per step1,3 \nDiagnostic Intervention (tNGS Illumina MiSeq) \nDiagnostic Intervention (Nanopore MinION) \nTarget \nEnrichment \nMultiplex PCR \nLibrary \nPrep \nQuality \nControl \nSequencing \nDNA Extraction \nTarget \nEnrichment \nMultiplex PCR \nLibrary Prep \nQuality \nControl \nSequencing \nDNA Extraction \nTafess K et al (2020)2 \n$4.15 \n \n \n  \n$28.56 \n \n  \n$13.78 \n \n  \n$19.52 \n  \n $3.64 \n  \n$4.15 \n  \n$16.11 \n  \n$2.52 \n  \n$47.08 \n  \n  $3.64 \n \n1. Values are written in 2021 USD unless otherwise sp",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Tables of tNGS costs per step, such as library preparation and sequencing, and of component costs per sample from one study comparing sequencing strategies with phenotypic testing.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-025",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 7C. Output from Included tNGS Studies for the Systematic Review: Costs per step",
      "pageStart": 793,
      "pageEnd": 793,
      "target": "../tables/table-7c.csv",
      "narrative": "draft",
      "content": "Costs per laboratory step (target enrichment by multiplex PCR, library preparation, quality control, sequencing and DNA extraction) of targeted next-generation sequencing on Illumina MiSeq and Nanopore MinION, from one included study (2021 USD).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/table-026",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 8A. Output from Included tNGS and WGS Studies for the Systematic Review: Costing Data, Cates et al. (2022)",
      "pageStart": 793,
      "pageEnd": 793,
      "target": "../tables/table-8a.csv",
      "narrative": "draft",
      "content": "Diagnostic cost per sample reported by Cates et al. (2022), broken down into labour, equipment, reagents and consumables, training, maintenance and total (2021 USD), for three testing scenarios: SSM+ or Xpert+, excluding Xpert/RIF- (Diagnostic Intervention (tNGS Illumina iSeq100)); All SSM+ or Xpert+ (Diagnostic Intervention (tNGS Illumina iSeq100)); All SSM+ or Xpert+, plus pDST (Diagnostic Intervention (tNGS Illumina iSeq100)).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/table-027",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 8B. Output from Included tNGS and WGS Studies for the Systematic Review: Costing Data, Cates et al. (2022)",
      "pageStart": 793,
      "pageEnd": 793,
      "target": "../tables/table-8b.csv",
      "narrative": "draft",
      "content": "Diagnostic cost per sample reported by Cates et al. (2022), broken down into labour, equipment, reagents and consumables, training, maintenance and total (2021 USD), for three testing scenarios: All culture+ (Diagnostic Intervention (tNGS Illumina iSeq100)); All culture+ excluding Xpert/RIF- (Diagnostic Intervention (tNGS Illumina iSeq100)); All culture+, plust pDST (Diagnostic Intervention (tNGS Illumina iSeq100)).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-794",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 794",
      "pageStart": 794,
      "pageEnd": 794,
      "target": "../sections/page-794.md",
      "narrative": null,
      "content": "789 \n \nTable 8C. Output from Included tNGS and WGS Studies for the Systematic Review: Costing Data, Cates et al. (2022) \n \nManuscript \nDiagnostic cost per sample1,2 \nDiagnostic Intervention (tNGS Illumina iSeq100) \nDiagnostic Intervention (WGS Illumina MiSeq) \nAll culture+ excluding Xpert/RIF-, plus pDST \nAll culture+ \nAll culture+ excluding Xpert/RIF- \nLabor \nEquipment \nReagents & \nConsumables \nTraining \nMaintenance \nTotal \nLabor \nEquipment \nReagents & \nConsumables \nTraining \nMaintenance \nTotal \nLabor  \nEquipment \nReagents & \nConsumables \nTraining \nMaintenance \nTotal \nCates L et \nal (2022)",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Tables of costing data and five-year budget impact estimates from one study comparing tNGS, WGS and phenotypic drug susceptibility testing scenarios for culture-positive samples.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-028",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 8C. Output from Included tNGS and WGS Studies for the Systematic Review: Costing Data, Cates et al. (2022)",
      "pageStart": 794,
      "pageEnd": 794,
      "target": "../tables/table-8c.csv",
      "narrative": "draft",
      "content": "Diagnostic cost per sample reported by Cates et al. (2022), broken down into labour, equipment, reagents and consumables, training, maintenance and total (2021 USD), for three testing scenarios: All culture+ excluding Xpert/RIF-, plus pDST (Diagnostic Intervention (tNGS Illumina iSeq100)); All culture+ (Diagnostic Intervention (WGS Illumina MiSeq)); All culture+ excluding Xpert/RIF- (Diagnostic Intervention (WGS Illumina MiSeq)).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/table-029",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 8D. Output from Included tNGS and WGS Studies for the Systematic Review: Costing Data, Cates et al. (2022)",
      "pageStart": 794,
      "pageEnd": 794,
      "target": "../tables/table-8d.csv",
      "narrative": "draft",
      "content": "Diagnostic cost per sample reported by Cates et al. (2022), broken down into labour, equipment, reagents and consumables, training, maintenance and total (2021 USD), for three testing scenarios: All culture+, plust pDST (); All culture+ excluding Xpert/RIF-, plus pDST (); N/A (Diagnostic Comparitor (pDST)).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/table-030",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 9A. Output from Included tNGS and WGS Studies for the Systematic Review: BIA, Cates et al. (2022)",
      "pageStart": 794,
      "pageEnd": 794,
      "target": "../tables/table-9a.csv",
      "narrative": "draft",
      "content": "Total budget impact estimates reported by Cates et al. (2022) for years 1 (start-up) to 5 and in total (2021 USD), for three testing strategies: pDST; tNGS for all culture+; tNGS for all culture+, excluding Xpert/RIF-.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-795",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 795",
      "pageStart": 795,
      "pageEnd": 795,
      "target": "../sections/page-795.md",
      "narrative": null,
      "content": "790 \n \nTable 9B. Output from Included tNGS and WGS Studies for the Systematic Review: BIA, Cates et al. (2022) \n \nManuscript \nTotal Budget Impact Estimates1 \ntNGS for all SSM+ or Xpert+ \ntNGS for all SSM+ or Xpert+, excluding Xpert/RIF- \ntNGS for all SSM+ or Xpert+, plus pDST \nYear 1 (start-up) \nYear 2 \nYear 3 \nYear 4 \nYear 5 \nTotal \nYear 1 (start-up) \nYear 2 \nYear 3 \nYear 4 \nYear 5 \nTotal \nYear 1 (start-\nup) \nYear 2 \nYear 3 \nYear 4 \nYear 5 \nTotal \nCates L et \nal (2022) \n$180,787.00  \n$137,132.00  \n$133,646.00  \n$130,265.00  \n$126,985.00  \n$708,816.00  \n$136,029.00  \n$93,717.00  \n$91,534.00",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Further budget impact tables from the same study, giving yearly and five-year totals for tNGS and WGS testing scenarios in 2021 US dollars.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-031",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 9B. Output from Included tNGS and WGS Studies for the Systematic Review: BIA, Cates et al. (2022)",
      "pageStart": 795,
      "pageEnd": 795,
      "target": "../tables/table-9b.csv",
      "narrative": "draft",
      "content": "Total budget impact estimates reported by Cates et al. (2022) for years 1 (start-up) to 5 and in total (2021 USD), for three testing strategies: tNGS for all SSM+ or Xpert+; tNGS for all SSM+ or Xpert+, excluding Xpert/RIF-; tNGS for all SSM+ or Xpert+, plus pDST.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/table-032",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 9C. Output from Included tNGS and WGS Studies for the Systematic Review: BIA, Cates et al. (2022)",
      "pageStart": 795,
      "pageEnd": 795,
      "target": "../tables/table-9c.csv",
      "narrative": "draft",
      "content": "Total budget impact estimates reported by Cates et al. (2022) for years 1 (start-up) to 5 and in total (2021 USD), for three testing strategies: tNGS for all culture+ plus pDST; tNGS for all culture+, excluding Xpert/RIF-, plus pDST; WGS for all culture+.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/table-033",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 9D. Output from Included tNGS and WGS Studies for the Systematic Review: BIA, Cates et al. (2022)",
      "pageStart": 795,
      "pageEnd": 795,
      "target": "../tables/table-9d.csv",
      "narrative": "draft",
      "content": "Total budget impact estimates reported by Cates et al. (2022) for years 1 (start-up) to 5 and in total (2021 USD), for three testing strategies: WGS for all culture+ excluding Xpert/RIF-; WGS for all culture+, plus pDST; WGS for all culture+, excluding Xpert/RIF-, plus pDST.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-796",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 796",
      "pageStart": 796,
      "pageEnd": 796,
      "target": "../sections/page-796.md",
      "narrative": null,
      "content": "791 \n \n \nTable 10A. Output from Included WGS Studies for the Systematic Review: Diagnostic Cost per Sample \n \nManuscript \nDiagnostic cost per sample1,7 \nDiagnostic Intervention (WGS) \nDiagnostic Comparator \nIllumina NextSeq \nIllumina MiSeq \nNanopore MinION \npDST \nCirillo DM et al (2016)10 \n  \n$185.77 \n \n  \n  \nPankhurst LJ et al (2016)5 \n  \n$110.87 ($109.67-112.38)4 \n  \n$49.00 ($48.60-49.47)4 \nHe G et al (2020) \n  \n$71.61 \n \n  \n  \n$45.47-102.303 \nVogel M et al (2021) \n  \n$141.00 - $277.006 \n  \n  \nRowlinson MC and Musser KA (2022)2 \n$68.00  \n$130.00 - $150.00 \n$63.00  \n  \n1. Values are written i",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Tables of WGS outputs from the systematic review, giving diagnostic cost per sample by platform compared with phenotypic testing, and component costs, with explanatory notes.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-034",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 10A. Output from Included WGS Studies for the Systematic Review: Diagnostic Cost per Sample",
      "pageStart": 796,
      "pageEnd": 796,
      "target": "../tables/table-10a.csv",
      "narrative": "draft",
      "content": "Diagnostic cost per sample of whole-genome sequencing (WGS) on Illumina NextSeq, Illumina MiSeq and Nanopore MinION, and of the phenotypic DST comparator, as reported by five included studies (2021 USD).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/table-035",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 10B. Output from Included WGS Studies for the Systematic Review: Component Costs",
      "pageStart": 796,
      "pageEnd": 796,
      "target": "../tables/table-10b.csv",
      "narrative": "draft",
      "content": "Component costs of whole-genome sequencing (equipment, sequencers, maintenance, labour, training, overhead, reagents and consumables) and of the phenotypic DST comparator, as reported by three included studies (2021 USD).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-797",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 797",
      "pageStart": 797,
      "pageEnd": 797,
      "target": "../sections/page-797.md",
      "narrative": null,
      "content": "792 \n \nTable 10C. Output from Included WGS Studies for the Systematic Review: Costs per step \n \nManuscript \nCosts per step1,3 \nDiagnostic Intervention (WGS) \nDNA Extraction \nTarget Enrichment \nMultiplex PCR \nLibrary Prep \nQuality Control \nSequencing \nVogel M et al (2021)2 \n$4.05  \n$1.31 - $2.88 \n$59.61 - $91.25 \n$3.83 - $14.70 \n$64.53 - $156.32 \n1. Values are written in 2021 USD unless otherwise specified. \n2. Range based on first 174 sample ($277.00), and subsequent samples for 500-cycles ($167.00) and 600-cycles ($141.00). \n3. Only studies with published data were for each outcome were inclu",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A table of WGS costs per step from one study, from DNA extraction to sequencing.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-036",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 10C. Output from Included WGS Studies for the Systematic Review: Costs per step",
      "pageStart": 797,
      "pageEnd": 797,
      "target": "../tables/table-10c.csv",
      "narrative": "draft",
      "content": "Costs per laboratory step of whole-genome sequencing (DNA extraction, target enrichment by multiplex PCR, library preparation, quality control and sequencing) from one included study (2021 USD).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-798",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 798",
      "pageStart": 798,
      "pageEnd": 798,
      "target": "../sections/page-798.md",
      "narrative": null,
      "content": "793 \n \nTable 11A. Output from Included Studies for the Scoping Review: Key Outcome(s) and Component Costs  \n \nManuscript \n  \n  \nKey outcome(s) \n Component costs per sample1,5 \n  \n  \n  \n  \n  \n  \n  \n Diagnostic Intervention \n  \n  \n  \n  \n  \n  \n  \nEquipment \nMaintenance \nStaffing \nOverhead \nReagents and \nconsumables \nQuality \nAssurance \nGachogo et al \n(2020) \nTo establish detailed cost profile \nfor HIVDR testing and to identify \ncost drivers \n$105.56 \n$2.44  \n$48.25  \n$15.14  \n$104.61  \n$4.12 \nKunasol et al \n(2022)2 \nTargeted Amplicon Deep \nsequencing (TADs) using Ion \nTorrent PGM with Illumina Mi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A table of key outcomes and per-sample component costs from scoping review studies on HIV, malaria and hepatitis sequencing, comparing platforms and methods.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-037",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 11A. Output from Included Studies for the Scoping Review: Key Outcome(s) and Component Costs",
      "pageStart": 798,
      "pageEnd": 798,
      "target": "../tables/table-11a.csv",
      "narrative": "draft",
      "content": "Key outcomes and component costs per sample (equipment, maintenance, staffing, overhead, reagents and consumables, quality assurance) of sequencing-based diagnostic interventions reported by five studies in the scoping review (USD).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-799",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 799",
      "pageStart": 799,
      "pageEnd": 799,
      "target": "../sections/page-799.md",
      "narrative": null,
      "content": "794 \n \nTable 11B. Output from Included Studies for the Scoping Review: Costs per step \n \nManuscript \n  \n  \nCosts per step1,2 \n  \n  \n  \n  \n  \n  \nDiagnostic Intervention  \n  \n  \n  \n  \n  \n  \nSample \ncollection \nTarget \nEnrichment \nMultiplex \nPCR \nSequencing \nDNA or \nRNA \nExtraction \nGel \nelectrophoresis \nSequencing \nanalysis \nDimitrov et al \n(2017) \n  \n$55.29  \n$56.40 \n$5.53 \n  \n  \nGachogo et \nal (2020) \n$2.48 \n$57.86  \n$165.88  \n$23.49  \n$10.66 \n$19.75  \nTaylor et al \n(2013) \n  \n$1.99 \n$1093.49 \n$1.99 \n  \n  \n1. All costs are in USD unless stated otherwise. \n2. Only manuscripts with data were inc",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A table of costs per step from three scoping review studies, from sample collection to sequencing analysis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-038",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 11B. Output from Included Studies for the Scoping Review: Costs per step",
      "pageStart": 799,
      "pageEnd": 799,
      "target": "../tables/table-11b.csv",
      "narrative": "draft",
      "content": "Costs per laboratory step (sample collection, target enrichment by multiplex PCR, sequencing, DNA or RNA extraction, gel electrophoresis and sequencing analysis) of sequencing-based diagnostic interventions reported by three studies in the scoping review (USD).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-800",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 800",
      "pageStart": 800,
      "pageEnd": 800,
      "target": "../sections/page-800.md",
      "narrative": null,
      "content": "795 \n \nTable 11C. Output from Included Studies for the Scoping Review: ICERs, Cost per Sample and Component Costs Total \n \nManuscript1,7 \n  \n  \n ICERs \n  \n  \n  \n  \n  \nCost per Sample \n  \nComponent costs total \n  \n Diagnostic Intervention \n  \n  \n  \n  \n  \nDiagnostic \nintervention \nDiagnostic Comparator \nDiagnostic intervention \nDiagnostic Comparator \nQALYs \n(months) \nQALYs sens analysis \n- effectiveness of \nHAART (months) \nAdded QALYs sens \nanalysis - maximum \nduration of HAART \n(months) \nCost-effectiveness \nratio (CER, $/QALY \ngained) \nCER sens analysis - \neffectiveness of \nHAART ($/QALY \ngaine",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A table of ICERs, cost per sample and equipment costs from scoping review studies, including cost per QALY from an HIV resistance testing model, with notes on assumptions.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-039",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 11C. Output from Included Studies for the Scoping Review: ICERs, Cost per Sample and Component Costs Total",
      "pageStart": 800,
      "pageEnd": 800,
      "target": "../tables/table-11c.csv",
      "narrative": "draft",
      "content": "Incremental cost-effectiveness results (QALYs and cost-effectiveness ratios with sensitivity analyses), cost per sample for the diagnostic intervention and comparator, and total equipment costs reported by eight studies in the scoping review (USD).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-801",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 801",
      "pageStart": 801,
      "pageEnd": 801,
      "target": "../sections/page-801.md",
      "narrative": null,
      "content": "796 \n \nTable 12A. Output from Included Studies for the Scoping Review: Cost per Sample, Alleweldt et al. (2021) \n \nManuscript \nCost per sample \nWGS \n  \n  \n  \n  \n  \n  \n  \nConventional \n  \n  \n  \n  \n  \n  \n  \nAvian \ninfluenza \n(APHA) \nAvian \ninfluenza \n(FLI) \nInfluenza \nA+B (EMC) \nFood-borne \nPathogens \n(IZSLER) \nFood-borne \nPathogens \n(INEI-ANLIS) \nFood-borne \nPathogens \n(MDH) \nFood-borne \nPathogens \n(PHAC) \nFood-borne \nPathogens \n(PHE) \nAvian \ninfluenza \n(APHA) \nAvian \ninfluenza \n(FLI) \nInfluenza \nA+B (EMC) \nFood-borne \nPathogens \n(IZSLER) \nFood-borne \nPathogens \n(INEI-ANLIS) \nFood-borne \nPathog",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Tables of WGS and conventional method costs per sample and consumable costs for avian influenza, influenza and food-borne pathogen laboratories from one multi-country study, converted to 2021 US dollars.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-040",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 12A. Output from Included Studies for the Scoping Review: Cost per Sample, Alleweldt et al. (2021)",
      "pageStart": 801,
      "pageEnd": 801,
      "target": "../tables/table-12a.csv",
      "narrative": "draft",
      "content": "Cost per sample reported by Alleweldt et al. (2021) for whole-genome sequencing (WGS) and conventional methods, for eight laboratory settings covering avian influenza, influenza A+B and food-borne pathogens (2021 USD).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/table-041",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 12B. Output from Included Studies for the Scoping Review: Component Costs, Alleweldt et al. (2021)",
      "pageStart": 801,
      "pageEnd": 801,
      "target": "../tables/table-12b.csv",
      "narrative": "draft",
      "content": "Equipment component costs reported by Alleweldt et al. (2021) for whole-genome sequencing (WGS) and conventional methods, for eight laboratory settings covering avian influenza, influenza A+B and food-borne pathogens (2021 USD).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/table-042",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 12C. Output from Included Studies for the Scoping Review: Cost per Sample, Alleweldt et al. (2021)",
      "pageStart": 801,
      "pageEnd": 801,
      "target": "../tables/table-12c.csv",
      "narrative": "draft",
      "content": "Consumables component costs reported by Alleweldt et al. (2021) for whole-genome sequencing (WGS) and conventional methods, for eight laboratory settings covering avian influenza, influenza A+B and food-borne pathogens (2021 USD).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-802",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 802",
      "pageStart": 802,
      "pageEnd": 802,
      "target": "../sections/page-802.md",
      "narrative": null,
      "content": "797 \n \nTable 12D. Output from Included Studies for the Scoping Review: Cost per Sample, Alleweldt et al. (2021) \n \nManuscript \nComponent Costs \n  \nWGS \n  \n  \n  \n  \n  \n  \n  \nConventional \nMethod \n  \n  \n  \n  \n  \n  \n  \nStaff (prof and tech): \nAvian influenza APHA \nStaff (prof \nand tech): \nAvian \ninfluenza FLI \nStaff (prof \nand tech): \nInfluenza \nA+B EMC \nStaff (prof \nand tech): \nFood-borne \npathogens \nIZSLER \nStaff (prof \nand tech): \nFood-borne \npathogens \nARG \nStaff (prof \nand tech): \nFood-borne \npathogens \nMDH \nStaff (prof \nand tech): \nFood-borne \npathogens \nPHAC \nStaff (prof \nand tech): \nFood-",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the component cost tables for WGS and conventional methods from the same multi-country study, covering staff and other costs by laboratory.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-043",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 12D. Output from Included Studies for the Scoping Review: Cost per Sample, Alleweldt et al. (2021)",
      "pageStart": 802,
      "pageEnd": 802,
      "target": "../tables/table-12d.csv",
      "narrative": "draft",
      "content": "Staff (professional and technical) component costs reported by Alleweldt et al. (2021) for whole-genome sequencing (WGS) and conventional methods, for eight laboratory settings covering avian influenza, influenza A+B and food-borne pathogens (2021 USD).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/table-044",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table 12E. Output from Included Studies for the Scoping Review: Cost per Sample, Alleweldt et al. (2021)",
      "pageStart": 802,
      "pageEnd": 802,
      "target": "../tables/table-12e.csv",
      "narrative": "draft",
      "content": "Other component costs reported by Alleweldt et al. (2021) for whole-genome sequencing (WGS) and conventional methods, for eight laboratory settings covering avian influenza, influenza A+B and food-borne pathogens (2021 USD).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-vfig-p803",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 803,
      "pageEnd": 803,
      "target": "../figures/vfig-p803.png",
      "narrative": "draft",
      "content": "A stacked bar chart of a modified CHEERS quality assessment for ten studies, each with 17 criteria; the blue segment is the number answered yes and the green segment the remainder. Criteria met range from 7 (Cirillo D.M. et al 2016) to 15 (Pankhurst LJ et al 2016). Brackets below the x-axis group the studies into three sets.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-803",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 803",
      "pageStart": 803,
      "pageEnd": 803,
      "target": "../sections/page-803.md",
      "narrative": null,
      "content": "798 \n \nFigure 3. Quality of Evidence for Systematic Review: Modified CHEERS \n \n \n \n \n \nFigure 4. Quality of Evidence for Scoping Review: Modified CHEERS \n0\n2\n4\n6\n8\n10\n12\n14\n16\n18\nChan W.S. et al\n(2020)\nTafess K. et al\n(2020)\nGliddon HD et\nal. (2021)\nColman R.E. et\nal (2019)\nCates L. et al\n(2022)\nRowlinson MC\nand Musser KA\n(2022)\nCirillo D.M. et al\n(2016)\nPankhurst LJ et\nal (2016)\nHe G. et al\n(2020)\nVogel M. et al\n(2021)\nNumber of criteria with \"yes (blue)\" vs \"no (grey)\" response\ntNGS \ntNGS & WGS \nWGS",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Bar charts showing modified CHEERS quality ratings for the studies in the systematic review.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-804",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 804",
      "pageStart": 804,
      "pageEnd": 804,
      "target": "../sections/page-804.md",
      "narrative": null,
      "content": "799 \n \n \n \n0\n5\n10\n15\n20\n25\nChaturbhuj\net al (2014)\nDimitrov et\nal (2017)\nDudley et al\n(2012)\nEkici et al\n(2014)\nGachogo et\nal (2020)\nKunasol et\nal (2022)\nMerel et al\n(2001)\nTaylor et al\n(2013)\nWeinstein\net al (2001)\nAlleweldt et\nal (2021)\nPatel et al\n(2016)\nRiaz et al\n(2021)\nWales et al\n(2017)\nNumber of criteria with \"yes (blue)\" vs \"no (grey)\" vs \"N/A (orange)\" \nresponse\ntNGS \nWGS",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A bar chart showing modified CHEERS quality ratings for the studies in the scoping review.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-805",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 805",
      "pageStart": 805,
      "pageEnd": 805,
      "target": "../sections/page-805.md",
      "narrative": null,
      "content": "7. REFERENCES \n1) Koh WJ, Ko Y, Kim CK, Park KS, and Lee NY. Rapid diagnosis of tuberculosis and \nmultidrug resistance using a MGIT 960 system. Ann Lab Med. 2012;32:264–9. \n2) Zaheen A, and Bloom BR. Tuberculosis in 2020 - New Approaches to a Continuing Global \nHealth Crisis. N Engl J Med. 2020 Apr 2;382(14):e26. doi: 10.1056/NEJMp2000325. \nPMID: 32242354. \n3) Barbier M, Wirth T. The Evolutionary History, Demography, and Spread of the \nMycobacterium tuberculosis Complex. Microbiol Spectr. 2016 Aug;4(4). doi: \n10.1128/microbiolspec.TBTB2-0008-2016. PMID: 27726798.World Health Organization \n(WHO",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The start of the reference list for the review of sequencing costs for drug-resistant TB, citing work on TB diagnostics, epidemiology and WHO reports.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-806",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 806",
      "pageStart": 806,
      "pageEnd": 806,
      "target": "../sections/page-806.md",
      "narrative": null,
      "content": "16) Cirillo DM, Miotto P, and Tortoli E. Evolution of Phenotypic and Molecular Drug \nSusceptibility Testing. Adv Exp Med Biol. 2017;1019:221-246. doi: 10.1007/978-3-319-\n64371-7_12. PMID: 29116638. \n17) Hepple, et al. Implementation of liquid culture for tuberculosis diagnosis in a remote \nsetting: lessons learned. Int J Tuberc Lung Dis. 2011;15(3):405–407. \n18) World Health Organization. Molecular line probe assays for rapid screening of patients at \nrisk of multidrug-resistant tuberculosis (MDR-TB) [Internet]. Geneva: World Health \nOrganization; 2008 [cited 2020 Jul 14]. https://www.who.int/",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of the reference list, citing sources on phenotypic and molecular drug susceptibility testing, line probe assays and sequencing in clinical microbiology.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-807",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 807",
      "pageStart": 807,
      "pageEnd": 807,
      "target": "../sections/page-807.md",
      "narrative": null,
      "content": "31) Hemelaar J, Elangovan R, Yun J, Dickson-Tetteh L, Fleminger I, Kirtley S, Williams B, \nGouws-Williams E, Ghys PD; WHO–UNAIDS Network for HIV Isolation \nCharacterisation. Global and regional molecular epidemiology of HIV-1, 1990-2015: a \nsystematic review, global survey, and trend analysis. Lancet Infect Dis. 2019 \nFeb;19(2):143-155. doi: 10.1016/S1473-3099(18)30647-9. Epub 2018 Nov 30. Erratum in: \nLancet Infect Dis. 2020 Mar;20(3):e27. PMID: 30509777. \n32) Bell LCK, Noursadeghi M. Pathogenesis of HIV-1 and Mycobacterium tuberculosis co-\ninfection. Nat Rev Microbiol. 2018 Feb;16(2):80-90.",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of the reference list, citing sources on HIV epidemiology, TB–HIV co-infection, costing methods and the included sequencing studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-808",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 808",
      "pageStart": 808,
      "pageEnd": 808,
      "target": "../sections/page-808.md",
      "narrative": null,
      "content": "BMC Res Notes. 2020 Sep 18;13(1):444. doi: 10.1186/s13104-020-05287-9. PMID: \n32948225; PMCID: PMC7501614. \n43) Gliddon HD, Frampton D, Munsamy V, Heaney J, Pataillot-Meakin T, Nastouli E, Pym \nAS, Steyn AJC, Pillay D, and McKendry RA. A Rapid Drug Resistance Genotyping \nWorkflow for Mycobacterium tuberculosis, Using Targeted Isothermal Amplification and \nNanopore Sequencing. Microbiol Spectr. 2021 Dec 22;9(3):e0061021. doi: \n10.1128/Spectrum.00610-21. Epub 2021 Nov 24. PMID: 34817282; PMCID: \nPMC8612157. \n44) He G, Li Y, Chen X, Chen J, and Zhang W. Prediction of treatment outcomes for \nmulti",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of the reference list, citing the included tNGS, WGS and HIV genotyping cost studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-809",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 809",
      "pageStart": 809,
      "pageEnd": 809,
      "target": "../sections/page-809.md",
      "narrative": null,
      "content": "2014 Feb 12;9(2):e87441. doi: 10.1371/journal.pone.0087441. PMID: 24533056; PMCID: \nPMC3922725. \n54) Dudley DM, Chin EN, Bimber BN, Sanabani SS, Tarosso LF, Costa PR, Sauer MM, Kallas \nEG, O'Connor DH. Low-cost ultra-wide genotyping using Roche/454 pyrosequencing for \nsurveillance of HIV drug resistance. PLoS One. 2012;7(5):e36494. doi: \n10.1371/journal.pone.0036494. Epub 2012 May 4. PMID: 22574170; PMCID: \nPMC3344889. \n55) Ekici H, Rao SD, Sönnerborg A, Ramprasad VL, Gupta R, Neogi U. Cost-efficient HIV-1 \ndrug resistance surveillance using multiplexed high-throughput amplicon sequencing: \nim",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The end of the reference list, citing HIV, malaria and hepatitis C sequencing studies and a cost-effectiveness analysis of WGS for TB.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-810",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 810",
      "pageStart": 810,
      "pageEnd": 810,
      "target": "../sections/page-810.md",
      "narrative": null,
      "content": "8. APPENDIX \n8.1 Results of Economic Evidence around using WGS for the Diagnosis of DR-\nTB \nSeven manuscripts looked at cost data for WGS (Tables 2A-B). Two manuscripts were from the \nUS, one in countries across Europe (UK, Ireland, Germany and France) as well as Canada, one in \nItaly, one in the Kyrgyz Republic, and one in China (Tables 2A-B). The year of cost valuation \nranged from 2014-2021 across studies (Tables 2A-B). Six manuscripts assessed diagnostic cost per \nsample for WGS, and two assessed this for conventional pDST (Tables 2A-B). Four of the \nmanuscripts considered component costs",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "An appendix summarizing economic evidence on WGS for diagnosing drug-resistant TB from seven studies, covering settings, cost years and cost measures. It then begins the systematic review search strategy, listing PubMed terms for drug-resistant TB.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-811",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 811",
      "pageStart": 811,
      "pageEnd": 811,
      "target": "../sections/page-811.md",
      "narrative": null,
      "content": "TB\"[tw] OR \"resistant pulmonary tuberculos*\"[tw] OR \"resistant TB\"[tw] OR \"resistant \ntuberculos*\"[tw] \n \nConcept #2 \n \n\"Costs and Cost Analysis\"[mesh] OR \"Costs and Cost Analys*\"[tw] OR \"Cost \nAnalys*\"[tw] OR \"Cost Benefit Analys*\"[tw] OR \"Cost Effectiveness\"[tw] OR \n\"economic*\"[tw] OR \"cost allocation\"[tw] OR \"cost efficiency analys*\"[tw] \n \nConcept #3 \n \n\"Sequence Analysis\"[mesh] OR \"Sequence Analys*\"[tw] OR \"Sequence \nDetermination*\"[tw] OR \"next-generation sequencing\"[tw] OR \" next-gen sequence \nanalys*\"[tw] OR \"next-gen sequencing\"[tw] OR \"genetic sequencing\"[tw] OR \"genomic \nsequencing\"",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the systematic review search strategy with PubMed cost and sequencing terms and the start of the Embase search.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-812",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 812",
      "pageStart": 812,
      "pageEnd": 812,
      "target": "../sections/page-812.md",
      "narrative": null,
      "content": "'cost'/exp OR 'cost effectiveness analysis'/exp OR 'cost benefit analysis'/exp OR ('cost \nanalys*' OR 'cost benefit' OR 'cost-benefit analys*' OR 'cost effectiveness' OR 'cost \nefficiency analys*' OR 'cost allocation'):ab,ti,kw \n \nConcept #3 \n \n'sequence analysis'/exp OR 'high throughput sequencing'/exp OR ('high through-put \nsequencing' OR 'Sequence Analys*' OR 'Sequence Determination*' OR 'next-generation \nsequencing' OR ' genetic sequencing' OR 'genomic sequencing'):ab,ti,kw \n \n#1 AND #2 AND #3 – 26 records \n \nScopus search \n \nConcept #1 \n \n\"multidrug resistant tuberculosis\" OR \"drug resist",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the search strategy with Embase cost and sequencing terms and the Scopus search terms for drug-resistant TB and costs.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-813",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 813",
      "pageStart": 813,
      "pageEnd": 813,
      "target": "../sections/page-813.md",
      "narrative": null,
      "content": "\"sequence analys* \" OR \"high throughput sequencing\" OR \"Sequence Determination*\" OR \n\"next-generation sequencing\" OR \"genetic sequencing\" OR \"genomic sequencing\" \n#1 AND #2 AND #3 – 56 records \n \n8.3. Scoping Review Search Strategy \n \nPubMed search \n \nConcept #1A \n \n\"Hepatitis\"[Mesh:NoExp] OR \"Hepatitis, Viral, Human\"[Mesh] OR \"hepatitis\"[tiab] \n \nConcept #1B \n \n\"HIV\"[mesh] OR \"HIV\"[tw] OR \"Human Immunodeficiency Virus*\"[tw] OR \"AIDS \nVirus*\"[tw] OR \"Acquired Immune Deficiency Syndrome Virus\"[tw] OR \"Acquired \nImmunodeficiency Syndrome Virus\"[tw] \n \nConcept #1C \n \n(\"Malaria\"[mesh] OR \"Malaria\"",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Ends the systematic review search strategy and begins the scoping review PubMed search, with terms for hepatitis, HIV, malaria and influenza.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-814",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 814",
      "pageStart": 814,
      "pageEnd": 814,
      "target": "../sections/page-814.md",
      "narrative": null,
      "content": "\"Costs and Cost Analysis\"[mesh] OR \"Costs and Cost Analys*\"[tw] OR \"Cost \nAnalys*\"[tw] OR \"Cost Benefit Analys*\"[tw] OR \"Cost Effectiveness\"[tw] OR \n\"economic*\"[tw] OR \"cost allocation\"[tw] OR \"cost efficiency analys*\"[tw] \n \nConcept #3 \n \n\"Sequence Analysis\"[mesh] OR \"Sequence Analys*\"[tw] OR \"Sequence \nDetermination*\"[tw] OR \"next-generation sequencing\"[tw] OR \" next-gen sequence \nanalys*\"[tw] OR \"next-gen sequencing\"[tw] OR \"genetic sequencing\"[tw] OR \"genomic \nsequencing\"[tw] \n \n#1A AND #2 AND #3 – 63 records \n#1B AND #2 AND #3 – 62 records \n#1C AND #2 AND #3 – 29 records \n#1D AND #2 AND #",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the scoping review search strategy with PubMed cost and sequencing terms, record counts and the start of the Embase search.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-815",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 815",
      "pageStart": 815,
      "pageEnd": 815,
      "target": "../sections/page-815.md",
      "narrative": null,
      "content": "'Malaria'/exp OR ('Malaria' OR 'Plasmodium Infection*' OR 'Remittent Fever' OR 'Marsh \nFever' OR 'malarial fever' OR 'malarial infection*' OR 'marsh fever' OR 'paludism' OR \n'Plasmodia infection*' OR 'plasmodial infection*' OR 'plasmodiosis'):ab,ti,kw \n \nConcept #1D \n \n'influenza'/exp OR ('Influenza*' OR 'Human Flu' OR 'Grippe'):ab,ti,kw \n \nConcept #2 \n \n'cost'/exp OR 'cost effectiveness analysis'/exp OR 'cost benefit analysis'/exp OR ('cost \nanalys*' OR 'cost benefit' OR 'cost-benefit analys*' OR 'cost effectiveness' OR 'cost \nefficiency analys*' OR 'cost allocation'):ab,ti,kw \n \nConcept #3",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the scoping review search strategy with Embase terms and record counts, then begins the Scopus search.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-816",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 816",
      "pageStart": 816,
      "pageEnd": 816,
      "target": "../sections/page-816.md",
      "narrative": null,
      "content": "\"Human immunodeficiency virus* \" OR \"HIV\" OR \"AIDS Virus* \" OR \"Acquired \nImmune Deficiency Syndrome Virus\" OR \"Acquired Immunodeficiency Syndrome Virus\" \n \nConcept #1C \n \n\"Malaria\" OR \"Plasmodium Infection*\" OR \"Remittent Fever\" OR \"Marsh Fever\" OR \n\"malarial fever\" OR \"malarial infection*\" OR \"marsh fever\" OR \"paludism \" OR \n\"Plasmodia infection*\" OR \"plasmodial infection*\" OR \"plasmodiosis\" \n \nConcept #1D \n \nInfluenza OR \"Human Flu\" OR \"Grippe\" \n \nConcept #2 \n \n\"cost\" OR \"cost effectiveness analys*\" OR \"cost analys*\" OR \"cost benefit\" OR \"cost-\nbenefit analys*\" OR \"cost effectiveness\" OR \"c",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the scoping review Scopus search strategy, with record counts for each disease.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-817",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 817",
      "pageStart": 817,
      "pageEnd": 817,
      "target": "../sections/page-817.md",
      "narrative": null,
      "content": "_(no text on this page)_",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A blank page with no extracted text.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-818",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 818",
      "pageStart": 818,
      "pageEnd": 818,
      "target": "../sections/page-818.md",
      "narrative": null,
      "content": "813 \n \nSupplementary Table 1A. Quality of Health Economic Studies Instrument Assessment of Included tNGS Studies for the \nSystematic Review \n \nManuscript \nWas the study population \nclearly described? \nWere competing \nalternatives clearly \ndescribed? \nWas a well-defined \nresearch question \nposed in answerable \nform?  \nWas the actual \nperspective chosen \nappropriate? \nWere all important and \nrelevant costs for each \nalternative identified? \nWere all costs measured \nappropriately in physical units? \nWere costs valued \nappropriately? \nWere all \noutcomes \nmeasured \nappropriately in \nphysical units?",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A supplementary table giving yes/no answers on each quality criterion for the included tNGS studies, using the health economic study quality instrument.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-819",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 819",
      "pageStart": 819,
      "pageEnd": 819,
      "target": "../sections/page-819.md",
      "narrative": null,
      "content": "814 \n \nRowlinson MC and Musser KA \n(2022)1 \nNo \nYes \nNo \nYes \nNo \nNo \nYes \nNo \nYes \n1. The paper by Rowlinson MC and Musser KA (2020) is a report. \n \nSupplementary Table 2A. Quality of Health Economic Studies Instrument Assessment of Included WGS Studies for the \nSystematic Review \n \nManuscript \nWas the study population \nclearly described? \nWere competing \nalternatives clearly \ndescribed? \nWas a well-defined research \nquestion posed in \nanswerable form?  \nWas the actual \nperspective chosen \nappropriate? \nWere all important and \nrelevant costs for each \nalternative identified? \nWere all costs m",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A supplementary table of quality criteria answers for the included WGS studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-820",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 820",
      "pageStart": 820,
      "pageEnd": 820,
      "target": "../sections/page-820.md",
      "narrative": null,
      "content": "815 \n \nRowlinson MC and Musser \nKA (2022)1 \nNo \nYes \nNo \nYes \nNo \nNo \nYes \nNo \nYes \n1. The paper by Rowlinson MC and Musser KA (2020) is a report \n \nSupplementary Table 3A. Quality of Health Economic Studies Instrument Assessment of Included Studies for the Scoping \nReview \n \nManuscript \nWas the study \npopulation \nclearly \ndescribed? \nWere competing \nalternatives \nclearly \ndescribed? \nWas a well-\ndefined research \nquestion posed \nin answerable \nform?  \nWas the actual \nperspective \nchosen \nappropriate? \nWere all \nimportant and \nrelevant costs for \neach alternative \nidentified? \nWere all costs",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A supplementary table of quality criteria answers for the studies in the scoping review.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-821",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 821",
      "pageStart": 821,
      "pageEnd": 821,
      "target": "../sections/page-821.md",
      "narrative": null,
      "content": "816 \n \nSupplementary Table 3B. Quality of Health Economic Studies Instrument Assessment of Included Studies for the Scoping \nReview \n \n \nManuscript \nWere the \nsources of all \ncosting \nincluded? \nDid the authors \ndescribe methods for \nconverting costs into \na common currency \nbase and the \nexchange rate? \nDid the \nauthors state \nthe time \nhorizon(s) \nover which \ncosts were \nbeing \nevaluated and \nsay why \nappropriate?  \nDid the study \ndiscuss the \ngeneralizability \nof the results to \nother settings \nand \npatient/client \ngroups? \nDid the article \nindicate that there \nwas no potential \nconflict of",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Supplementary Table 3B rates the included studies of a scoping review against a quality instrument for health economic studies, covering items such as cost sources, currency conversion, time horizon, comparators and discounting. Most studies met only some criteria, with one meeting all.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-822",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 822",
      "pageStart": 822,
      "pageEnd": 822,
      "target": "../sections/page-822.md",
      "narrative": null,
      "content": "9. ACKNOWLEDGEMENTS \n \nThere are a number of individuals whom we wish to thank for their support in completing this work. \nFirstly, we would like to thank Donna Hesson for her help with designing our search strategy and for \nbeing an altogether great resource. Donna is one of the Public Health Informationists at Johns Hopkins \nUniversity, Bloomberg School of Public Health. Secondly, we want to thank several members of FIND, \nfor providing their technical expertise, as well as for making themselves available to consult whenever \nwe had questions. Thirdly, we want to thank the authors from sever",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Acknowledges the people who helped with the review on targeted next-generation sequencing costs, including an information specialist at Johns Hopkins, FIND staff and authors of included studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-823",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 823",
      "pageStart": 823,
      "pageEnd": 823,
      "target": "../sections/page-823.md",
      "narrative": null,
      "content": "Web Annex B.18. Stakeholder experiences and perspectives of \nusing Targeted next generation sequencing for diagnosing MDR-\nTB: interviews results \n \n \n(Draft 1, 15 December 2022. Final draft 17 April 2023) \n \n \n \n \n \nAuthor: \nDr Natalie Leon \nIndependent public health research consultant \nnatalieheleneleon@gmail.com \n \n \n \nA project for:  \nUnit: TB Prevention, Diagnosis, Treatment, Care & Innovation (PCI)",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title page of Web Annex B.18, a 2023 report on stakeholder experiences of using targeted next-generation sequencing (tNGS) to diagnose MDR-TB, naming the author and the WHO unit it was prepared for.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-824",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 824",
      "pageStart": 824,
      "pageEnd": 824,
      "target": "../sections/page-824.md",
      "narrative": null,
      "content": "Abstract \n \n \nBACKGROUND \nTimely drug resistance detection is essential to global tuberculosis management. Novel molecular \ntesting tools are being developed to rapidly detect multi-drug resistant Tuberculosis (MDR-TB) which is \nneeded for more effective and timely treatment of MDR-TB. Targeted next generation sequencing \n(tNGS), is new molecular-based rapid diagnostic technology (used on a sputum sample), that can \nprovide comprehensive diagnosis of MDR-TB in a short turnaround time of a few days, compared to the \nstandard culture-based phenotypic MDR-TB diagnostics that take multiple weeks.",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The abstract explains that tNGS on sputum can diagnose drug-resistant TB within days instead of weeks, but faces implementation barriers. It states that interviews with implementers aimed to inform a new WHO guideline by exploring acceptability, feasibility, values, preferences and equity. It begins reporting that implementers saw the technology as a major advance but found platforms complex, with few chances to catch errors early.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-825",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 825",
      "pageStart": 825,
      "pageEnd": 825,
      "target": "../sections/page-825.md",
      "narrative": null,
      "content": "3. Specialized laboratory infrastructure and human resources are required which are \npotentially challenging: As tNGS is a molecular-based testing platform, the platform requires \nhighly specialised laboratory infrastructure that includes multiple rooms to prevent \ncontamination and specialized cold storage facilities. Highly specialized molecular/medical \nscientists are needed to perform the tests. In these LMIC settings, such specialized  laboratory \ninfrastructure and staff may only be available at centralized laboratories and not necessarily at \nregional laboratories.  \n4. Specialist requi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the abstract's findings on tNGS feasibility, including specialized laboratories and staff found only in central facilities and needs for stable power, internet, computing, clean water and temperature control. It also notes supply chain delays and data management concerns. Turning to values and equity, participants worried about the cost of equipment, reagents and maintenance and called for comprehensive costing.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-826",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 826",
      "pageStart": 826,
      "pageEnd": 826,
      "target": "../sections/page-826.md",
      "narrative": null,
      "content": "consumables, extra general laboratory consumables, and the additional infrastructure needs \n(such as the extra space, temperature control, and internet connectivity needs). Cost-\neffectiveness calculations should also be comprehensive, and should include assessment of the \nimpact of the use of tNGS testing on improving TB outcomes.  \n3. Synergies should be found to optimize the use of tNGS in diagnostic algorithms: There was \nsome concern that more and more rapid tests are adopted, without consideration for how to \ndevelop synergies  between the use of various tests, to optimize use of the new",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Finishes the abstract's findings, calling for full cost-effectiveness analysis, synergies within diagnostic algorithms and stronger laboratory and health systems, including clinician training. It concludes that adoption in low- and middle-income countries should weigh ease of use, organizational context and stakeholder values as well as effectiveness.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-827",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 827",
      "pageStart": 827,
      "pageEnd": 827,
      "target": "../sections/page-827.md",
      "narrative": null,
      "content": "Background \n \nTuberculosis (TB) remains a major global public health threat. One of the most challenging forms of the \ndisease is multidrug resistant TB (MDR-TB), due to higher morbidity and mortality, complexity of \ntreatment and higher cost (1). WHO estimated that in 2016, close to half a million people were \ndiagnosed with MDR-TB (2). It is estimated that only two out of every three patients with MDR-TB are \ndiagnosed, of those diagnosed, only three out of four are treated, and only half of those treated are \ncured. This results in the majority (75%) of patients with MDR-TB not being cured;",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives background on MDR-TB, noting that only about two in three cases are diagnosed and most patients are not cured, partly because of diagnostic and treatment delays. It explains that culture-based drug susceptibility testing takes weeks and that the End TB Strategy calls for universal testing. It then describes the tNGS workflow, stressing the precise library preparation needed before sequencing.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p828-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 828,
      "pageEnd": 828,
      "target": "../images/img-p828-1.webp",
      "narrative": "draft",
      "content": "A linear workflow diagram of targeted next-generation sequencing (tNGS) with four steps joined by arrows: DNA extraction, library preparation, sequencing and data analysis.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-828",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 828",
      "pageStart": 828,
      "pageEnd": 828,
      "target": "../sections/page-828.md",
      "narrative": null,
      "content": "different commercially available models. Computational resources are needed for data analysis, report \ngeneration and data storage (5).  \nFigure 1: tNGS workflow \n \nTargeted tNGS therefore hold the potential, especially in high-burden TB settings, for effective \ntreatment and control of the spread of MDR-TB, but there are obstacles to the uptake of novel, rapid \ntechnologies. Obstacles to similar rapid diagnostics have been documented and include concerns \nabout costs, integration into existing laboratory workflows, technical training and skill requirements for \nutilization of the technology,",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the description of the tNGS workflow and notes known barriers to rapid diagnostics, such as cost, workflow integration, training and interpreting sequence data. It explains why health worker perspectives on implementability matter. It restates the study's objective of informing WHO guidance on tNGS.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-829",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 829",
      "pageStart": 829,
      "pageEnd": 829,
      "target": "../sections/page-829.md",
      "narrative": null,
      "content": "Methods \n \nStudy design \nThis was a cross-sectional observational qualitative study using semi-structured interviews. \n \nSetting \nThe main stakeholders were implementers of the FIND trial in the main reference laboratories that \nimplemented the tNGS trial, located in Georgia, India, and South Africa.   \n \nParticipants  \nWe were interested to learn about how the experiences and perceptions of the stakeholders who were \ninvolved with implementing the FIND trial in the three countries. Participant selection was therefore \npurposive and representative; purposive as the aim was to include in-countr",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the qualitative methods, using semi-structured interviews with implementers of the FIND trial in reference laboratories in Georgia, India and South Africa, plus coordinators and other global sites. It explains purposive selection of staff with hands-on experience. It also describes recording, verbatim quotes and interview summaries as the main data source.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-830",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 830",
      "pageStart": 830,
      "pageEnd": 830,
      "target": "../sections/page-830.md",
      "narrative": null,
      "content": "check it for accuracy. Most welcomed this opportunity and provided minor edits to improve the \nsummary. The edited versions were used as the final version of the interview summary.  \n \nThe interviewer read and reread the summaries, to become familiar with the overall data set. Using a \nthematic analysis approach, key issues were identified within and across interview data sets, and this \nwas synthesized into themes and sub-themes. Key implementation areas identified related to views on \nacceptability and feasibility, including challenges with setting up and running the technology, ideas \nabout",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains that participants checked interview summaries, which were analysed thematically under acceptability, feasibility, and values, preferences and equity. It notes confidentiality agreements, so countries are anonymized as A, B and C. Findings begin with strong acceptability because tNGS is comprehensive, uses sputum and is fast.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-831",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 831",
      "pageStart": 831,
      "pageEnd": 831,
      "target": "../sections/page-831.md",
      "narrative": null,
      "content": "to culture samples), and the rapidness (quick results compared to phenotypic testing times; 3-5 days as \ncompared to 4-6 weeks). There was also the sense that the technology is arriving at the right time \ngiven that resistance to newer TB drugs is likely to increase as use if these drugs become routine.  \n \nThe tNGS technology represents major advancement in TB diagnostics.  The view from all participants \nwas that tNGS technology represents a major advancement in TB diagnostics. The ability of the \ntechnology to provide a more comprehensive diagnosis of TB drug resistance was considered a maj",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that all participants saw tNGS as a major advance, with results in 3–5 days rather than 4–6 weeks and testing for more drugs, including new and repurposed ones. Quotes describe it as revolutionary while noting it is still evolving. Participants felt now is the time to build expertise before resistance to newer drugs rises.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-832",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 832",
      "pageStart": 832,
      "pageEnd": 832,
      "target": "../sections/page-832.md",
      "narrative": null,
      "content": "the test because the resistance is low. But in another 4-5 years we still need this technology. So, it’s \ntime to build on this technology, now, so that we are experts at it when it happens, when \nresistances start coming.” (Country C, participant 1).  \n \nFeasibility considerations of tNGS technology \n \nThe overall sentiment was that the tNGS technology needs to be further developed before it can be \nconsidered  fully ready for operational use. Implementers experienced a range of challenges with \nimplementing the tNGS technology which influenced their views on the feasibility of the technology",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports the overall view that tNGS needs further development before routine use. It lists challenges such as start-up problems, specialized staff and laboratories, supply chains, complex library preparation, data management and security, and the need to keep the WHO mutation catalogue updated. It begins describing how skilled molecular scientists were available only at central laboratories.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-833",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 833",
      "pageStart": 833,
      "pageEnd": 833,
      "target": "../sections/page-833.md",
      "narrative": null,
      "content": "included laboratory technicians. The scientists had experience in molecular diagnostics, were highly \ntrained (some had PHDs), and some had advance experience in testing other rapid molecular TB \ndiagnostics. Molecular scientists were only available at reference/central laboratories, and not at \nregional laboratories, which means that in these settings this test could not be decentralized to lower-\nlevel laboratories. \nThe tNGS technology was thought to be technically challenging even for specialist staff with molecular \ntesting experience, making it difficult to consider the use of tNGS in th",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains that tNGS required highly trained molecular scientists found only in reference laboratories, so it could not be decentralized. Quotes say routine technicians could not run the test in its current form and that automation would help. It also describes requirements for high-capacity computers and internet access for data analysis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-834",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 834",
      "pageStart": 834,
      "pageEnd": 834,
      "target": "../sections/page-834.md",
      "narrative": null,
      "content": "platforms generate large data files which rely on high capacity, stable internet connectivity.  \n“To implement this technology, computing power, storage infrastructure and internet \nconnectivity are extremely critical…The two platforms we tested are reliant on internet \nconnectivity. And we’re talking about uploading massive sizes of files of a server for analysis. If \na country doesn’t have the infrastructure to support internet and uploading info into a cloud for \nanalysis, you can forget it. I don’t know if they are considering having a localized machine to do \nthat analysis.) (Country A, p",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the reliance of tNGS platforms on stable, high-capacity internet for uploading large data files. It reports supply chain disruptions, with specialist kits ordered from overseas causing delays. It also describes the demanding precision of sample and library preparation and the need to avoid cross-contamination.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-835",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 835",
      "pageStart": 835,
      "pageEnd": 835,
      "target": "../sections/page-835.md",
      "narrative": null,
      "content": "the correct volume is added”, which is “demanding of precision” and requires “the utmost \nconcentration”, and with” no room for error” (Country A, participant 1). \nSome of the problems implementers experiences were contamination of control samples, recognition, \nstruggles with uploading data from the sequencer to a web-based data analysis tool, problems with \ndata analysis software, problem with computer storage capacity, problems with sensitivity of \nequipment to temperature and disturbances. In one setting there was multiple failed runs where the \nsource of the problem proved difficult to pi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes problems implementers faced, such as contaminated controls, data upload and software problems, temperature sensitivity and failed runs eventually traced to water quality. Participants called for simplifying the many library preparation steps so that less specialized staff could perform them. It notes that the trial stopped at producing reports, so clinicians' use of results was not evaluated.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-836",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 836",
      "pageStart": 836,
      "pageEnd": 836,
      "target": "../sections/page-836.md",
      "narrative": null,
      "content": "the utility of the tNGS reports for clinical decision-making.  \nThe process of producing a diagnostic report with test results is automated for all the three platforms. \nReports differed in the level of detail they provided, and reports were presented in a format that could \nbe used for clinical decision-making. For each drug, information on resistance is expressed in terms of \nwhether there is resistance or susceptibility to the drug, and the confidence level of that test result \n(high, medium, or low). Identification of mutations are based on the WHO reference catalogue of \nmutations. Data f",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes tNGS reports, which give resistance or susceptibility for each drug with a confidence level based on the WHO mutation catalogue, plus lineage data useful for surveillance. Participants noted uncertainty about resistance to newer drugs such as bedaquiline. They called the WHO catalogue essential and wanted automatic links between tNGS software and laboratory information systems.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-837",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 837",
      "pageStart": 837,
      "pageEnd": 837,
      "target": "../sections/page-837.md",
      "narrative": null,
      "content": "Differences on feasibility views across different tNGS platforms \n \nThe overall sentiment that is that all three the tNGS platforms need to be further developed before it is \nfully ready for operational use. However, there were differences in the sentiments about feasibility \nchallenges, with some being regarded as having more substantial challenges. The high level of \ntechnical complexity of the sample preparation stages (mainly the library preparation stage) was \nconsidered a key challenge for the Deeplex platform. Sample preparation prior to sequencing was \nconsidered less of a challenge fo",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Compares feasibility views across platforms, with library preparation seen as more complex for Deeplex than for Oxford Nanopore. Both were thought to lack early error detection, risking wasted runs. Quotes describe the products as immature and not yet market-ready.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-838",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 838",
      "pageStart": 838,
      "pageEnd": 838,
      "target": "../sections/page-838.md",
      "narrative": null,
      "content": "Areas of concern that would need to be addressed in future development of the tNGS platform include \nthe need to reduce the level of complexity of the sequencing procedures and develop more automated \ninterlinkage with the data analysis processes. Mechanisms would need to be in place to ensure a \ncontinuous supply of specialist products, especially the reagents. Local vendors are needed for the \nmaintenance of equipment and to provide technical support.  \n \nA lower complexity tNGS platform is needed to decentralize the use of tNGS technology: A key \nconcern was the high level of complexity of",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists areas for future tNGS development, such as reducing complexity, automating links to data analysis, securing supplies and local technical support. Participants wanted a lower-complexity platform, for example with robotic liquid handling, so testing could be decentralized. Some hoped for point-of-care tests close to patients.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-839",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 839",
      "pageStart": 839,
      "pageEnd": 839,
      "target": "../sections/page-839.md",
      "narrative": null,
      "content": "“If the technology can become like a ‘plug and play’, it can be placed closer to the point of care” \n(Country C, participant 1). \n \nEquitable access \n \nCentralized, reference laboratory placement requires tradeoffs in terms of equitable access:  \nStakeholders agreed given the complexity of tNGS technology (and the specialist laboratory and HR \nrequirements), it is only suitable for placement at centralized, reference laboratories that can perform \nmolecular testing. This has equity considerations as it may mean less access for some regions of the \ncountry without reference laboratorys. The MDR",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Discusses equitable access, noting that placing tNGS only in reference laboratories may disadvantage some regions and that batching in low-caseload settings causes delays. A participant called for 'plug and play' single-sample testing at affordable cost. It begins listing the costs that comprehensive costing should include, such as computing and storage.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-840",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 840",
      "pageStart": 840,
      "pageEnd": 840,
      "target": "../sections/page-840.md",
      "narrative": null,
      "content": "should also include equipment maintenance and servicing costs. There may also additional costs \nassociated with securing high internet capacity and interlinking the tNGS information platform and the \nlaboratory’s own information system.  \n \nCost effectiveness should include test effectiveness and impact on TB treatment outcomes: While \nthere is appreciation for dramatic improvements in scope and turnaround time of diagnosis of MDR \nTB, the real value should be measured in terms of the accuracy of the test and the impact on TB \ntreatment outcome. Test results should be accurate and valid, and t",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Adds that costing should include maintenance, internet capacity and links to laboratory information systems. Participants argued cost-effectiveness should reflect accuracy, indeterminate results and impact on treatment outcomes, accepting short delays for reliable results. A participant hoped tNGS would replace rather than add to existing tests.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-841",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 841",
      "pageStart": 841,
      "pageEnd": 841,
      "target": "../sections/page-841.md",
      "narrative": null,
      "content": "Rational adoption of new rapid tests such as tNGS should also include considerations for if and how it \nwould fit with a transition from phenotypic testing towards genotypic testing; how phenotypic testing \nmay still be needed in future scenarios. And costing of the use of tNGS should include the costs of the \nscenario where both genotypic and phenotypic testing for MDR TB would still be required.  \n \nA similar point was made by another participant who pointed to the need to be clear about the \npurpose of the rapid diagnostic in the overall cascade and the need for it to be easily implemented:",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Discusses how tNGS fits a shift from phenotypic to genotypic testing and the cost of needing both. A participant called for a simple molecular technology to manage most patients quickly, leaving complex cases for further testing. Others stressed that rapid diagnostics only pay off if laboratories and health systems become more efficient.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-842",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 842",
      "pageStart": 842,
      "pageEnd": 842,
      "target": "../sections/page-842.md",
      "narrative": null,
      "content": "from rapid diagnostics, one participant explained the importance of  “reengineering” improvements in \nthe laboratory and the broader health care system: \n“A colleague of mine who was involved with the LPA evaluation said that we have to \nreengineer the lab to do these tests. But if we don’t reengineer the health care system, the \nvalue of the time that’s gained in the lab is lost in the health care system. I think that’s really \nquite important. You can have   test that can give you an answer in 2 days…You ship that test \nresult to the clinic. Then the clinic has to find the patient. And the p",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Quotes a global expert who argued that time gained in the laboratory is lost if the health system cannot find and treat patients. It calls for better sample referral, communication between clinicians and laboratories and clinician training on using results. A summary table of findings begins with acceptability.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-843",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 843",
      "pageStart": 843,
      "pageEnd": 843,
      "target": "../sections/page-843.md",
      "narrative": null,
      "content": "rapidness, as well as the timeliness of the \ntest. \n \nFeasibility \nThe overall sentiment was  that the tNGS \ntechnology needs to be further developed \nbefore it can be considered  fully ready for \noperational use. \n1. Setting up the new technology \n2. High technical complexity of the test \n3. Specialized laboratory infrastructure and \nhuman resources \n4. Specialist requirements for operating the \ntest \n5. Supply chain challenges \n6. Data management and storage \nrequirements \n7. Continuous updating of the WHO \nmutations reference library \n8. There are different feasibility concerns \nfor the dif",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the summary of findings, listing feasibility challenges and values, preferences and equity considerations such as placement, affordability, algorithm synergies and health system strengthening. It begins the discussion, noting the positive views alongside concerns about centralization, cost and equitable access.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-844",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 844",
      "pageStart": 844,
      "pageEnd": 844,
      "target": "../sections/page-844.md",
      "narrative": null,
      "content": "rapid MDR-TB diagnostics. The wish is for affordable, point of care rapid MDR-TB diagnostics, that is \nmore automated and simplified (‘plug and play’), to allow for decentralized laboratory use, closer to \nthe point of diagnosing patients. There was a sense of caution about the need to rationalize and \nintegrate new rapid diagnostics. There is also the concern that diagnostic gains from rapid tNGS \ntechnology, can only be optimized if there is investment in reengineering and strengthening laboratory \nsystems and the broader health system.  \n \nAs noted earlier, participants acknowledged that wh",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Discusses participants' wish for affordable, automated point-of-care MDR-TB diagnostics and caution about integrating new tests. It notes that some feasibility challenges, such as supply chains and support, may persist and that the findings echo the 2018 WHO guidance on tNGS. It calls for evaluations of impact on treatment outcomes and mixed-methods research including providers and patients.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-845",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 845",
      "pageStart": 845,
      "pageEnd": 845,
      "target": "../sections/page-845.md",
      "narrative": null,
      "content": "comprehensive view of how to optimise diagnostic gains from tNGS. Policy and practice initiatives may \nbe required to strengthen laboratory and health system support functions, to allow for rational, \nintegrated, and equitable use of tNGS technology.  \n \nStrength and limitations \nThe primary qualitative study responds to the gap in qualitative research on tNGS technology. To our \nknowledge, this is the first study on the experience and perceptions of implementers of tNGS for MDR-\nTB diagnostics. The findings from this study is timely as it can be used to guide further development of \nthe tNGS",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Closes the discussion of a primary qualitative study on implementing targeted next-generation sequencing (tNGS) for MDR-TB diagnosis, calling for policy support to strengthen laboratory and health system functions. It sets out strengths, such as being the first study of tNGS implementers and interviewing them soon after the FIND trial, and limitations, such as limited depth across several groups and uncertainty over which problems stem from the technology being new. It concludes that tNGS is seen as a major advance but needs further development, and introduces considerations for manufacturers and decision-makers.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-846",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 846",
      "pageStart": 846,
      "pageEnd": 846,
      "target": "../sections/page-846.md",
      "narrative": null,
      "content": "1. Planning, preparation, and implementation support will be needed to implement tNGS in LMIC \nsettings. \na. Ensure the procurement of equipment is in place, including the required in country \ncustom clearances.  \nb. Training programmes should allow sufficient opportunity for local staff to do multiple \nsequencing runs by themselves, under supervision of the trainers.  \nc. Include a checklist of challenges and errors to anticipate in the set-up and running of \nthe sequencing platform, and instructions for trouble shooting. \nd. Include a checklist of all specialist consumables and the quantitie",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists considerations for implementing tNGS in low- and middle-income countries, including advance planning for equipment procurement and customs clearance, hands-on training, troubleshooting and consumables checklists and local vendor support. It also calls for specialist molecular staff and for platforms to be made field-ready with simpler sample preparation.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-847",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 847",
      "pageStart": 847,
      "pageEnd": 847,
      "target": "../sections/page-847.md",
      "narrative": null,
      "content": "automation and standardization to increase efficiency and effectiveness of \nimplementation. \nc. There needs to be more of a seamless linkage between the sequencing technology and \nthe data analysis technology of the tNGS platform.  \nd. Data analysis and data storage requirements need to be fully supported, including \nsystems for backing up data.  \ne. Data ownership and data security considerations will need to be put in place. \nf. Mechanisms are needed for interlinking data systems, to allow for two-way \ncommunication between the information system of the tNGS platform and the \ninformation sys",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the considerations, calling for more automation, seamless links between sequencing and data analysis, data storage, ownership and security arrangements and interoperability with laboratory information systems. It also stresses affordability, sustainable funding and full costing, and concludes that guidelines should consider ease of use, organizational context and stakeholder values as well as effectiveness.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-848",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 848",
      "pageStart": 848,
      "pageEnd": 848,
      "target": "../sections/page-848.md",
      "narrative": null,
      "content": "technology in diagnosing MDR-TB, but also the influences  such as the ease of use of the technology, as  \norganizational contexts and needs, and the values, preferences, and aspirations of stakeholders. \n \nContributions of authors  \n \nDr Natalie Leon is the lead author of the protocol and the guarantor of the review.  \n \nAcknowledgements \n \nThe following individuals and groups are acknowledged for their contribution and support of this work. \nParticipants in the interview study are thanked for their generosity in sharing their insights and their \ntime. Experts are thanked for responding to the",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Credits the study's lead author, thanks interview participants, experts and WHO staff, and states no conflicts of interest, before beginning the reference list. The references include work on drug-resistant TB and on health workers and policy implementation in South Africa.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-849",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 849",
      "pageStart": 849,
      "pageEnd": 849,
      "target": "../sections/page-849.md",
      "narrative": null,
      "content": "14. \nErasmus E, Gilson L. How to start thinking about investigating power in the organizational \nsettings of policy implementation. Health policy and planning. 2008;23: 361-68. \n15. \nYoussef A, Chaudhary ZK, Wiljer D, Mylopoulos M, Sockalingam S. Mapping Evidence of \nPatients' Experiences in Integrated Care: A Scoping Review. Gen Hosp Psychiatry. 2019;61:1-9. \n \n \nAppendix 1: Interview guide for primary qualitative study \n \nQUALITATIVE RESEARCH ON EXPERIECES AND PERCEPTIONS WHO TNGS RAPID DIAGNOSTICS \nINTERVIEW GUIDE (Draft, V3). 9 Sept 2022 \nLABORATORY STAFF AND MANAGEMENT IN TRIAL STUDY SITE",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Ends the reference list and presents the interview guide for laboratory staff and managers at trial sites, with questions on their roles, experience with molecular and sequencing technologies and preparatory work for tNGS. It then moves on to their experiences across sample and library preparation.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-850",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 850",
      "pageStart": 850,
      "pageEnd": 850,
      "target": "../sections/page-850.md",
      "narrative": null,
      "content": "2) sequencing 3) data analysis and reporting, what has the experience been like for you?  \na) What has worked well? And why? \nb) What has not worked well? And why? \n \n9. For each of the tNGS solutions? What was the experience like for you? \na. What has worked well? And why? \nb. What has not worked well? And why? \n \n10. How would you summarise what the requirements were for implementing each of the tNGS \nsolutions?  \na. What the challenges were \nb. And how have you team responded to those challenges. \n \n11. How does your experience of tNGS with each of the three solutions compare with other DR-",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the interview guide with questions on what worked well or poorly in sequencing and data analysis, how tNGS compares with other drug-resistance tests and what enabling factors would help it improve TB care. It closes with contact details for the researcher.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-851",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 851",
      "pageStart": 851,
      "pageEnd": 851,
      "target": "../sections/page-851.md",
      "narrative": null,
      "content": "Appendix 2: Information and consent form \n \nEXPERIENCES AND PERCEPTIONS OF TARGETTED NEXT GENERATION SEQUENCING (tNGS) \nRAPID DIAGNOSTICS FOR DETECTING TB DRUG RESISTANCE:  \nAN INTERVIEW STUDY TO INFORM THE WHO tNGS GUIDELINE \nInformation and consent form (V3, 2022) \n \nStudy background and rationale \nTargeted New Generation Sequencing (tNGS) is a novel, rapid molecular diagnostic test that provide an \nend-to-end solution for detecting resistance to multiple TB drugs at the same time, but little is known \nabout its implementation considerations. To inform the World Health Organization (WHO) gui",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents the information and consent form for the interview study on tNGS for detecting TB drug resistance, explaining its purpose of informing WHO guidance by exploring acceptability, feasibility, equity, values and preferences at FIND trial sites and elsewhere. It describes the independent researcher, procedures, confidentiality and the absence of expected benefits or harms.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-852",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 852",
      "pageStart": 852,
      "pageEnd": 852,
      "target": "../sections/page-852.md",
      "narrative": null,
      "content": "I agree for this interview to be recorded:  Please pick one: YES or NO \n \nI (name and surname) ………………………………………………., give my consent to \nparticipate in this interview.  \n \n \nSignature …………………………………….                                                     Date:       /       / \n2022 \n \nPaste signature/write initials, then save as PDF and email or print, sign and email \nscanned printed document to natalieheleneleon@gmail.com",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The signature section of the consent form, where participants agree to recording and give consent to be interviewed.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-853",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 853",
      "pageStart": 853,
      "pageEnd": 853,
      "target": "../sections/page-853.md",
      "narrative": null,
      "content": "Web Annex B.19. Stakeholder experiences and perspectives of \nusing Targeted next generation sequencing (tNGS) for diagnosing \nmulti-drug resistant Tuberculosis  (MDR-TB): a qualitative \nevidence synthesis \n \n \n(Draft, 15 December 2022, Final draft 17 April 2023) \n \n \n \n \n \nAuthor: \nDr Natalie Leon \nIndependent public health research consultant \nnatalieheleneleon@gmail.com \n \n \n \n \nA project for:  \nUnit: TB Prevention, Diagnosis, Treatment, Care & Innovation (PCI)",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title page of Web Annex B.19, a qualitative evidence synthesis by Natalie Leon on stakeholder experiences of using tNGS to diagnose MDR-TB, dated 2022–2023 and prepared for WHO's TB unit.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-854",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 854",
      "pageStart": 854,
      "pageEnd": 854,
      "target": "../sections/page-854.md",
      "narrative": null,
      "content": "Abstract \n \nBACKGROUND \nTimely drug resistance detection is essential to global tuberculosis management. Novel molecular \ntesting tools are being developed to rapidly detect multi-drug resistant Tuberculosis (MDR-TB) which is \nneeded for more effective and timely treatment of MDR-TB. Targeted next generation sequencing \n(tNGS), is new molecular-based rapid diagnostic technology (used on a sputum sample), that can \nprovide comprehensive diagnosis of MDR-TB in a short turnaround time of a few days, compared to the \nstandard culture-based phenotypic MDR-TB diagnostics that take multiple weeks. Th",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The abstract of the qualitative evidence synthesis, explaining that tNGS can give comprehensive drug-resistance results within days but faces technical, organizational and economic barriers. It describes a rapid review that searched Medline and other sources in 2022, with single-reviewer screening and planned thematic synthesis and CERQual assessment. It reports that no eligible studies were found and points to the companion primary study and a 2022 Cochrane review of other rapid TB tests as indirect evidence.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-855",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 855",
      "pageStart": 855,
      "pageEnd": 855,
      "target": "../sections/page-855.md",
      "narrative": null,
      "content": "inform the new WHO guideline on tNGS technology. Of interest, is that the Cochrane review also \nreports on obstacles to effective utilization of rapid TB diagnostics from the perspective of patients as \nrecipients of rapid diagnostics, and they recommend programmes to support patient health literacy. \nPatient recipients’ experiences and perspectives of rapid TB diagnostics is an understudied area that \ndeserves more attention.   \n \nCONCLUSION \nA rapid review of qualitative evidence on the experiences and perceptions of  tNGS technology for \nMDR-TB diagnostics was conducted to inform the develo",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Concludes the abstract by noting the Cochrane review's attention to patient obstacles and calling for more qualitative research alongside tNGS trials and roll-out. The background describes TB and drug resistance as major threats and explains how tNGS works on sputum samples to report resistance to multiple drugs. It notes that some platforms also report lineage information useful for tracking resistant strains.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-856",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 856",
      "pageStart": 856,
      "pageEnd": 856,
      "target": "../sections/page-856.md",
      "narrative": null,
      "content": "resistance TB.  tNGS diagnostic test results would enable timely clinical decisions on appropriate \ntreatment for MDR-TB, which can improve TB treatment outcomes and stop the transmission of \nresistant strains of TB (5). \nDifferent tNGS platforms have different laboratory workflows, but the  general tNGS workflow \nincludes: preparation of the TB sputum sample, DNA extraction from the sputum sample and DNA \npurification, library preparation, sequencing and data analysis, as shown in the workflow Diagram 1 \nbelow (5). DNA extraction procedures require specialized laboratory safety levels (biosaf",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the general tNGS workflow from sample preparation and DNA extraction under biosafety level 3 to library preparation, sequencing and data analysis, stressing its complexity. It notes the lack of evidence on implementation experiences. It explains that a separate primary qualitative study was carried out alongside the FIND trial in India, Georgia and South Africa.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-857",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 857",
      "pageStart": 857,
      "pageEnd": 857,
      "target": "../sections/page-857.md",
      "narrative": null,
      "content": "Objective \nThis study seeks to provide evidence to inform the new WHO guideline for use of tNGS technology for \nMDR-TB diagnostics. The objective is to identify and synthesize qualitative evidence on \nimplementation considerations of tNGS technology for MDR-TB, related to acceptability, feasibility, \nvalues and preferences and equity, from the perspectives of health workers as implementers.  \n \nMethods  \nOverall study design \nThis was a rapid review of qualitative evidence, guided by the rapid review recommendations for \nstreamlining methods used for systematic reviews (16). Rapid review techn",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "States the review objective of synthesising qualitative evidence on acceptability, feasibility, values, preferences and equity of tNGS from health workers' perspectives. It describes the rapid review design and the inclusion and exclusion criteria, focusing on implementers but open to other stakeholders.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p858-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 858,
      "pageEnd": 858,
      "target": "../images/img-p858-1.webp",
      "narrative": "draft",
      "content": "A PRISMA flow diagram of search and screening with two parallel streams. The systematic search found 3 records, all reviewed in full text and excluded for wrong method, leaving 0 eligible records. The open, hand and expert search found 27 records, excluded for wrong method or being out of scope, leaving 0 eligible records.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-858",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 858",
      "pageStart": 858,
      "pageEnd": 858,
      "target": "../sections/page-858.md",
      "narrative": null,
      "content": "Search methods for identification of studies \nElectronic searches  \nWe used a combination of a systematic search strategy and an open searching approach. The search \nstrategy used a set of search terms associated with NGS technology as an intervention, with MDR-TB \nand qualitative research (and associated names). We searched Medline with no year or language \nlimits.  We designed a search strategy and ran on Aug 19, 2022, and  rerun on Oct 10, to include WGS \nrelated studies for MDR-TB.  (See Appendix 1 for the search strategy terms).  We complemented this \nsearch  with open searching on Google",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the search methods, combining a Medline search run in August and October 2022 with open searching, hand searching and contacting experts, and single-author screening with WHO consultation. It reports that no eligible studies were found, mostly because papers were quantitative, discussion pieces or based on hypothetical views.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-859",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 859",
      "pageStart": 859,
      "pageEnd": 859,
      "target": "../sections/page-859.md",
      "narrative": null,
      "content": "Discussion and recommendations \n \nThe aim of this study was to identify and synthesize qualitative evidence on implementation \nconsiderations of the new tNGS technology for diagnosing multi-drug resistance TB, to inform the new \nWHO guideline on rapid TB diagnostics. This review did not yield any direct evidence from primary \nqualitative perception studies on tNGS technology as no studies could be found.  \n \nResearch and practice implications \nGiven this is an empty review, there are clearly research gaps waiting to be filled. While there are \nseveral studies testing the accuracy of tNGS techn",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Discusses the empty review, noting that tNGS is not yet widely operational and that future effectiveness studies offer chances to study implementers' experiences. It recommends including qualitative process evaluations and draws on a Cochrane review of other rapid TB diagnostics that highlights implementation factors. It also notes that patients' experiences of rapid TB diagnostics are understudied.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-860",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 860",
      "pageStart": 860,
      "pageEnd": 860,
      "target": "../sections/page-860.md",
      "narrative": null,
      "content": "Strength and limitations  \nThe QES found no eligible studies most likely due the newness of the tNGS technology. We anticipated \nfinding very little information, and a pilot search prior to embarking on the review may have confirmed \nthe lack of information. The value of having conducted the review and coming up empty, it that it \nconfirms there is gap in qualitative studies on implementation of TNGS for MDR-TB. This knowledge is \nuseful to encourage researchers to include qualitative and or mixed method process evaluations \nalongside future operational research on tNGS technology.  \nThe autho",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Discusses strengths and limitations, noting that the empty result confirms a research gap and that the review used multiple search methods but rapid-review shortcuts such as single-reviewer screening. It closes with acknowledgements and a statement of no conflicts of interest.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-861",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 861",
      "pageStart": 861,
      "pageEnd": 861,
      "target": "../sections/page-861.md",
      "narrative": null,
      "content": "References \n \n1. Tiberi S, Utjesanovic N, Galvin J, et. al. Drug resistant TB – latest developments in \nepidemiology, diagnostics and management. International Journal of Infectious Diseases. 2022. \n2. World Health Organization. Global Tuberculosis Report 2022. Geneva: World Health \nOrganization; 2022. \n3. Farhat M, Fresch L, Calderon R, Loerger T, Snyder M, Meehan C, et al. GWAS for quantitative \nresistance phenotypes in Mycobacterium tuberculosis reveals resistance genes and regulatory \nregions. Nature Communications. 2019;10(2128). \n4. World Health Organization. Global Tuberculosis Report 2",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The reference list for the qualitative evidence synthesis, citing WHO TB reports and guidelines, a WHO technical guide on next-generation sequencing for TB, studies on resistance genetics and guidance on rapid reviews.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-862",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 862",
      "pageStart": 862,
      "pageEnd": 862,
      "target": "../sections/page-862.md",
      "narrative": null,
      "content": "Appendix 1: Search strategy for QES \n \nDatabase: Ovid MEDLINE(R) ALL <1946 to August 19, 2022> \nSearch Strategy: \n-------------------------------------------------------------------------------- \n1     \"targeted next generation sequencing\".mp. (2890) \n2     \"targeted NGS\".mp. (782) \n3     1 or 2 (3284) \n4     Interviews as Topic/ or interview*.mp. or Interview/ (458448) \n5     survey*.mp. or Health Surveys/ or Health Care Surveys/ or \"Surveys and Questionnaires\"/ \n(1192334) \n6     Qualitative Research/ (76141) \n7     Focus group discussion*.mp. or Focus Groups/ (42217) \n8     \"mixed methods\".t",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives the Medline search strategy used for the qualitative evidence synthesis, combining terms for targeted sequencing, qualitative and survey methods and tuberculosis, which yielded three records.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-863",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 863",
      "pageStart": 863,
      "pageEnd": 863,
      "target": "../sections/page-863.md",
      "narrative": null,
      "content": "Web Annex B20. Accuracy of Mycobacterium tuberculosis antigen-\nbased skin tests: a systematic review and meta-analysis \n \nReview team: Yohhei Hamada1, Elena Surkova2, Irina Kontsevaya3, Ting Ting Wang1, \nVictoria Liu4, Liliya Eugenevna Ziganshina5,6, Molebogeng X Rangaka1,7 \n \n1. Institute for Global Health, University College London, UK \n2. Royal Brompton Hospital. Part of Guy's and St Thomas' NHS Foundation Trust London, UK \n3. Research Center Borstel, Germany \n4. London School of Hygiene and Tropical Medicine, UK \n5. Institute of Fundamental Medicine and Biology of Kazan Federal University,",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title page of Web Annex B.20, a systematic review and meta-analysis of the accuracy of Mycobacterium tuberculosis antigen-based skin tests, listing the review team and advisers and noting it was prepared for a February 2022 WHO guideline meeting.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-864",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 864",
      "pageStart": 864,
      "pageEnd": 864,
      "target": "../sections/page-864.md",
      "narrative": null,
      "content": "Executive summary \n \nBackground \nDiagnosis and treatment of TB infection (TBI) is a pillar of the WHO End TB strategy to achieve a global \nreduction in TB incidence by 2035.  Accurate diagnostic tests are needed to better identify individuals who \nwould benefit from preventive treatment. If found to have at least equivalent diagnostic accuracy \ncompared to existing screening tests, the interferon gamma release assays (IGRA) and purified protein \nderivative (PPD) tuberculin skin test (TST), the new Mycobacterium tuberculosis (MTB) specific skin-based \ntests have the potential to improve access",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The executive summary of the skin test review, explaining that new TB antigen-based skin tests could improve access to TB infection testing if as accurate as IGRA and the tuberculin skin test. It describes updating an earlier review with searches of English, Russian and Chinese databases to July 2021, random-effects meta-analyses, a post-hoc pooled analysis requested by WHO for a class-based recommendation, and notes that no longitudinal studies were found.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-865",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 865",
      "pageStart": 865,
      "pageEnd": 865,
      "target": "../sections/page-865.md",
      "narrative": null,
      "content": "Three C-Tb studies were conducted in South Africa, while two were in Spain and UK, respectively. 33/34 \nstudies for Diaskintest were done in Russian Federation and the remaining one in Ukraine. All C-TST studies \nwere done in China.  \n \nStudy quality varied, with low concerns about applicability for all studies but high risk of bias in Diaskintest \nstudies, largely due to the study design (using data collected under routine clinical practice rather than \ndesigned as diagnostic accuracy studies) and a lack of clarity on the population selection criteria.  \n \nSensitivity \nTwo studies assessed te",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that C-Tb studies came from South Africa, Spain and the UK, Diaskintest studies almost all from Russia and C-TST studies from China, and that Diaskintest studies had high risk of bias. It summarises sensitivity results showing Diaskintest, the tuberculin skin test and IGRAs performed similarly in adults with active TB, C-Tb had similar sensitivity to comparators, and agreement with IGRA was generally higher than with the skin test.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-866",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 866",
      "pageStart": 866,
      "pageEnd": 866,
      "target": "../sections/page-866.md",
      "narrative": null,
      "content": "Diaskintest with IGRA was 94.62% (95% CI 90.49–97.02; I² = 56.2%) in five studies in participants with any \ntuberculosis status. By contrast, the agreement between Diaskintest and TST5 mm showed considerable \nheterogeneity; the pooled agreement was estimated in children with active tuberculosis (97.39% [96.39–\n98.12]) and children without TB (17.62% [6.79-38.60]). \n \nConsidering all studies without restricting to 3-way head-to-head studies, the pooled agreement of C-Tb \nwith TST was similar, 81% (95% CI, 76-85%) at TST5mm in HIV-infected and 76% (95% CI, 71-81%) at TST15mm \nin HIV-uninfected T",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports further agreement results, such as about 95% agreement of Diaskintest with IGRA and highly variable agreement with the tuberculin skin test in children, and agreement of C-Tb with IGRA and the skin test in various groups. It notes limitations such as few responses from authors, and concludes that the new skin tests perform broadly like IGRA.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-867",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 867",
      "pageStart": 867,
      "pageEnd": 867,
      "target": "../sections/page-867.md",
      "narrative": null,
      "content": "was seen in Diaskintest performance probably because of the observational nature of the studies affecting \nthe quality of studies and the impact of BCG vaccination. Data quality was higher for C-Tb and C-TST \nstudies. Although the review suggests TBST may enable precise and accessible TBI testing, more research is \nneeded to address fully diagnostic accuracy and predictive performance in different at-risk populations in \npost-licensure studies.  \n \nBackground \n \nApproximately 25-27% of the world’s population is estimated to have TB infection1 2 with a lifetime risk of \nprogression to active di",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the conclusion, noting variable Diaskintest results likely due to observational designs and BCG vaccination and better data for C-Tb and C-TST, with more research needed. The background explains that about a quarter of the world's population has TB infection with a 5–10% lifetime risk of disease. It argues that current infection tests are imperfect and that new skin tests may perform well in children and people with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p868-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 868,
      "pageEnd": 868,
      "target": "../images/img-p868-1.webp",
      "narrative": "draft",
      "content": "An inverted pyramid showing the hierarchy of reference standards for tests of TB infection, with an arrow running from Weaker at the bottom to Stronger at the top. From top to bottom the levels are: efficacy of TPT based on results of test for TB infection; predictive value for TB disease; correlation with exposure gradient; sensitivity/specificity using surrogate reference standards; and concordance with existing tests.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-vfig-p868",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 868,
      "pageEnd": 868,
      "target": "../figures/vfig-p868.png",
      "narrative": "draft",
      "content": "A PICO table. Population: PLHIV, children under 5 years, household and other close contacts, and other at-risk groups such as immunocompromised people. Intervention: novel M. tuberculosis antigen-based skin tests (Diaskintest, C-Tb, C-TST, DPPD, others). Comparator: TST or IGRA. Outcomes: efficacy of TB preventive treatment based on test results, and predictive value for progression to TB disease (the table continues beyond the image).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-868",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 868",
      "pageStart": 868,
      "pageEnd": 868,
      "target": "../sections/page-868.md",
      "narrative": null,
      "content": "We previously conducted a systematic review to synthesise current evidence on the diagnostic \nperformance of TBST compared to that of currently available in vitro IGRA tests and TST.17 The review \nsuggested that those tests perform similarly to TST or IGRA.  We have updated the review to inform the \ndevelopment of WHO guidelines.  \n \nHierarchy of reference standards \nThe study of the diagnostic performance of tests for TB infection is hampered by a lack of an adequate \nreference standard. Existing tests for TB infection measure the cell-mediated immune response (memory T \ncell response) to exp",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes that the earlier review found new skin tests performed similarly to the tuberculin skin test and IGRA, and explains the hierarchy of reference standards used because no gold standard exists for TB infection. It sets out the review aim and begins a table of populations, tests, comparators and outcomes.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-869",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 869",
      "pageStart": 869,
      "pageEnd": 869,
      "target": "../sections/page-869.md",
      "narrative": null,
      "content": "mellitus, steroid use, \nsmoker);  \no High risk of prior TB \nexposure (prisoners, health \nworkers, immigrants from \nhigh TB burden countries, \nindividuals with CXR \nabnormalities; homeless \npeople and people who use \ndrugs, inhabitants of high TB \nburden settings)1 \n− BCG vaccinated vs non-vaccinated (in \nidentified groups at risk of TB infection- \nstratified or in combination, as \nappropriate) \n3. Correlation \nwith exposure \ngradient; \n− 4.Sensitivity/Sp\necificity for TB \ninfecton2;  \n− 5. Concordance \nwith TST3;  \n− 6.Concordance \nwith IGRA;  \n− 7.Proportion \nstarted on TPT.  \n−  \n \n \n1 > 100",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the table of review questions, covering at-risk populations, BCG status and outcomes such as concordance with other tests and proportion started on preventive treatment. It begins the methods, setting out inclusion and exclusion criteria for study designs.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-870",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 870",
      "pageStart": 870,
      "pageEnd": 870,
      "target": "../sections/page-870.md",
      "narrative": null,
      "content": "1 \nLongitudinal studies that report index \ntest result in the population eligible \nfor LTBI testing, preventive therapy \ngiven and cases of incident TB during \nthe study period \nMust be free of active disease at \nbaseline; Must report method of TB \ndiagnosis (microbiological or clinical) \n2 \nLongitudinal studies reporting \ndevelopment of incident TB in the \npopulation tested with index test \nduring the study period \nMust be free of active TB at baseline; \nMust report method of TB diagnosis \n(microbiological or clinical) \n3 \nStudies reporting index test result in \ncontacts of active TB cases \nM",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Tabulates inclusion criteria by objective, such as longitudinal studies of incident TB, contact studies by proximity, and studies of confirmed TB or low-risk groups for sensitivity and specificity. It describes the search, including Russian, Ukrainian and Chinese sources screened by native speakers or translation.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-871",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 871",
      "pageStart": 871,
      "pageEnd": 871,
      "target": "../sections/page-871.md",
      "narrative": null,
      "content": "reviewers independently.  \nDiscrepancies in inclusion/exclusion between the 2 reviewers were resolved by discussion between the 2 \nreviewers or if needed with additional reviewers. We used the systematic review management platform \nRayyan20 for study screening and tracking of exclusion reasons. Data extraction was carried out using \nspecific data extraction sheet in Microsoft Excel.  \n \nCase definitions:  \n \nIncident TB disease: any new case of TB (new or relapse) diagnosed subsequent to initial symptoms and \nsigns screening \nPrevalent TB: any case of known TB disease at the time of the diagno",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes resolving screening disagreements, using Rayyan and Excel for data management, and case definitions for incident, prevalent and active TB. It explains quality assessment with QUADAS-2, with an added domain on manufacturer involvement.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-872",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 872",
      "pageStart": 872,
      "pageEnd": 872,
      "target": "../sections/page-872.md",
      "narrative": null,
      "content": "The GRADE framework22 was used to systematically assess the quality of evidence and strength of \nrecommendations regarding the use of novel TBST. \n \nData Analysis: \n \nWhere possible, outcome measures were stratified by: type of test and population. Outcome and effect \nmeasures of interest were evaluated separately as described in Table 4.  \n \nSensitivity in those with microbiologically-confirmed active TB and specificity in those at low risk of TB \ninfection (restricted to studies from low TB burden countries) was calculated where possible. For \nspecificity, we also applied less restrictive cr",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains use of GRADE and the analysis plan, including sensitivity in confirmed TB, specificity in low-risk groups in low-burden countries, test agreement with exact confidence intervals and meta-analysis where two or more studies existed. It sets out less restrictive proxy measures of specificity, restricted to presumed healthy people.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-873",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 873",
      "pageStart": 873,
      "pageEnd": 873,
      "target": "../sections/page-873.md",
      "narrative": null,
      "content": "• \nOther diseases without TB symptoms and not suspected of having active TB and no reported \nhistory of contacts (this was assumed unless specified otherwise)   \n \nThe term specificity (proxy) is used when describing the estimates above. \nWe did not conduct a test for publication bias because none of the quantitative syntheses included \nsufficient numbers of studies (≥ 10 studies). \nOur primary analysis did not pool different TBST because of substantial heterogeneity in study designs. \nHowever, based on a request from WHO as the purpose of the review is for a class-based recommendation, \nas po",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains that proxy specificity also included people with other diseases and that publication bias was not tested because of too few studies. It notes that, at WHO's request, data across different skin tests were pooled post hoc despite heterogeneity and high risk of bias in Diaskintest studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-874",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 874",
      "pageStart": 874,
      "pageEnd": 874,
      "target": "../sections/page-874.md",
      "narrative": null,
      "content": "Table 4: Effect measures according to objective \nOutcome \nEffect measure \n1. Efficacy of preventive therapy based on \nthe test result  \n \nIncidence rates for disease progression \nstratified by test result  \nIncidence Rate Ratios \nNegative and Positive Predictive Values for \ndisease progression \n2. Predictive value of novel recombinant \nskin tests for incident TB among risk-\nstratified populations \n \nIncidence rates for disease progression in \nrisk-stratified populations \nIncidence Rate Ratios \nNegative and Positive Predictive Values for \nprogression with confidence intervals \n3. The associatio",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents Table 4, listing the effect measures for each review objective, such as incidence rate ratios and predictive values for progression, odds ratios by contact proximity, sensitivity and specificity, concordance and the proportion starting preventive treatment.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-875",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 875",
      "pageStart": 875,
      "pageEnd": 875,
      "target": "../sections/page-875.md",
      "narrative": null,
      "content": "Results \n5.1 Systematic literature review \nFigure 2 presents the selection process. Our previous review identified 37 studies reporting TBST: 29 \nDiaskintest, five C-Tb, two C-TST, one DPPD, that were published from inception until 20 October 2020.17    \n \nOur updated search covering the period until 30 July 2021 identified four reports (three in Russian23-25 on \nDiaskintest and one in English on C-TST26) via the database search. The report on C-TST reported the results \nof two studies. Additional two studies (one Russian27 and one Ukrainian28) on Diaskintest were identified \nvia other methods",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports the study selection, with the earlier review's 37 studies and new reports bringing totals to five C-Tb, 34 Diaskintest, four C-TST and one DPPD study, noting that DPPD data are shown but not pooled.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p876-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 876,
      "pageEnd": 876,
      "target": "../images/img-p876-1.webp",
      "narrative": "draft",
      "content": "A PRISMA flow diagram of study selection with two streams. Via databases and registers, records were identified from Medline (853), EMBASE (1317), CKNI (565), CBLD (282) and e-library (405); 746 duplicates were removed, 2676 records screened, 2651 excluded, 25 reports sought, 2 not retrieved, 23 assessed and 19 excluded, giving 7 studies from the updated search. Via other methods (WHO public call 20, manufacturers 15), 21 reports were assessed and 19 excluded. Combined with studies from the review by Krutikov et al, the review included 5 C-Tb, 4 C-TST, 34 Diaskintest and 1 DPPD studies.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-876",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 876",
      "pageStart": 876,
      "pageEnd": 876,
      "target": "../sections/page-876.md",
      "narrative": null,
      "content": "Figure 2 Study selection",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A page carrying only the caption of Figure 2 on study selection.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-877",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 877",
      "pageStart": 877,
      "pageEnd": 877,
      "target": "../sections/page-877.md",
      "narrative": null,
      "content": "5.2 Characteristics of included studies \n \nAll but one Diaskintest study was conducted in Russia (Table 5). All were cross-sectional assessments \nperformed as a part of a routine care provision, and cohorts were recruited prospectively or constructed \nretrospectively. None of the studies randomised different tests into different groups or arms. Four studies43 \n44 57 60 were head-to-head comparisons of Diaskintest with TST 5 mm cut-off (TST5mm) and IGRA in the same \nstudy. Four studies34 51 53 55 enrolled a total of 346 (4·9%) adults and 23 (0·3%) children with HIV, and 17 \nstudies recruited ch",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the included studies, noting that nearly all Diaskintest studies were routine-care cross-sectional studies in Russia, with few head-to-head comparisons and small numbers of people with HIV. It reports high BCG coverage where recorded and varied positivity thresholds. It also outlines the C-Tb, C-TST and DPPD studies, including a Brazilian DPPD study in BCG-vaccinated people.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-878",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 878",
      "pageStart": 878,
      "pageEnd": 878,
      "target": "../sections/page-878.md",
      "narrative": null,
      "content": "of preventive treatment, or the proportion of participants with positive results starting TB preventive \ntreatment. Table S3 in supplement summarises studies available for assessment of each review objective \n \nTable 7 shows the cut-off defined for each test by test manufacturers. It should be noted, however, that \nthey were not always followed, and the cut-off used in each study is indicated in the text and presented in \nsupplementary tables.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes that no studies assessed incident TB, effectiveness of preventive treatment or uptake of treatment, and that manufacturers' test cut-offs were not always followed.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-879",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 879",
      "pageStart": 879,
      "pageEnd": 879,
      "target": "../sections/page-879.md",
      "narrative": null,
      "content": "Table 5 Characteristics of Diaskintest studies \n \nStudy \nCountry \nIndex Test  \nComparators \nAge, years (a) \nSample size \n(in review) \nStudy population (b)  \nReview Objective Addressed \nTest \nconcordance \nSensitivity \nSpecificity \n(c ) \nDose-response \nassociation (d) \nAksenova 2011 \nRussia \nDiaskintestAI \nTST5mm \nNS \n1551(63) \nChildren; TB Screening  \nX \nX \n  \n  \nBaryshnikova \n2017 \nRussia \nDiaskintestAI \nTST5mm \nNS \n811(163) \nChildren; active PTB \nX \n  \n  \n  \nBaryshnikova \n2021 \nRussia \nDiaskintest7mm \nIGRA (TSPOT.TB) \n18-65 \n4756 (645) \nChildren under routine \nsurveillance in TB service \nX",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents Table 5, listing Diaskintest studies by country, threshold, comparator, age, sample size, population and review objectives addressed. Most studies were in Russia, with children and medical students among the populations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-880",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 880",
      "pageStart": 880,
      "pageEnd": 880,
      "target": "../sections/page-880.md",
      "narrative": null,
      "content": "Losovskaya \n2014* \nRussia \nDiaskintestAI \nTST5mm, IGRA (QFT-TB \nGIT) \n0.5-15 \n50 (46) \nChildren; TB investigations \nX \n  \n  \n  \nLosovskaya \n2016* \nRussia \nDiaskintestAI \nTST5mm, IGRA (QFT-TB \nGIT) \n3-6 \n63 (63) \nChildren; TB investigation \nX \n  \n  \n  \nMishin 2016 \nRussia \nDiaskintestAI \nNone \nNS \n529 (103) \nHIV- Adults; PTB; Healthy control;  \n  \nX \n  \n  \nNakonechnaya \n2020 \nRussia \nDiaskintestAI \nIGRA (QFT-TB GIT) \n1-17 \n62 (62) \nChildren with TB and other \npulmonary conditions \nX \n \n \n \nNikitina 2019 \nRussia \nDiaskintestAI \nIGRA (QFT-TB GIT) \nAdults: 42, 18-84 \nChildren: 10, 3-16 \n181 (68)",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues a table of study characteristics for Diaskintest evaluations, almost all from Russia, listing comparator tests, participant ages, sample sizes and populations. Most studies involve children or adults under investigation for tuberculosis, with some in people with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-881",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 881",
      "pageStart": 881,
      "pageEnd": 881,
      "target": "../sections/page-881.md",
      "narrative": null,
      "content": "Starshinova \n2018* \nRussia \nDiaskintest5mm \nTST5mm, IGRA (T.SPOT-\nTB, QFT-TB GIT) \nChildren 8.1, Adults: 37 \n860 (860) \nHIV- ; BCG-vaccinated; TB \nScreening; Children; Adults \n  \nX \n  \n  \nStarshinova \n2019a* \nRussia \nDiaskintest5mm \nTST5mm , IGRA (QFT-TB \nGIT, T-SPOT.TB) \n18-65 \n187 (135) \nAdults, Culture+ TB; TB \nunexposed; IGRA+/Diaskintest+ \n  \nX \n  \n  \nStarshinova \n2019b \nRussia \nDiaskintestAI \nNone \nTB hospital: 42 (0.23), \nGeneral hospital: 43 \n(0.27)  \n154 (154) \nHealthcare professionals in TB \nhospitals and general hospitals \n  \n  \n  \nX \nStogova 2020a \nRussia \nDiaskintestAI \nTST5mm \n48",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the Diaskintest study characteristics table with further Russian studies, including some in health workers and people with suspected tuberculosis. Footnotes flag studies used in the three-way comparison with IGRA and TST and explain the abbreviations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-882",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 882",
      "pageStart": 882,
      "pageEnd": 882,
      "target": "../sections/page-882.md",
      "narrative": null,
      "content": "Table 6 Characteristics of C-Tb, C-TST, and DPPD studies \n \nStudy \nCountry \nIndex Test  \nComparators \nAge, years (a) \nSample size (in \nreview) \nStudy population (b)  \nReview Objective Addressed \nTest \nconcordance \nSensitivity \nSpecificity \nDose-response \nassociation (c) \nAggerbeck 2013 \nUnited Kingdom \nC-Tb  \nTST (multiple \nthresholds), QFT \nCases: 33, 18-60, \nControls:34, 18-65 \n189 (189) \nActive Tb (3 participants selected \non the basis of positive IGRA); TB \nunexposed adults  \nX \n  \nX \n  \nAggerbeck 2018 \nSouth Africa \nC-Tb  \nTST5mm/15mm,  QFT \n17, 0-65 \n1190 (1190) \nChild case-contacts unde",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 6 lists characteristics of studies on the C-Tb, C-TST and DPPD skin tests, from countries such as the United Kingdom, South Africa, Spain and China. It gives comparators, ages, sample sizes, populations and which review objectives each study addressed, such as concordance, sensitivity and specificity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-883",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 883",
      "pageStart": 883,
      "pageEnd": 883,
      "target": "../sections/page-883.md",
      "narrative": null,
      "content": "Table 7 Cut off of TBST defined by their manufacturers \n \nTest \nSource \nCut off \nDiaskintest Package insert \nNegative response: The absence of \ninfiltration and hyperaemia or the \npresence of ‘prick response’ up to 2 mm \n \n \nAmbiguous response: The presence of \nhyperaemia without infiltrate. \n \nPositive response: The presence of \ninfiltrate (papule) of any size \nC-Tb \nThe most recent study by \nAggereck et al. (The test is not \nyet commercialized. No package \ninsert is available) \nInduration ≥ 5mm  \nC-TST \nPackage insert \nA positive result is interpreted \nby an average diameter of redness or \ni",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 7 sets out the positivity cut-offs for the Diaskintest, C-Tb, C-TST and DPPD skin tests as defined by manufacturers or the latest studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-884",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 884",
      "pageStart": 884,
      "pageEnd": 884,
      "target": "../sections/page-884.md",
      "narrative": null,
      "content": "5.3 Risk of bias in individual studies   \n \nThe quality of studies evaluating Diaskintest performance was difficult to assess due to \ninconsistencies and incomplete reporting of study methods and sample recruitment. This resulted in \nseveral “unknown” assessments against the quality criteria. Of the 16 studies evaluating the \nsensitivity of Diaskintest, risk of bias was high in 5 (31.3%) studies where test assessors were not \nblinded to TB culture results,34 46 49 51 53 and unclear in  at least one of the four risk of bias criteria in \n14 (87.5%) studies as information on patient selection or",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes risk of bias in the included studies, noting that poor reporting left many quality criteria unclear and that many Diaskintest studies lacked blinding to culture or comparator results. It then reports pooled results for Diaskintest, including sensitivity in people without HIV of 66% at any induration and 88% at a 5 mm cut-off.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p885-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 885,
      "pageEnd": 885,
      "target": "../images/img-p885-1.webp",
      "narrative": "draft",
      "content": "A forest plot of test sensitivity (%) in two head-to-head studies (Starshinova 2019 (a) and Starshinova 2018) for four tests, with random-effects pooled estimates: TST 5mm 88.24% (78.20–94.01), IGRA (QFT) 89.66% (78.83–95.28), IGRA (T-SPOT.TB) 90.91% (79.95–96.16) and Diaskintest 5mm 91.18% (81.72–95.98). Heterogeneity is I² = 0% for each test.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-885",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 885",
      "pageStart": 885,
      "pageEnd": 885,
      "target": "../sections/page-885.md",
      "narrative": null,
      "content": "22 \n \n \n \n \n \n \n \nFigure 3. Test sensitivity in three-way head-to-head studies comparing Diaskintest, IGRA and \nTST \n \nTP = True positive \nFN = False negative \n \nIncludes HIV-uninfected adults with microbiologically-confirmed active TB.  \n \n \nFigure 4 Test agreement in head-to-head Diaskintest studies comparing all three tests",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Captions for figures on test sensitivity and agreement in studies comparing Diaskintest, IGRA and TST directly against each other.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p886-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 886,
      "pageEnd": 886,
      "target": "../images/img-p886-1.webp",
      "narrative": "draft",
      "content": "A forest plot of percentage agreement between tests in two studies (Losovskaya 2014 and 2016), with random-effects pooled estimates: Diaskintest vs. IGRA 87.16% (79.47–92.24), Diaskintest vs. TST 5mm 55.45% (46.08–64.45) and IGRA vs. TST 5mm 51.38% (42.05–60.60). Agreement is much higher between Diaskintest and IGRA than between either test and TST.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-886",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 886",
      "pageStart": 886,
      "pageEnd": 886,
      "target": "../sections/page-886.md",
      "narrative": null,
      "content": "23 \n \n \nIncludes HIV-uninfected children under investigation for TB and those with active TB (clinical and \nconfirmed). AI = Any induration ; Conc = Concordant; N = Total, concordant + discordant; % \nAgreement: represents agreement with IGRA as the comparator  \n \nHighly variable methods and sub-populations precluded meaningful meta-analysis for most risk \ngroups; sensitivity estimates from individual studies ranged from 40%-71% in HIV-infected adults51 53 \nand from 92% to 100% in HIV-uninfected children,31 33 35 (Supplement Table S4-S12).  \nIn the study that stratified results by age,36 in chi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that varied methods and populations prevented meta-analysis for most Diaskintest risk groups, with sensitivity ranging from 40–71% in adults with HIV to 92–100% in children without HIV. It notes specificity could not be estimated for Diaskintest, then turns to C-Tb, where pooled specificity in low-burden settings was similar to IGRA and higher than TST at 15 mm.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-887",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 887",
      "pageStart": 887,
      "pageEnd": 887,
      "target": "../sections/page-887.md",
      "narrative": null,
      "content": "24 \n \n95% CI 90-95%); the analysis was not possible for TST5mm due to insufficient data.  \n \nThree studies provided suitable head-to-head data for agreement comparisons between C-Tb, IGRA, \nand TST. Pooled test agreement between C-Tb and IGRA was 79.80% (95% CI 76.10–83.07), similar to \nthat between IGRA and TST5 mm/15 mm (74.67% [64.01–83.01]) and C-Tb and TST5 mm/15 mm (78·92% [74·65–82·63]; \nfigure 7).16 29 30\n \n \nC-Tb results from studies that only compared two tests are shown in supplement section 4 (Tables \nS21-S26, Figure S2, Figure S3). These showed a pooled agreement of C-Tb with TST",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports pooled agreement between C-Tb, IGRA and TST in head-to-head studies, all around 75–80%, and similar agreement in two-test comparisons. It also gives agreement among people without tuberculosis and how often C-Tb was negative among IGRA-negative participants.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p888-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 888,
      "pageEnd": 888,
      "target": "../images/img-p888-1.webp",
      "narrative": "draft",
      "content": "A forest plot of test sensitivity (%) in four head-to-head studies (Ruhwald 2017, Aggerbeck 2018, Aggerbeck 2019, Hoff 2016), with random-effects pooled estimates: TST 15mm 77.18% (66.44–85.25), TST 5mm 82.68% (74.91–88.42), TST 5mm HIV+/15mm HIV− 78.18% (67.75–85.94), IGRA 71.67% (63.44–78.68) and C-Tb 74.52% (70.39–78.25). Heterogeneity is high for TST 15mm (I² = 81%), TST 5mm/15mm (84%) and IGRA (78%) and 0% for C-Tb.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-888",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 888",
      "pageStart": 888,
      "pageEnd": 888,
      "target": "../sections/page-888.md",
      "narrative": null,
      "content": "25 \n \nFigure 5. Test sensitivity in head-to-head studies comparing C-Tb, IGRA and TST \n \n \n \nTP = True positive; FN = False negative. Results include individuals with microbiologically-confirmed active TB.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Caption for a figure on the sensitivity of C-Tb, IGRA and TST compared directly in people with microbiologically confirmed tuberculosis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p889-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 889,
      "pageEnd": 889,
      "target": "../images/img-p889-1.webp",
      "narrative": "draft",
      "content": "A forest plot of test specificity (%) in two studies in low-incidence settings (Aggerbeck 2013 and Ruhwald 2017), with random-effects pooled estimates: TST 15mm 93.31% (90.22–95.48), IGRA 99.15% (79.66–99.97) and C-Tb 97.85% (93.96–99.25). The x-axis runs from 80 to 100.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-889",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 889",
      "pageStart": 889,
      "pageEnd": 889,
      "target": "../sections/page-889.md",
      "narrative": null,
      "content": "26 \n \nFigure 6. Test specificity in head-to-head studies comparing C-Tb, IGRA and TST  \n \nIndividuals without active TB in studies conducted in TB low-incidence settings. TN = True negative FP = False \npositive",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Caption for a figure on the specificity of C-Tb, IGRA and TST compared directly in people without active tuberculosis in low-incidence settings.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p890-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 890,
      "pageEnd": 890,
      "target": "../images/img-p890-1.webp",
      "narrative": "draft",
      "content": "A forest plot of percentage agreement between tests in three studies (Aggerbeck 2018, Aggerbeck 2019, Hoff 2016), with random-effects pooled estimates for seven comparisons: C-Tb vs. IGRA 79.80%, C-Tb vs. TST 5mm HIV+/15mm HIV− 78.92%, C-Tb vs. TST 5mm (HIV+) 83.85%, C-Tb vs. TST 15mm (HIV−) 76.39%, IGRA vs. TST 5mm HIV+/15mm HIV− 74.67%, IGRA vs. TST 5mm (HIV+) 79.54% and IGRA vs. TST 15mm (HIV−) 74.74%. Pooled agreement lies between about 75% and 84% in all comparisons.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-890",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 890",
      "pageStart": 890,
      "pageEnd": 890,
      "target": "../sections/page-890.md",
      "narrative": null,
      "content": "27 \n \nFigure 7 Test agreement in head-to-head C-Tb studies comparing all three tests \n \n \n \nIncludes individuals with bacteriologically-confirmed active TB.  Ruhwald 2017 and Aggerbeck 2013, although did \nthree- test comparisons, did not report data suitable for estimation of %TST-IGRA agreement. Conc = \nConcordant; N = Total, concordant + discordant; % Agreement: represents agreement with IGRA as the \ncomparator",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Caption for a figure on agreement among C-Tb, IGRA and TST in direct comparison studies, with notes on which studies lacked suitable data.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-891",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 891",
      "pageStart": 891,
      "pageEnd": 891,
      "target": "../sections/page-891.md",
      "narrative": null,
      "content": "28 \n \n  \n5.6 C-TST \n \nThree studies evaluated sensitivity of the C-TST.26 61 62 Sensitivity at the ≥5mm induration threshold \nranged from 77% to 90%, with a pooled estimate of 86% (95%CI: 83-89%) (Table S34). In one study, \nthe sensitivity of C-TST (90% [73-98%]) was similar to that of TPOT.TB (89% [78-95%] and TST 10 mm  \n(87% [76-94%] and slightly higher than TST 10 mm  (82% [69-90%].26 \nIn two studies including healthy individuals as well as TB patients and patients with other pulmonary \ndiseases,26 the pooled agreement between C-TST and IGRA was higher (85.96% [78.82-90.97%]) than \nthat be",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports C-TST results, with a pooled sensitivity of 86% at a 5 mm cut-off and higher agreement with IGRA than with TST, largely due to BCG vaccination effects. A post-hoc meta-analysis combining all novel skin tests then gives pooled sensitivity of 76% in people with HIV-negative or unknown status and lower estimates in people with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-892",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 892",
      "pageStart": 892,
      "pageEnd": 892,
      "target": "../sections/page-892.md",
      "narrative": null,
      "content": "29 \n \n73%, 5 studies) in HIV-positive individuals. \nWhile this may be seen as the lower sensitivity in HIV-positive individuals, these estimates are based \non different studies, and thus not conclusive. \n \n \nWhen combining all TBST, the pooled difference in specificity between TBST and IGRA was 2.29% \n(95%CI: -1.60-6.18%, 6 studies on Diaskintest, C-Tb and C-TST)  and that between TBST and TST was \n33.47% (95%CI: 18.16-48.78%, 14 studies including Diaskintest, C-Tb, and C-TST). The difference in \nagreement between TBST vs IGRA and TBST vs TST is consistent with 3-way head-to-head analysis of",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the combined analysis, finding little difference in specificity between the novel skin tests and IGRA but a much larger difference with TST, most likely due to BCG. Pooled agreement with IGRA was high, while agreement with TST was lower in people without tuberculosis, and the discussion begins on study quality concerns.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-893",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 893",
      "pageStart": 893,
      "pageEnd": 893,
      "target": "../sections/page-893.md",
      "narrative": null,
      "content": "30 \n \noften insufficient. Ascertainment of TB was inadequate; the diagnosis often pragmatically made on \nclinical and/or radiological findings rather than microbiologically-confirmed. Although Russian \nnational TB guidelines define Diaskintest positivity as induration of any size, more than a third of \nstudies used the 5mm cut-off, making comparison between studies and products difficult.  \nIncorporation bias is a risk in studies that selected study participants based on TST-positivity or had \nfollowed Russian national TB recommendations and used Diaskintest for TB diagnosis. There are also \nc",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Discusses limitations, including inconsistent procedures, poor reporting, reliance on clinical diagnosis, varied Diaskintest cut-offs, incorporation bias and possible conflicts of interest from industry funding. It concludes that the novel skin tests appear comparable to existing tests and could allow accessible screening without laboratories or blood draws.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-894",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 894",
      "pageStart": 894,
      "pageEnd": 894,
      "target": "../sections/page-894.md",
      "narrative": null,
      "content": "31 \n \n \nReferences \n \n \n1. Hay SI, Abajobir AA, Abate KH, et al. Global, regional, and national disability-adjusted life-years \n(DALYs) for 333 diseases and injuries and healthy life expectancy (HALE) for 195 countries \nand territories, 1990&#x2013;2016: a systematic analysis for the Global Burden of Disease \nStudy 2016. The Lancet 2017;390(10100):1260-344. doi: 10.1016/S0140-6736(17)32130-X \n2. Houben RMGJ, Dodd PJ. The Global Burden of Latent Tuberculosis Infection: A Re-estimation \nUsing Mathematical Modelling. PLOS Medicine 2016;13(10):e1002152. doi: \n10.1371/journal.pmed.1002152 \n3. Comst",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The start of the reference list for the systematic review on novel tuberculosis skin tests, citing global burden studies and early work on Diaskintest.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-895",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 895",
      "pageStart": 895,
      "pageEnd": 895,
      "target": "../sections/page-895.md",
      "narrative": null,
      "content": "32 \n \n15. Ruhwald M, Aggerbeck H, Gallardo RV, et al. Safety and efficacy of the C-Tb skin test to diagnose \nMycobacterium tuberculosis infection, compared with an interferon gamma release assay \nand the tuberculin skin test: a phase 3, double-blind, randomised, controlled trial. The \nLancet Respiratory Medicine 2017;5(4):259-68. \n16. Hoff ST, Peter JG, Theron G, et al. Sensitivity of C-Tb: a novel RD-1-specific skin test for the \ndiagnosis of tuberculosis infection. European Respiratory Journal 2016;47(3):919-28. \n17. Krutikov M, Faust L, Nikolayevskyy V, et al. The diagnostic performance of",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list, citing trials of C-Tb, a meta-analysis of novel skin tests and several Russian and Ukrainian studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-896",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 896",
      "pageStart": 896,
      "pageEnd": 896,
      "target": "../sections/page-896.md",
      "narrative": null,
      "content": "33 \n \nTUBERCULOSIS MEDICAL CENTERS (Ukrainian). Ukrainian Journal of Occupational Health \n2015;1(42) \n29. Aggerbeck H, Ruhwald M, Hoff ST, et al. C-Tb skin test to diagnose Mycobacterium tuberculosis \ninfection in children and HIV-infected adults: A phase 3 trial. PLoS ONE [Electronic Resource] \n2018;13(9):e0204554. \n30. Aggerbeck H, Ruhwald M, Hoff ST, et al. Interaction between C-Tb and PPD given concomitantly \nin a split-body randomised controlled trial. International Journal of Tuberculosis & Lung \nDisease 2019;23(1):38-44. \n31. Aksenova V, Barishnikova L, Klevno N, et al. New screening an",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list with C-Tb trials and Russian studies of Diaskintest in children and adolescents.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-897",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 897",
      "pageStart": 897,
      "pageEnd": 897,
      "target": "../sections/page-897.md",
      "narrative": null,
      "content": "34 \n \n44. Lozovskaya ME, Belushkov VV, Gurina OP, Vasilyeva YB, Klochkova LV. Comparative evaluation of \ninnovative diagnostic tests for latent and active TВ infection in children. Pediatr 2014; 5: 46–\n50.  \n45. Ministry of Health, and Social Development of the Russian Federation. Order of the Ministry of \nHealth and Social Development of the Russian Federation of October 29, 2009 N 855 ‘On \nAmendments to Appendix N 4 to the Order of the Ministry of Health of Russia of March 21, \n2003 N 109’. 2009. https://www.garant.ru/products/ipo/prime/doc/4089228/ (accessed Feb \n7, 2020).  \n46. Mishin V, M",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list, including a Russian health ministry order and further Russian comparative studies of Diaskintest.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-898",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 898",
      "pageStart": 898,
      "pageEnd": 898,
      "target": "../sections/page-898.md",
      "narrative": null,
      "content": "35 \n \n58. Starshinova AA, Zinchenko YS, Istomina E V., et al. Diagnosis of Latent Tuberculosis Infection in \nPersonnel of Various Institutions and Determination of the Risk Group for Tuberculosis. \nBioprep Prev Diagnosis, Treat 2019; 19: 178–84.  \n59. Vaganova US, Russkih OE. Ways of rising detectability of Tuberculosis among health workers. J \nPharm Qual Assur 2015; 2: 61–7.  \n60. A S. A comparison of intradermal test with recombinant tuberculosis allergen (diaskintest) with \nother immunologic tests in the diagnosis of tuberculosis infection. Int J Mycobacteriology \n2018;7:32-9. \n61. Zhang H,",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Ends the reference list with studies on Diaskintest in health workers, Chinese trials of the ESAT6-CFP10 skin test and the Cochrane handbook.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-899",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 899",
      "pageStart": 899,
      "pageEnd": 899,
      "target": "../sections/page-899.md",
      "narrative": null,
      "content": "36 \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \nAnnex 1: Pooled results of the systematic review. \n \nFigure 1. Sensitivity of TBST in head-to-head studies. \n{Ruhwald, 2017 #2778}{Ruhwald, 2017 #2778}",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The heading of Annex 1 with pooled review results and the caption of a figure on sensitivity of novel skin tests in head-to-head studies, with stray citation codes.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p900-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 900,
      "pageEnd": 900,
      "target": "../images/img-p900-1.webp",
      "narrative": "draft",
      "content": "A forest plot of sensitivity (%) in six head-to-head studies grouped by test, with random-effects pooled estimates: TST 81.61% (72.34–88.27), IGRA (QFT) 77.04% (66.36–85.10) and TBST 78.10% (70.61–84.11). The test for subgroup differences is not significant (p = 0.74).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-900",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 900",
      "pageStart": 900,
      "pageEnd": 900,
      "target": "../sections/page-900.md",
      "narrative": null,
      "content": "37 \n \n  \nStarshinova 2018 (a) and Starshinova 2019 (a) studied DST5mm. The rest studied C-Tb. \nTST cut-off was 5mm for HIV+ and 15 mm for HIV- in Aggerbeck, 2018, Aggerbeck 2019, Hoff 2016, \nand Ruhwald 2017. TST cut-off was 5mm in Starshinova 2018 (a) and Starshinova 2019 (a).  \nTBST: novel skin tests for TB infection; TST: tuberculin skin test: QFT: quantiferon; CT: confidence \ninterval; TP: true positive; FN: false negative.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Footnotes for the head-to-head sensitivity figure, stating which studies tested Diaskintest or C-Tb and which TST cut-offs were used.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p901-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 901,
      "pageEnd": 901,
      "target": "../images/img-p901-1.webp",
      "narrative": "draft",
      "content": "A forest plot of TBST sensitivity (%) in individuals with HIV-negative or unknown status, grouped by test: DST (any induration size) 66.37% (59.03–72.99), DST (induration ≥5mm) 88.41% (78.49–94.09), C-Tb (≥5mm) 75.78% (70.37–80.47) and C-TST (≥5mm) 86.29% (82.77–89.19). The overall random-effects pooled sensitivity is 75.96% (70.32–80.82) across 1276 participants.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-901",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 901",
      "pageStart": 901,
      "pageEnd": 901,
      "target": "../sections/page-901.md",
      "narrative": null,
      "content": "38 \n \nFigure 2. Sensitivity of TBST in all studies in individuals with HIV-negative or unknown status \n \nThe pooled sensitivity when DST5mm was excluded: 74.21% [95%CI 68.30; 79.36] \nTBST: novel skin tests for TB infection; DST: Diaskintest; TST: tuberculin skin test: QFT: quantiferon; \nCI: confidence interval; TP: true positive; FN: false negative.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Caption for a figure on sensitivity of novel skin tests in people with HIV-negative or unknown status, giving a pooled sensitivity of about 74% when Diaskintest at 5 mm is excluded.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p902-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 902,
      "pageEnd": 902,
      "target": "../images/img-p902-1.webp",
      "narrative": "draft",
      "content": "A forest plot of TBST sensitivity (%) in HIV-positive individuals, grouped by test: DST (≥5mm) 70.27% (53.87–82.71), DST (any induration size) 39.77% (30.11–50.30) and C-Tb (≥5mm) 68.75% (61.85–74.91). The overall random-effects pooled sensitivity is 63.49% (52.60–73.16) across 317 participants.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p902-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 902,
      "pageEnd": 902,
      "target": "../images/img-p902-2.webp",
      "narrative": "draft",
      "content": "A forest plot of TBST sensitivity (%) in children, grouped by test: DST AI (Aksenova 2011, Nikitina 2019, Borodulina 2014) pooled at 94.74% (70.61–99.26) and DST 5mm (Starshinova 2018) at 100.00%. The overall random-effects pooled sensitivity is 97.06% (81.86–99.59) across 34 participants.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-vfig-p902",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 902,
      "pageEnd": 902,
      "target": "../figures/vfig-p902.png",
      "narrative": "draft",
      "content": "A page render with footnote text and Figure 4, a forest plot of TBST sensitivity (%) in children: DST AI (three studies) pooled at 94.74% (70.61–99.26) and DST 5mm (Starshinova 2018) at 100.00%. The overall random-effects pooled sensitivity is 97.06% (81.86–99.59) across 34 participants.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-902",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 902",
      "pageStart": 902,
      "pageEnd": 902,
      "target": "../sections/page-902.md",
      "narrative": null,
      "content": "39 \n \nFigure 3. Sensitivity of TBST in HIV-positive individuals  \n \n \n \n \n \ns 0.6131 [0.4790; 0.7 319] \nPooled sensitivity when DST5mm was excluded:  61.31% [47.90; 73.19] \nTBST: novel skin tests for TB infection; DST: Diaskintest; TST: tuberculin skin test: QFT: quantiferon; \nCI: confidence interval; TP: true positive; FN: false negative. \n \n \nFigure 4 Sensitivity of TBST in children \n \n \nss\n \nss \n \n \nTBST: novel skin tests for TB infection; DST: Diaskintest; AI: Any induration size; CI: confidence \ninterval; TP: true positive; FN: false negative. \nAggerbeck 2018 estimated the sensitivity of",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Captions for figures on sensitivity of novel skin tests in people with HIV, pooled at about 61%, and in children, noting one study excluded because few cases were confirmed.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p903-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 903,
      "pageEnd": 903,
      "target": "../images/img-p903-1.webp",
      "narrative": "draft",
      "content": "A forest plot of percentage agreement between TBST and IGRA in studies including participants without active TB, grouped by test pair: C-Tb 5mm vs QFT 90.45%, DST AI vs QFT 90.67%, DST 5mm vs QFT 97.18%, DST papule 7+mm vs T-SPOT.TB 60.93% and C-TST 5mm vs T-SPOT 85.48%. The overall random-effects pooled agreement is 88.96% (82.60–93.19) across 2510 participants, with significant subgroup differences (p < 0.01).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-903",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 903",
      "pageStart": 903,
      "pageEnd": 903,
      "target": "../sections/page-903.md",
      "narrative": null,
      "content": "40 \n \nFigure 5 Agreement of TBST vs IGRA in all studies including participants without active TB \n \n \nPooled agreement when only DSTAI was used for DST: 89.28 [95%CI 84.13; 92.90] \nTBST \nnovel skin tests for TB infection QFT \nquantiferon \nDST \nDiaskintest \n \n \nCI \nconfidence interval \nAI \nany induration size \n \nntb \nnon-TB \nUS \nUnder surveillance for TB \nPTB/TBI study sample includes individuals with pulmonary TB or latent TB infection",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Caption for a figure on agreement between novel skin tests and IGRA in people without active tuberculosis, with a pooled value of about 89%.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p904-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 904,
      "pageEnd": 904,
      "target": "../images/img-p904-1.webp",
      "narrative": "draft",
      "content": "A forest plot of percentage agreement between TBST and IGRA in studies including people with active TB, grouped by test pair: C-Tb 5mm vs QFT 79.59%, DST AI vs QFT 97.28%, DST 5mm vs QFT 97.01% and C-TST 5mm vs T-SPOT 85.42%. The overall random-effects pooled agreement is 85.71% (79.53–90.26) across 870 participants, with significant subgroup differences (p < 0.01).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-904",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 904",
      "pageStart": 904,
      "pageEnd": 904,
      "target": "../sections/page-904.md",
      "narrative": null,
      "content": "41 \n \nFigure 6. Agreement of TBST vs IGRA in all studies including people with active TB \n \n \nPooled agreement when DST 5mm was excluded: 81.83 [95%CI 77.56; 85.45] \nTBST \nnovel skin tests for TB infection QFT \nquantiferon \nDST \nDiaskintest \n \n \nCI \nconfidence interval \nAI \nany induration size",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Caption for a figure on agreement between novel skin tests and IGRA in people with active tuberculosis, pooled at about 82% with Diaskintest at 5 mm excluded.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p905-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 905,
      "pageEnd": 905,
      "target": "../images/img-p905-1.webp",
      "narrative": "draft",
      "content": "A forest plot of percentage agreement between TBST and TST in studies including participants without active TB, grouped by test pair: C-Tb5mm vs TST5mm 85.95%, C-Tb5mm vs TST15mm 81.61%, DST AI vs TST5mm 47.38%, DST5mm vs TST5mm 41.90% and C-TST vs TST 5mm 59.59%. The overall random-effects pooled agreement is 59.42% (45.34–72.10) across 4857 participants, with wide variation among DST studies.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-905",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 905",
      "pageStart": 905,
      "pageEnd": 905,
      "target": "../sections/page-905.md",
      "narrative": null,
      "content": "42 \n \nFigure 7 Agreement of TBST vs TST in all studies including participants without active TB \n \n \n \nTsss \nPooled agreement when DST5mm was excluded: 63.20% [95%CI: 47.38; 76.62] \nUI  \nunder investigation for TB  \nCI \nconfidence interval \n \n \nnTB  \nnon-TB  \n \n \nTST \ntuberculin skin test  \n \n \nG  \ngeneral population       \n \n \nUT  \nactive TB or under investigation for TB, results not reported separately  \nDST     Diaskintest \nPE \npotentially exposed (health care workers) \n \nTtt",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Caption for a figure on agreement between novel skin tests and TST in people without active tuberculosis, pooled at about 63%, with abbreviation notes and stray characters.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p906-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 906,
      "pageEnd": 906,
      "target": "../images/img-p906-1.webp",
      "narrative": "draft",
      "content": "A forest plot of percentage agreement between TBST and TST in studies including people with active TB, grouped by test pair: C-Tb5mm vs TST5mm 80.84%, C-Tb5mm vs TST15mm 76.39%, DST AI vs TST5mm 92.56%, DST5mm vs TST5mm 95.36% and C-TST vs TST 5mm 81.25%. The overall random-effects pooled agreement is 88.29% (82.14–92.51) across 2260 participants.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-906",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 906",
      "pageStart": 906,
      "pageEnd": 906,
      "target": "../sections/page-906.md",
      "narrative": null,
      "content": "43 \n \nFigure 8 Agreement of TBST vs TST in all studies including people with active TB \n \nPooled agreement when DST5mm was excluded: 86.98% [80.76-91.41%] \n \nactNR  active TB, diagnostic method not reported  TST \ntuberculin skin test  \n \n \nCI \nconfidence interval \n \nCC  \nculture-confirmed TB \n \nCCPCR culture or PCR confirmed  \n \nCLM  clinically or microbiologically confirmed TB \nCCS  \nculture or smear confirmed TB  \nDST     Diaskintest",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Caption for a figure on agreement between novel skin tests and TST in people with active tuberculosis, pooled at about 87%, with abbreviations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p907-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 907,
      "pageEnd": 907,
      "target": "../images/img-p907-1.webp",
      "narrative": "draft",
      "content": "A forest plot of percentage agreement between TBST and IGRA in children without active TB, by test pair: C-Tb vs QFT 89.52%, DST AI vs QFT 90.67%, DST 5mm vs QFT 97.18% and DST papule 7+mm vs T-SPOT.TB 60.93%. The overall random-effects pooled agreement is 89.26% (79.25–94.76) across 779 participants.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-907",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 907",
      "pageStart": 907,
      "pageEnd": 907,
      "target": "../sections/page-907.md",
      "narrative": null,
      "content": "44 \n \nTttt \nFigure 9 Agreement of TBST vs IGRA in children without active TB \n \n \n \nPooled agreement when DST5mm was excluded: 54.02 [26.46; 79.31] \nTBST \nnovel skin tests for TB infection QFT \nquantiferon \nDST \nDiaskintest \n \n \nCI \nconfidence interval \nAI \nany induration size \n \nntb \nnon-TB \nUS \nUnder surveillance for TB \nPTB/TBI study sample includes individuals with pulmonary TB or latent TB infection",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Caption for a figure on agreement between novel skin tests and IGRA in children without active tuberculosis, pooled at about 54% with wide uncertainty.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p908-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 908,
      "pageEnd": 908,
      "target": "../images/img-p908-1.webp",
      "narrative": "draft",
      "content": "A forest plot of percentage agreement between TBST and TST in children without active TB, by test pair: DST AI vs TST 5mm 39.21% (seven studies ranging from 6.25% to 85.71%), DST 5mm vs TST 5mm 76.74% and C-Tb vs TST 15mm 90.37%. The overall random-effects pooled agreement is 56.73% (30.85–79.40) across 2202 participants.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p908-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 908,
      "pageEnd": 908,
      "target": "../images/img-p908-2.webp",
      "narrative": "draft",
      "content": "A forest plot of percentage agreement between TBST and TST in children with active TB, by test pair: DST AI vs TST5mm 97.04% (three studies) and DST 5mm vs TST 5mm 98.85% (one study). The overall random-effects pooled agreement is 97.39% (96.39–98.12) across 1342 participants.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-908",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 908",
      "pageStart": 908,
      "pageEnd": 908,
      "target": "../sections/page-908.md",
      "narrative": null,
      "content": "45 \n \n \n \n \n \nFigure 10 Agreement of TBST vs TST in children without  active TB \n \n \nPooled agreement when DST5mm was excluded 89.81 [83.30; 93.96] \n \n \n \n \n \n \n \nTsssUI  \nunder investigation for TB  \nCI \nconfidence interval \n \n \nnTB  \nnon-TB  \n \n \nTST \ntuberculin skin test  \n \n \nG  \ngeneral population       \n \n \nUT  \nactive TB or under investigation for TB, results not reported separately  \nDST     Diaskintest \n \n \nNegC \nNegative control \nCL \nClose contacts \n \n \nFigure 11 Agreement of TBST vs TST in children with active TB \n \nactNR  active TB, diagnostic method not reported  TST \ntuberculin s",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Captions for figures on agreement between novel skin tests and TST in children without and with active tuberculosis, with abbreviation notes.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-909",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 909",
      "pageStart": 909,
      "pageEnd": 909,
      "target": "../sections/page-909.md",
      "narrative": null,
      "content": "46 \n \nCCS  \nculture or smear confirmed TB  \nDST     Diaskintest",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A near-empty page holding only two leftover abbreviations from a figure legend.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p910-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 910,
      "pageEnd": 910,
      "target": "../images/img-p910-1.webp",
      "narrative": "draft",
      "content": "A forest plot of percentage agreement between TBST and TST in children aged under 5 years without active TB: DST AI vs TST 5mm 76.57% (Dovgalyuk 2012) and C-Tb vs TST 15mm 88.51% (Aggerbeck 2018). The overall random-effects pooled agreement is 81.97% (71.38–89.23) across 484 participants.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p910-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 910,
      "pageEnd": 910,
      "target": "../images/img-p910-2.webp",
      "narrative": "draft",
      "content": "A forest plot of percentage agreement between TBST and IGRA in BCG-vaccinated individuals without active TB: DST 5mm vs QFT 97.18% (Starshinova 2018, children) and C-TST vs T-SPOT.TB 97.44% (Xu 2021, adults). The overall random-effects pooled agreement is 97.22% (93.96–98.75) across 216 participants.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-910",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 910",
      "pageStart": 910,
      "pageEnd": 910,
      "target": "../sections/page-910.md",
      "narrative": null,
      "content": "47 \n \nFigure 12 Agreement of TBST vs TST in children aged < 5 years without active TB \n \n \n \nCI \nconfidence interval \n \nTST \ntuberculin skin test  \n \n \nG  \ngeneral population      \n \nCL \nClose contacts \nDST     Diaskintest \n \n \n \n \nFigure 13 Agreement of TBST vs QFT in BCG-vaccinated individuals without active TB \n \n \nCI \nconfidence interval \n \nQFT \nquantiferon  \n \n \nntb \nnon-TB \nDST     Diaskintest",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Captions for figures on agreement of novel skin tests with TST in children under five without active tuberculosis and with QuantiFERON in BCG-vaccinated people without active tuberculosis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p911-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 911,
      "pageEnd": 911,
      "target": "../images/img-p911-1.webp",
      "narrative": "draft",
      "content": "A forest plot of percentage agreement between TBST and QFT in BCG-vaccinated children with active TB: DST 5mm vs QFT 97.01% (Starshinova 2018) and DST AI vs QFT 100.00% (Nakonechnaya 2020). The overall random-effects pooled agreement is 97.65% (93.90–99.11) across 170 participants.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p911-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 911,
      "pageEnd": 911,
      "target": "../images/img-p911-2.webp",
      "narrative": "draft",
      "content": "A forest plot of percentage agreement between TBST and TST in BCG-vaccinated individuals without active TB: DST 5mm vs TST 5mm 52.30% (Starshinova 2018 28.34%, Yablonskiy 2013 76.74%) and C-TST vs TST 5mm 16.67% (Xu 2021). The overall random-effects pooled agreement is 38.76% (14.09–70.94) across 519 participants.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-911",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 911",
      "pageStart": 911,
      "pageEnd": 911,
      "target": "../sections/page-911.md",
      "narrative": null,
      "content": "48 \n \nFigure 14 Agreement of TBST vs QFT in BCG-vaccinated individuals with active TB \n \n \nCI \nconfidence interval \n \nQFT \nquantiferon  \n \n \nntb \nnon-TB \nDST     Diaskintest \n \n \nFigure 15 Agreement of TBST vs TST in BCG-vaccinated individuals without active TB \n \n \n \nCI \nconfidence interval \n \nTST \ntuberculin skin test  \n \n \nntb \nnon-TB  \n \n \nUI \nunder investigation for TB \nDST     Diaskintest",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Captions for figures on agreement of novel skin tests with QuantiFERON and with TST in BCG-vaccinated people with and without active tuberculosis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p912-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 912,
      "pageEnd": 912,
      "target": "../images/img-p912-1.webp",
      "narrative": "draft",
      "content": "A forest plot of the difference in specificity (%) between TBST and IGRA, by test pair: DST 5mm vs QFT 4.50 (−13.13 to 22.12), C-Tb vs QFT 0.70 (−7.02 to 8.42) and C-TST vs T-SPOT.TB 3.54 (−1.37 to 8.44). The overall random-effects pooled difference is 2.29 (−1.60 to 6.18), with confidence intervals crossing zero.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-912",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 912",
      "pageStart": 912,
      "pageEnd": 912,
      "target": "../sections/page-912.md",
      "narrative": null,
      "content": "49 \n \nFigure 16 Difference in specificity -TBST vs IGRA (using data on DST vs QFT from Starshinova 2018) \n \n \nCI \nconfidence interval \n \nQFT \nquantiferon  \n \n \nDST     Diaskintest \n \nNote: Starshinova 2018a and Starshinova 2018b presented data on DST 5mm vs QFT and DST 5mm \nand TSPOT.TB using the same cohort of participants. Because of the overlap in the cohort, both DST \n5mm vs QFT and DST 5mm and TSPOT.TB cannot be pooled together. Hence, the graph includes \nonly estimates for DST 5mm vs QFT. The next graph includes DST 5mm vs TSPOT.TB. \n*Alternate specificity measure. Estimates difference i",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Caption for a figure on the difference in specificity between novel skin tests and IGRA, explaining that overlapping cohorts from one study meant only the Diaskintest–QuantiFERON comparison was pooled here.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p913-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 913,
      "pageEnd": 913,
      "target": "../images/img-p913-1.webp",
      "narrative": "draft",
      "content": "A forest plot of the difference in specificity (%) between TBST and IGRA, by test pair: DST 5mm vs T-SPOT.TB −9.49 (−34.96 to 15.98), C-Tb vs QFT 0.70 (−7.02 to 8.42) and C-TST vs T-SPOT.TB 3.54 (−1.37 to 8.44). The overall random-effects pooled difference is −2.21 (−10.27 to 5.86).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p913-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 913,
      "pageEnd": 913,
      "target": "../images/img-p913-2.webp",
      "narrative": "draft",
      "content": "A forest plot of the difference in specificity (%) between TBST and TST, by test pair: DST 5mm vs TST 5mm 57.26, DST AI vs TST 5mm 29.93, C-Tb vs TST 15mm 4.61, C-Tb vs TST 5/15mm −2.00 and C-TST vs TST 5mm 39.90. The overall random-effects pooled difference is 33.47 (18.16 to 48.78), favouring higher specificity for TBST, with high residual heterogeneity (I² = 96%).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-913",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 913",
      "pageStart": 913,
      "pageEnd": 913,
      "target": "../sections/page-913.md",
      "narrative": null,
      "content": "50 \n \nFigure 17 Difference in specificity - TBST vs IGRA (using data on DST vs TSPOT.TB from Starshinova \n2018) \n \n \nCI \nconfidence interval \n \nQFT \nquantiferon  \n \n \nDST     Diaskintest \n \nNote: Starshinova 2018a and Starshinova 2018b presented data on DST 5mm vs QFT and DST 5mm \nand TSPOT.TB using the same cohort of participants. Because of the overlap in the cohort, both DST \n5mm vs QFT and DST 5mm and TSPOT.TB cannot be pooled together. Hence, the graph includes \nonly estimates for DST 5mm vs TSPOT.TB. The previous graph includes DST 5mm vs QFT. \n*Alternate specificity measure. Estimates d",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Captions for figures on specificity differences between novel skin tests and IGRA, using the Diaskintest–T-SPOT.TB data, and between novel skin tests and TST, with a pooled difference of about 22 percentage points when Diaskintest at 5 mm is excluded.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-914",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 914",
      "pageStart": 914,
      "pageEnd": 914,
      "target": "../sections/page-914.md",
      "narrative": null,
      "content": "51 \n \n*Alternate specificity measure. Estimates difference in proportion negative in healthy populations \n(see the method section for the definition).",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A footnote explaining that an alternative specificity measure estimates the difference in the proportion testing negative in healthy populations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p915-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 915,
      "pageEnd": 915,
      "target": "../images/img-p915-1.webp",
      "narrative": "draft",
      "content": "A forest plot of the difference in specificity (%) between TBST and QFT in children: DST 5mm vs QFT 10.60 (2.30 to 18.90; Starshinova 2018a) and C-Tb vs QFT 6.50 (−5.25 to 18.25; Aggerbeck 2018). The overall random-effects pooled difference is 9.21 (−15.46 to 33.88).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p915-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 915,
      "pageEnd": 915,
      "target": "../images/img-p915-2.webp",
      "narrative": "draft",
      "content": "A forest plot of the difference in specificity (%) between TBST and TST in children: DST 5mm vs TST 5mm 70.70 (Starshinova 2018a), DST AI vs TST 5mm 12.30 (Dovgalyuk 2013) and C-Tb vs TST 15mm −2.00 (Aggerbeck 2018). The overall random-effects pooled difference is 27.15 (−68.57 to 122.87), with significant subgroup differences (p < 0.01).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-915",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 915",
      "pageStart": 915,
      "pageEnd": 915,
      "target": "../sections/page-915.md",
      "narrative": null,
      "content": "52 \n \nFigure 19 Difference in specificity - TBST vs IGRA in children \n \n  \nCI \nconfidence interval \n \nQFT \nquantiferon  \n \n \nDST     Diaskintest \n \n*Alternate specificity measure. Estimates difference in proportion negative in healthy populations \n(see the method section for the definition). \n  \nFigure 20 Difference in specificity - TBST vs TST in children \n \n \nPooled difference when DST5mm  was excluded: 20.17 [95%CI 4.93; 35.42] \nCI \nconfidence interval \n \nTST \ntuberculin skin test  \n \n \nDST     Diaskintest \n \n \n*Alternate specificity measure. Estimates difference in proportion negative in h",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Captions for figures on specificity differences between novel skin tests and IGRA, and between novel skin tests and TST, in children, with a pooled difference versus TST of about 20 points.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p916-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 916,
      "pageEnd": 916,
      "target": "../images/img-p916-1.webp",
      "narrative": "draft",
      "content": "A forest plot of the difference in specificity (%) between TBST and T-SPOT.TB in BCG-vaccinated individuals: DST 5mm vs T-SPOT.TB −12.61 (Starshinova 2018a, adults −5.90 and children −18.70) and C-TST vs T-SPOT.TB −2.60 (Xu 2021, adults). The overall random-effects pooled difference is −9.67 (−31.12 to 11.79).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p916-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 916,
      "pageEnd": 916,
      "target": "../images/img-p916-2.webp",
      "narrative": "draft",
      "content": "A forest plot of the difference in specificity (%) between TBST and TST 5mm in BCG-vaccinated individuals: DST 5mm vs TST 5mm 60.12 (Starshinova 2018a, adults 48.30 and children 70.70) and C-TST vs TST 5mm 83.30 (Xu 2021, adults). The overall random-effects pooled difference is 67.39 (24.04 to 110.73).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-916",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 916",
      "pageStart": 916,
      "pageEnd": 916,
      "target": "../sections/page-916.md",
      "narrative": null,
      "content": "53 \n \n \nFigure 21 Difference in specificity - TBST vs IGRA in BCG-vaccinated individuals  \n \nCI \nconfidence interval \n \nQFT \nquantiferon  \n \n \nDST     Diaskintest \n \n*Alternate specificity measure. Estimates difference in proportion negative in healthy populations \n(see the method section for the definition). \n  \n \nFigure 22 Difference in specificity- TBST vs TST in BCG-vaccinated individuals \n \n \nPooled difference when DST5mm  was excluded: 20.17 [95%CI 4.93; 35.42] \nCI \nconfidence interval \n \nTST \ntuberculin skin test  \n \n \nDST     Diaskintest \n \n \n \n*Alternate specificity measure. Estimates",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Captions for figures on specificity differences between novel skin tests and IGRA or TST in BCG-vaccinated people, with explanatory footnotes.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p917-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 917,
      "pageEnd": 917,
      "target": "../images/img-p917-1.webp",
      "narrative": "draft",
      "content": "A forest plot of TBST specificity (%) in healthy individuals with negative IGRA results, by test pair: DST 5mm vs QFT 99.06% (Starshinova 2018), C-Tb vs QFT 98.01% (three studies, 91.43% to 99.32%) and C-TST vs T-SPOT.TB 95.48% (Xu 2021). The x-axis runs from 80 to 100; subgroup differences are not significant (p = 0.18).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p917-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 917,
      "pageEnd": 917,
      "target": "../images/img-p917-2.webp",
      "narrative": "draft",
      "content": "A forest plot of TBST specificity (%) in healthy children with negative IGRA results: DST 5mm vs QFT 99.06% (Starshinova 2018) and C-Tb vs QFT 91.43% (Aggerbeck 2018). The overall random-effects pooled specificity is 96.74% (85.52–99.33) across 176 participants.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p917-3",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 917,
      "pageEnd": 917,
      "target": "../images/img-p917-3.webp",
      "narrative": "draft",
      "content": "A forest plot of TBST specificity (%) in BCG-vaccinated individuals with negative IGRA results from three studies: Xu 2021 (C-TST vs T-SPOT.TB) 97.44%, Starshinova 2018 (DST 5mm vs QFT) 99.06% and Aggerbeck 2013 (C-Tb vs QFT) 99.32%. The random-effects pooled specificity is 98.97% (96.86–99.67) across 292 participants.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-917",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 917",
      "pageStart": 917,
      "pageEnd": 917,
      "target": "../sections/page-917.md",
      "narrative": null,
      "content": "54 \n \nFigure 23 Specificity in healthy individuals with negative IGRA results \n \nPooled estimated when DST5mm vs TSPOT.TB was used instead of DST5mm vs QFT: 98.46 [95%CI 94.36; \n99.60] \nCI \nconfidence interval \n \nTST \ntuberculin skin test  \n \n \nDST     Diaskintest \n \n \nFigure 24 Specificity in healthy children with negative IGRA results \n \n \nCI \nconfidence interval \n \nQFT \nquantiferon  \n \n \nDST     Diaskintest \n \n \nFigure 25 Specificity in BCG-vaccinated individuals with negative IGRA results \n \nPooled estimated when DST5mm vs TSPOT.TB was used instead of DST5mm vs QFT: 99.49 [97.97; 99.87]",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Captions for figures on specificity of novel skin tests in healthy adults, children and BCG-vaccinated people with negative IGRA results, with pooled estimates near 98–99%.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-918",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 918",
      "pageStart": 918,
      "pageEnd": 918,
      "target": "../sections/page-918.md",
      "narrative": null,
      "content": "55 \n \n \nWeb Annex B 21. Safety of Mycobacterium tuberculosis antigen-\nbased skin tests: a systematic review and meta-analysis \n \nReview team: Yohhei Hamada1, Irina Kontsevaya2, Elena Surkova3, Ting Ting \nWang1, Victoria Liu4, Liliya Eugenevna Ziganshina5,6, Molebogeng X Rangaka1,7 \n \n8. Institute for Global Health, University College London, UK \n9. Research Center Borstel, Germany \n10. Royal Brompton Hospital. Part of Guy's and St Thomas' NHS Foundation Trust \nLondon, UK \n11. London School of Hygiene and Tropical Medicine, UK \n12. Institute of Fundamental Medicine and Biology of Kazan Federal",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title page of Web Annex B21, a systematic review and meta-analysis on the safety of tuberculosis antigen-based skin tests, naming the review team and advisers and noting it was prepared for a WHO guideline meeting in February 2022.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-919",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 919",
      "pageStart": 919,
      "pageEnd": 919,
      "target": "../sections/page-919.md",
      "narrative": null,
      "content": "56 \n \n \n \n \nAbbreviations \n \nAE \nadverse events \nBCG \nbacille Calmette-Guérin \nCI \nconfidence interval \nCI \nconfidence interval \nDAIDS \ndivision of AIDS \nDST \ndiaskintest \nGDG \nguideline development group \nIGRA \ninterferon gamma release assay \nISR \ninjection site reaction \nMTB \nmycobacterium tuberculosis \nPICO \npopulation, intervention, comparison, outcome \nPPD \npurified protein derivative \nQFT \nquantiferon \nRR \nrisk ratio \nTB \ntuberculosis \nTBST \nnovel skin-based tests for TB infection using TB specific antigens (TBST) \nTST \ntuberculin skin test \nWHO \nWorld Health Organization",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The list of abbreviations used in the safety review of novel tuberculosis skin tests.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-920",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 920",
      "pageStart": 920,
      "pageEnd": 920,
      "target": "../sections/page-920.md",
      "narrative": null,
      "content": "57 \n \nExecutive summary \n \nBackground \nThe diagnostic tests in current use for identification of TB infection are the tuberculin skin test (TST) \nand interferon-gamma release assay (IGRA). The safety of TST, an in-vivo test, is well established and \nassociated with only uncommon adverse reactions. Over the last decade, novel skin-based tests for \nTB infection using TB specific antigens (TBST) have been developed. Evidence on the safety of these \ntests has not been systematically reviewed. \n \nMethod \nWe systematically searched for studies reporting the outcomes of our interest: local reactions",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The executive summary of the safety review explains that the tuberculin skin test is known to be safe but that newer antigen-based skin tests had not been systematically assessed. It describes searches of English, Russian and Chinese databases up to July 2021, plus manufacturer data and a WHO call for data. It notes that some studies were at high risk of bias owing to lack of blinding or passive collection of adverse events.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-921",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 921",
      "pageStart": 921,
      "pageEnd": 921,
      "target": "../sections/page-921.md",
      "narrative": null,
      "content": "58 \n \n \nThe pooled risk of any injection site reaction (ISR) due to C-Tb (N= 2931, 5 studies) was not \nsignificantly different from TST (RR 1.05; 95%CI 0.70-1.58). Over 95% of ISR were reported as mild or \nmoderate; common ISR included pain, itching and rash. In one study that allowed, 49/153 (37.6%) of \nparticipants given C-Tb developed any systemic adverse events (e.g. fever and headache) in \ncomparison with 37.6% (56/149) in those given TST (RR 0.85 95%CI 0.6-1.2). The remaining studies \ndid not allow the comparison of systemic adverse events. \n \nIn a single paper in China reporting combine",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarises key findings: injection site reactions with C-Tb were no more common than with TST and were mostly mild or moderate, while one Chinese report found more reactions with C-TST. Diaskintest safety data were too inconsistent to pool, and overall no safety signal was seen that would affect the choice between the novel tests and TST. The background notes that about a quarter of the world's population has tuberculosis infection.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-922",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 922",
      "pageStart": 922,
      "pageEnd": 922,
      "target": "../sections/page-922.md",
      "narrative": null,
      "content": "59 \n \nsuch as people living with HIV and those who are known to be recently infected (i.e. household \ncontacts of people with TB) are important targets for treatment of TB infection, also known as TB \npreventive treatment.4 Currently, available tests for LTBI are imperfect, as they cannot accurately \ndistinguish between active TB disease and TB infection, nor are they useful predictors of progression \nto active disease.5 More accurate diagnostic tests are critical to the achievement of the targets of \nEnd TB Strategy.6 \n \nThe diagnostic tests in current use are the tuberculin skin test (TST) a",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains why better tests for tuberculosis infection are needed, describing the drawbacks of TST, such as false positives after BCG vaccination, low sensitivity in people with HIV and the need for two visits. It notes that the safety of the novel skin tests had not previously been reviewed and states the review's objective.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-923",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 923",
      "pageStart": 923,
      "pageEnd": 923,
      "target": "../sections/page-923.md",
      "narrative": null,
      "content": "60 \n \ncurrently available in vitro IGRA tests and TST to inform the development of WHO guidelines.  \n \nAims and Objectives \n \nAim \nTo evaluate the safety of novel TBST for detection of TB infection compared to currently available in \nvitro IGRA tests or TST. \nPICO \nDo TBST cause more adverse reactions compared to TST or IGRA? \nOutcomes: \n- \nRelative risks of both local and systemic reactions adverse reactions, where possible, graded \nby type, severity and seriousness and stratified by sub-group;  \n- \nFrequency (%) of adverse events (if a control group receiving a comparator test is \nunavailabl",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Sets out the review's aim and PICO question on whether novel skin tests cause more adverse reactions than TST or IGRA, the outcomes sought, and inclusion criteria naming the index tests, such as C-Tb, Diaskintest and C-TST, and comparators.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-924",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 924",
      "pageStart": 924,
      "pageEnd": 924,
      "target": "../sections/page-924.md",
      "narrative": null,
      "content": "61 \n \n \nBased on published literature, we expected the most frequent adverse reactions to be injection site \nreactions, although systemic reactions like headache, fever and even lymphadenitis have also been \nreported.11 12  Induration/swelling/inflammation are the intent of the skin tests and may not \nnecessarily be “adverse” reactions in all instances, however some clinical trials consider these as \nadverse reactions if exuberant (e.g. induration size ≥ 50mm and erythema  ≥ 80mm)12. We report \nthem as defined by the authors.12 \n \nData on the severity of adverse reactions were collected accord",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains how adverse reactions were defined, noting that induration is the intended response and counting only exaggerated reactions as adverse where trials did so. Severity was graded with the DAIDS scale and serious events defined in standard terms, followed by exclusion criteria and the search strategy.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-925",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 925",
      "pageStart": 925,
      "pageEnd": 925,
      "target": "../sections/page-925.md",
      "narrative": null,
      "content": "62 \n \nStudy Screening and Data Collection Process: \nIn support of data synthesis for the GDG, we developed a broad search strategy for English papers to \nallow title and abstract screening informing multiple systematic reviews addressing other objectives.  \nThe initial list of English titles and abstracts were reviewed by two independent reviewers (YH and \nLEZ) to identify studies reporting any new skin tests regardless of the outcomes of interest. This was \nfollowed by screening of titles and abstracts by two independent reviewers (YH and IK) and \nscreening of full-text articles. Two Russian",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes how studies in English, Russian and Chinese were screened and data extracted by pairs of independent reviewers, with disagreements resolved by discussion. It explains how risk of bias was judged for randomised and non-randomised studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-926",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 926",
      "pageStart": 926,
      "pageEnd": 926,
      "target": "../sections/page-926.md",
      "narrative": null,
      "content": "63 \n \nrandomized participants into different doses of TBST, this was not considered randomized studies. \nWe used the GRADE framework19 to systematically assess the quality of evidence regarding the use \nof novel skin tests.  \n \nData Analysis: \nThe summary measures for dichotomous outcomes were relative risks for dichotomous data or \nproportions in the absence of a control group, with 95% confidence intervals. We conducted a meta-\nanalysis using a random-effects model if included studies were considered to be clinically \nhomogenous (ie ignoring statistical heterogeneity in places). We used the",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the use of GRADE and the statistical methods, including random-effects meta-analysis of risk ratios and pooled proportions and combining trial and observational data where appropriate. It begins describing the included studies, which covered several recombinant antigen preparations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-927",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 927",
      "pageStart": 927,
      "pageEnd": 927,
      "target": "../sections/page-927.md",
      "narrative": null,
      "content": "64 \n \nin C-Tb.42 43 One study reported on ESAT-6 developed in-house.44 These studies are \nsummarized in Tables 2 and 3. We did not find studies that reported safety data on DPPD. \nHereafter, this report focuses on C-Tb, C-TST, and Diaskintest since the other tests or \nantigens are different from those translated into the final products or have not reached the \nlate stage of the product development.   \n \nCharacteristics of individual studies \nC-Tb \nOut of the seven studies that evaluated C-Tb, five reported data on injection site reactions \n(ISR) compared to TST (Table 2) 10 12 21 22 24 and in",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes that no safety data were found for DPPD and that the report focuses on C-Tb, C-TST and Diaskintest. It describes the seven C-Tb studies, five of which were double-blind comparisons with TST at low risk of bias, though systemic events could not be attributed to one test when both were given.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-928",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 928",
      "pageStart": 928,
      "pageEnd": 928,
      "target": "../sections/page-928.md",
      "narrative": null,
      "content": "65 \n \nOut of the five studies that evaluated C-TST, participants received both C-TST and TST in \nthree studies (Table 2).25 27 Tests were administered to either of the arms; allocation was \nnon-blinded and the choice was determined a priori without randomization. In one of the \nremaining two studies, a sub-set of the participants received C-TST in one arm and placebo \nin the other arm26 and, in the other study,25 participants received only C-TST; thus, these \nstudies did not provide safety data for the comparison of ISR between the two tests. None \nof the studies provided data for the comparis",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the C-TST studies, all in HIV-negative adults in China, which were non-blinded and at serious risk of bias for outcome measurement. It notes that most Diaskintest studies collected adverse events passively and poorly defined them, placing them at high risk of bias.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-929",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 929",
      "pageStart": 929,
      "pageEnd": 929,
      "target": "../sections/page-929.md",
      "narrative": null,
      "content": "66 \n \n \nInjection site reactions \nFor C-Tb, the frequency of any ISR reactions ranged from 23.7% to 53.1% (Table 4). Most \n(>95%) ISR were reported as mild to moderate by the investigators (mild: easily tolerated; \nmoderate: sufficient to interfere with daily activities; severe: sufficient to prevent normal \nactivity). Common ISR included itching, pain, and rash. One study reported only mild \nreactions and in four studies, <5% were of severe intensity. The pooled RR did not show \nevidence of a significant difference in the frequency of any ISR between C-Tb and TST \n(Figure 2, Figure A1).  Howe",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that injection site reactions with C-Tb occurred in about a quarter to half of recipients, mostly mild or moderate, with no significant pooled difference from TST despite high heterogeneity. It notes more frequent haematoma at the C-Tb site in European studies and gives limited C-TST comparison data.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-930",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 930",
      "pageStart": 930,
      "pageEnd": 930,
      "target": "../sections/page-930.md",
      "narrative": null,
      "content": "67 \n \nSafety data on the Diaskintest were poorly reported lacking standardization of types of \nadverse events and assessment of severity based on a priori criteria thus precluding pooling \nof data (Table 6). Two studies reported the frequency of ISR in participants given Diaskintest \nand TST at the same time in different arms; one included adults with active TB (n=53) and \nthe other TB care health workers (n=25).34 In the former study in adults with active TB, six \ndeveloped hyperallergic reactions with vesicles/necrosis and lymphangitis due to Diaskintest \ncompared to two due to TST (RR, 3.0;",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that Diaskintest safety data were poorly standardised, with two small studies noting more hyperallergic reactions with Diaskintest than TST in people with tuberculosis and in health workers. It begins reporting systemic adverse events in C-Tb studies, where only a small share were linked to the skin tests.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-931",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 931",
      "pageStart": 931,
      "pageEnd": 931,
      "target": "../sections/page-931.md",
      "narrative": null,
      "content": "68 \n \nrelated to either agent separately, they were ascribed to C-Tb.” In two studies, the \nfrequency of systemic adverse events deemed at least possibly related to C-Tb among \nparticipants were 5% (7/151)21 and 14% (36/253).24   \n \nFour studies on C-TST reported data on systemic adverse events (Table 8). In two studies, \nparticipants received both C-TST and TST.26 27 In the phase 2a study (n=144),25 nine systemic \nadverse events related to the test were reported. For the phase 2b study,27 only proportions \ncould be extracted without raw data. The authors did not respond to requests for data.",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives systemic adverse event data for C-Tb, C-TST and Diaskintest, including a pooled fever frequency of about 2.6% for Diaskintest. It reports no test-related serious adverse events among nearly 3000 C-Tb recipients and describes findings in people with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-932",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 932",
      "pageStart": 932,
      "pageEnd": 932,
      "target": "../sections/page-932.md",
      "narrative": null,
      "content": "69 \n \n \nChildren \nFive studies reported adverse events in children who received Diaskintest.28-30 36 37 As \nmentioned above, adverse events were not systematically ascertained (Table 6).  \nPregnant women \nOnly one study included pregnant women.31 The study by  Borisova et al used Diaskintest in 267 \npregnant women with TB.31 In 124 patients (46.4%) Diaskintest was performed in the first half of \npregnancy (but after 12 weeks), the rest in the second half.  The study reported that ‘no embryo \ntoxicity was registered’ without further details. Since the study included pregnant women with signs \na",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes that adverse events in children given Diaskintest were not systematically recorded and that one Russian study in pregnant women with tuberculosis reported no embryo toxicity. It rates evidence as high quality for injection site reactions and concludes that no safety signal affects the choice between the novel tests and TST.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-933",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 933",
      "pageStart": 933,
      "pageEnd": 933,
      "target": "../sections/page-933.md",
      "narrative": null,
      "content": "70 \n \n \nReferences \n \n1. Hay SI, Abajobir AA, Abate KH, et al. Global, regional, and national disability-adjusted life-years \n(DALYs) for 333 diseases and injuries and healthy life expectancy (HALE) for 195 countries \nand territories, 1990&#x2013;2016: a systematic analysis for the Global Burden of Disease \nStudy 2016. The Lancet 2017;390(10100):1260-344. doi: 10.1016/S0140-6736(17)32130-X \n2. Houben RMGJ, Dodd PJ. The Global Burden of Latent Tuberculosis Infection: A Re-estimation \nUsing Mathematical Modelling. PLOS Medicine 2016;13(10):e1002152. doi: \n10.1371/journal.pmed.1002152 \n3. Comstoc",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The start of the reference list for the safety review, citing global burden estimates, WHO preventive treatment guidance and studies on IGRAs.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-934",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 934",
      "pageStart": 934,
      "pageEnd": 934,
      "target": "../sections/page-934.md",
      "narrative": null,
      "content": "71 \n \nhttps://rsc.niaid.nih.gov/sites/default/files/daidsgradingcorrectedv21.pdf Accessed 11 \nSeptember 2020. \n14. Europen Medial Agency. Serious adverse reaction. \nhttps://www.ema.europa.eu/en/glossary/serious-adverse-reaction accessed 3 December \n2021.  \n15. Ouzzani M, Hammady H, Fedorowicz Z, et al. Rayyan—a web and mobile app for systematic \nreviews. Systematic Reviews 2016;5(1):210. doi: 10.1186/s13643-016-0384-4 \n16. Sterne JAC, Savović J, Page MJ, et al. RoB 2: a revised tool for assessing risk of bias in randomised \ntrials. BMJ 2019;366:l4898. doi: 10.1136/bmj.l4898 \n17. Sterne JA, Her",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the safety review reference list with methods sources such as DAIDS grading, risk-of-bias tools and GRADE, and trials of C-Tb and the ESAT6-CFP10 skin test.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-935",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 935",
      "pageStart": 935,
      "pageEnd": 935,
      "target": "../sections/page-935.md",
      "narrative": null,
      "content": "72 \n \n28. Aksenova VA, Klevno NI, Baryshnikova LA, et al. Diaskintest for tuberculosis in children and \nadolescents. [Russian]. Problemy tuberkuleza i boleznei legkikh 2009(10):13-16. \n29. Belushkov V.V. LME, Novik G.A., Gurina O.P., Shibakova N.D. IMPORTANCE OF DIASKINTEST AND \nQUANTIFERON TEST IN THE DIAGNOSTICS OF TUBERCULOSIS IN CHILDREN (in Russian) \nFUNDAMENTAL RESEARCH 2012;7:34-39. \n30. Bаrminа N, Bаryshnikovа L. Ways to enhance the efficiency of tuberculosis prevention in the \nchildren exposed to tuberculous infection illustrated by the example of Perm Kray (in \nTuberculosis and Lung",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the reference list with Russian and Ukrainian studies of Diaskintest, including in children, pregnant women and people with HIV, and Chinese studies of ESAT-6.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-936",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 936",
      "pageStart": 936,
      "pageEnd": 936,
      "target": "../sections/page-936.md",
      "narrative": null,
      "content": "73 \n \n41. Sun QF, Xu M, Wu JG, et al. Efficacy and safety of recombinant Mycobacterium tuberculosis \nESAT-6 protein for diagnosis of pulmonary tuberculosis: a phase II trial. Medical Science \nMonitor 2013;19:969-77. \n42. Arend SM, Franken WP, Aggerbeck H, et al. Double-blind randomized Phase I study comparing \nrdESAT-6 to tuberculin as skin test reagent in the diagnosis of tuberculosis infection. \nTuberculosis 2008;88(3):249-61. \n43. Lillebaek T, Bergstedt W, Tingskov PN, et al. Risk of sensitization in healthy adults following \nrepeated administration of rdESAT-6 skin test reagent by the Mant",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Ends the reference list with early trials of recombinant ESAT-6 skin test reagents.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-937",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 937",
      "pageStart": 937,
      "pageEnd": 937,
      "target": "../sections/page-937.md",
      "narrative": null,
      "content": "74 \n \n \n \nTable 1: Variables of interest \nCategory \nVariables \nStudy design \nStudy design, country, setting, period of \nrecruitment, sample size \nPopulation summary measures \nAge, gender, history of \nimmunosuppression, HIV status, pregnancy \nand lactating status, co-morbidities \nIndex test \nRecombinant antigen skin test used \nComparator \nTypes of tests used  \nOutcome \nType, severity, and seriousness of adverse \nevents.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 1 lists the variables extracted, covering study design, population characteristics, index and comparator tests, and the type, severity and seriousness of adverse events.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-938",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 938",
      "pageStart": 938,
      "pageEnd": 938,
      "target": "../sections/page-938.md",
      "narrative": null,
      "content": "Table 2. Summary of C-Tb, C-TST, and other studies.  \nTest \nStudy design \nControl \nCountry \nStudy population \nAge \n%Female \nC-Tb \nDose-finding trial \nNo control \nUK \nAdults with TB \nMean: 33 \nn/d \nC-Tb \nDouble-blind split body RCT \nTST \nUK \nBCG vaccinated healthy adults \nMean: 34 \n60.9% \nC-Tb \nDouble-blind split body RCT \nTST \nSouth Africa \nIndividuals suspected of TB, \nactive TB, contact and children. \nChildren < 5rs (20%) \nChildren 5-17 yrs (31%) \nMedian: 17 \n51% \nC-Tb \nDouble-blind split body RCT \nTST \nSouth Africa \nAdults with TB \nmean: 36 \n36% \nrdESAT-6 \nDouble-blind split body RCT \nTST",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 2 summarises the C-Tb, C-TST and other recombinant antigen studies, giving design, control, country, population, age and proportion female.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-939",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 939",
      "pageStart": 939,
      "pageEnd": 939,
      "target": "../sections/page-939.md",
      "narrative": null,
      "content": "Table 3. Summary of Diaskintest studies  \n \nStudy design \nControl \nCountry \nStudy population \nAge \nAksenova, 2009 \nCross-sectional \nn/d** \nRussia \nChildren and adolescents with TB  \n2-17 \nSlogotskaya, 2011a \nCross-sectional \nn/d** \nRussia \nTB/HIV coinfected adults \nmean: 32 \nSlogotskaya 2011b \nCross-sectional \nn/d** \nRussia \nPeople with TB including people living with HIV \nn/d \nStreltsova, 2011 \nCross-sectional \nTST \nRussia \nAdults with TB \nmean: 26 \nBelushkov, 2012 \nCross-sectional \nn/d \nRussia \nChildren suspected of  TB \nmean: 7.1 \nPatsyuk, 2017 \nCross-sectional \nn/d \nRussia \nPeople with TB",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 3 summarises the Diaskintest safety studies, mostly cross-sectional studies from Russia in children, adults with tuberculosis or HIV, pregnant women and health workers.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-940",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 940",
      "pageStart": 940,
      "pageEnd": 940,
      "target": "../sections/page-940.md",
      "narrative": null,
      "content": "Table 4. The frequencies of local injection site reactions in studies on C-Tb \nStudy \nTest \nN \nAny ISR \nItching \nPain \nRash \nErythema \nSwelling \nVesicle \ninduration \nAggerbeck, \n2013a \n \nC-Tb \n26 \n- \n16 (61.5%) \n6 (23.1%) \n1 (3.8%) \n- \n1 (3.8%) \n- \n- \nTST \n- \n- \n- \n- \n- \n- \n- \n- \n- \nAggerbeck, \n2013b \n \nC-Tb \n151 \n48 (31.8%) \n- \n- \n- \n- \n- \n- \n- \nTST \n151 \n31 (20.5%) \n- \n- \n- \n- \n- \n- \n- \nAggerbeck, \n2018 \n \nC-Tb \n1188 \n282 (23.7%) \n210 (17.7%) \n90 (7.6%) \n58 (4.9%) \n3 (0.3%) \n5 (0.4%) \n24 (2%) \n15 (1.3%) \nTST \n1190 \n290 (24.4%) \n221 (18.6%) \n81 (6.8%) \n63 (5.3%) \n3 (0.3%) \n4 (0.3%) \n24 (2%)",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 4 gives the frequency of injection site reactions such as itching, pain, rash and swelling in C-Tb and TST arms of each study, plus aggregated C-TST data with an unclear denominator.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-941",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 941",
      "pageStart": 941,
      "pageEnd": 941,
      "target": "../sections/page-941.md",
      "narrative": null,
      "content": "_(no text on this page)_",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A blank page with no text.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-942",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 942",
      "pageStart": 942,
      "pageEnd": 942,
      "target": "../sections/page-942.md",
      "narrative": null,
      "content": "Table 6. Frequencies of adverse events in studies on Diaskintest \n Study \nPopulation \nDiaskintest vs TST \nFever \nOthers \nAksenova, 2009# \nChildren and adolescents with TB \nNA \n2/63 (3.2%) \n \nBarmina, 2018# \nChild and adolescent household contact \nNA \nNA \nAny adverse events:\nBelushkov, 2012 \nChildren suspected of  TB \nNA \n4/88 (4.5%) \nHyper allergic reacti\nBorisova, 2017 \nPregnant women with TB \nNA \nNA \nAny adverse events:\nDotsenko 2015 \nTB care workers \nItching 3/25 (12%) in DST site vs 7/25 (28%) \nin TST site; \nhyperallergic reaction with local \nlymphadenitis, lymphangitis and pain 1/25 \n(4.0",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 6 lists adverse events reported in studies of Diaskintest in children, adults, pregnant women and people with TB or HIV, and Table 7 begins listing fever, headache and dizziness in participants given C-Tb or the tuberculin skin test. Most reactions shown are described as mild to moderate.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-943",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 943",
      "pageStart": 943,
      "pageEnd": 943,
      "target": "../sections/page-943.md",
      "narrative": null,
      "content": "HIV- and 76.6% \nin HIV+ group  \n \nRuhwald, 2017 \nNA \n137/979 (14%) \nNA \nMild-moderate: \n100%",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The end of Table 7, giving the share of mild systemic reactions by HIV status in one study and figures from Ruhwald 2017.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-944",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 944",
      "pageStart": 944,
      "pageEnd": 944,
      "target": "../sections/page-944.md",
      "narrative": null,
      "content": "Table 8. Systemic adverse events in studies on C-TST \nStudy \nSystemic reactions \nLi 2016a \nNo systemic reactions in 6 participants  \nLi 2016b \n9 adverse events in 144 patients. \nAbnormal percentage of eosinophil (n=1); Increased blood \npressure (n=2); fever (n=1); phlebitis (n=2); vasculitis (n=3)  \nXu 2021* \nFever (7.1%); nausea (0.3%); headache (0.7%); allergy (0.2%); \nmuscle pain (0.8%) \n*This paper reported two control trials but only aggregated data were reported.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 8 lists systemic adverse events reported in studies of the C-TST skin test, such as fever, raised blood pressure and phlebitis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-945",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 945",
      "pageStart": 945,
      "pageEnd": 945,
      "target": "../sections/page-945.md",
      "narrative": null,
      "content": "Table 9  \nQuestion: Do TBST cause more adverse reactions compared to tuberculin skin tests? \nSetting: Global \nBibliography: Aggerbeck 2013; Aggerbeck 2018; Aggerbeck 2019; Hoff 2016, Ruhwald 2017; Streltsova 2011 \nCertainty assessment \n№ of patients \nEffect \nCertainty \nImportanc\ne \n№ of \nstudie\ns \nStudy \ndesign \nRisk \nof bias \nInconsistenc\ny \nIndirectnes\ns \nImprecisio\nn \nOther \nconsideration\ns \nTBST \ntuberculi\nn skin \ntests \nRelativ\ne \n(95% \nCI) \nAbsolut\ne \n(95% CI) \nAny injection site reactions \n6 \n5 \nrandomise\nd trials and \n1 NRS \nnot \nseriou\ns \nnot seriousa not serious \nnot \nseriousb \nnone",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 9, a GRADE evidence profile asking whether TB antigen-based skin tests cause more adverse reactions than tuberculin skin tests, finding similar rates of injection site reactions (high certainty) and systemic reactions (moderate certainty).",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p946-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 946,
      "pageEnd": 946,
      "target": "../images/img-p946-1.webp",
      "narrative": "draft",
      "content": "A PRISMA flow diagram of study selection for the safety review. Via databases and registers (Medline 853, EMBASE 1317, CKNI 565, CBLD 282, e-library 405), 746 duplicates were removed, 2676 records screened, 2322 excluded, 354 reports sought, 27 not retrieved, 327 assessed and 304 excluded (mostly for no safety data, n = 261). Via other methods (WHO public call 20, manufacturers 26), 46 reports were sought and 38 assessed; the right edge of the image is cut off. In total 29 studies were included: 15 reports in English, 10 in Russian and 1 in Ukrainian.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-946",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 946",
      "pageStart": 946,
      "pageEnd": 946,
      "target": "../sections/page-946.md",
      "narrative": null,
      "content": "Figures \n \n \nFigure 1. Study selection",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A heading page for the figures, introducing Figure 1 on study selection.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p947-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 947,
      "pageEnd": 947,
      "target": "../images/img-p947-1.webp",
      "narrative": "draft",
      "content": "A forest plot of risk ratios (TBST vs TST) for any injection site reaction, from five C-Tb studies (pooled RR 1.05) and one DST study (Streltsova 2011, RR 3.00). The overall random-effects pooled risk ratio is 1.09 (0.74–1.61).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-947",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 947",
      "pageStart": 947,
      "pageEnd": 947,
      "target": "../sections/page-947.md",
      "narrative": null,
      "content": "Figure 2. Any injection site reactions \nProportion of HIV+: Aggerbeck 2018 (25%), Aggerbeck 2019 (20%); Hoff 2016 (39.5%). Other studies included HIV-negative \nindividuals. Aggerbeck 2018 included children < 5 years old (20%) and 5-17 years old (31%); Ruhwald included children < 5 \nyears old (3.5%) and 5-17 years old (8.8%); Other studies included adults. Hoff 2016, Aggerbeck 2019, and Streltsova, 2011 \nincluded people with TB only.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Figure 2 shows injection site reactions in skin test studies, with notes on the share of participants living with HIV, children and people with TB in each study.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p948-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 948,
      "pageEnd": 948,
      "target": "../images/img-p948-1.webp",
      "narrative": "draft",
      "content": "A forest plot of risk ratios (TBST vs TST) for itching/pruritus, from four C-Tb studies (pooled RR 0.87), one C-TST study (Li 2016b, RR 1.00) and one DST study (Dotsenko 2015, RR 0.43). The overall random-effects pooled risk ratio is 0.86 (0.78–0.96).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p948-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 948,
      "pageEnd": 948,
      "target": "../images/img-p948-2.webp",
      "narrative": "draft",
      "content": "A forest plot of risk ratios (TBST vs TST) for pain, from four C-Tb studies (pooled RR 1.05) and one C-TST study with no events (Li 2016b). The overall random-effects pooled risk ratio is 1.05 (0.88–1.25).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-948",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 948",
      "pageStart": 948,
      "pageEnd": 948,
      "target": "../sections/page-948.md",
      "narrative": null,
      "content": "Figure 3. Itching/ Pruritus \nProportion of HIV+: Aggerbeck 2018 (25%), Aggerbeck 2019 (20%); Hoff 2016 (39.5%). Other studies included HIV-negative \nindividuals. \nAggerbeck 2018 included children < 5 years old (20%) and 5-17 years old (31%); Ruhwald included children < 5 years old \n(3.5%) and 5-17 years old (8.8%); Other studies included adults. Hoff 2016 and Aggerbeck 2019 included people with TB \nonly.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Figure 3 shows itching or pruritus after skin testing, with notes on the HIV, age and TB status of study populations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p949-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 949,
      "pageEnd": 949,
      "target": "../images/img-p949-1.webp",
      "narrative": "draft",
      "content": "A forest plot of risk ratios (TBST vs TST) for rash, from four C-Tb studies (Aggerbeck 2018, Aggerbeck 2019, Hoff 2016, Ruhwald 2017). The overall random-effects pooled risk ratio is 0.80 (0.52–1.23).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-949",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 949",
      "pageStart": 949,
      "pageEnd": 949,
      "target": "../sections/page-949.md",
      "narrative": null,
      "content": "Figure 4. Pain \nProportion of HIV+: Aggerbeck 2018 (25%), Aggerbeck 2019 (20%); Hoff 2016 (39.5%). Other studies included HIV-negative \nindividuals. \nAggerbeck 2018 included children < 5 years old (20%) and 5-17 years old (31%); Ruhwald included children < 5 years old \n(3.5%) and 5-17 years old (8.8%); Other studies included adults. Hoff 2016, Aggerbeck 2019, and Streltsova, 2011 included \npeople with TB only. \n \n \n \n \n \n \n \n \nFigure 5. Rash",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Figure 4 shows pain at the injection site, with notes on study populations, and introduces Figure 5 on rash.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p950-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 950,
      "pageEnd": 950,
      "target": "../images/img-p950-1.webp",
      "narrative": "draft",
      "content": "A forest plot of risk ratios (TBST vs TST) for erythema, from four C-Tb studies, with Hoff 2016 carrying 83% of the weight. The overall random-effects pooled risk ratio is 0.82 (0.67–1.00).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-950",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 950",
      "pageStart": 950,
      "pageEnd": 950,
      "target": "../sections/page-950.md",
      "narrative": null,
      "content": "Proportion of HIV+: Aggerbeck 2018 (25%), Aggerbeck 2019 (20%); Hoff 2016 (39.5%). Other studies included HIV-negative \nindividuals. \nAggerbeck 2018 included children < 5 years old (20%) and 5-17 years old (31%); Ruhwald included children < 5 years old \n(3.5%) and 5-17 years old (8.8%); Other studies included adults. Hoff 2016, Aggerbeck 2019, and Streltsova, 2011 included \npeople with TB only. \n \n \n \n \n \n \n \nFigure 6. Erythema \nProportion of HIV+: Aggerbeck 2018 (25%), Aggerbeck 2019 (20%); Hoff 2016 (39.5%). Other studies included HIV-negative \nindividuals. \nAggerbeck 2018 included children",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes on the study populations accompanying the skin test adverse event figures on rash and the following figure, giving shares of participants with HIV, children and people with TB.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p951-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 951,
      "pageEnd": 951,
      "target": "../images/img-p951-1.webp",
      "narrative": "draft",
      "content": "A forest plot of risk ratios (TBST vs TST) for swelling/oedema, from three C-Tb studies (Aggerbeck 2018, Hoff 2016, Ruhwald 2017). The overall random-effects pooled risk ratio is 1.04 (0.68–1.59).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p951-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 951,
      "pageEnd": 951,
      "target": "../images/img-p951-2.webp",
      "narrative": "draft",
      "content": "A forest plot of risk ratios (TBST vs TST) for vesicle/blister, from four C-Tb studies, with study risk ratios from 0.47 to 1.24. The overall random-effects pooled risk ratio is 0.75 (0.36–1.56).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-951",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 951",
      "pageStart": 951,
      "pageEnd": 951,
      "target": "../sections/page-951.md",
      "narrative": null,
      "content": "Figure 7. Swelling/Oedema \nProportion of HIV+: Aggerbeck 2018 (25%); Hoff 2016 (39.5%). Other studies included HIV-negative individuals. \nAggerbeck 2018 included children < 5 years old (20%) and 5-17 years old (31%); Ruhwald included children < 5 years old \n(3.5%) and 5-17 years old (8.8%); Other studies included adults. Hoff 2016 included people with TB only. \n \nProportion of HIV+: Aggerbeck 2018 (25%), Aggerbeck 2019 (20%); Hoff 2016 (39.5%). Other studies included HIV-negative \nindividuals.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Figure 7 shows swelling or oedema after skin testing, with notes on the HIV, age and TB status of study populations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p952-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 952,
      "pageEnd": 952,
      "target": "../images/img-p952-1.webp",
      "narrative": "draft",
      "content": "A forest plot of risk ratios (TBST vs TST) for induration, from three C-Tb studies (Aggerbeck 2018, Aggerbeck 2019, Ruhwald 2017), indicating more induration events with TBST. The overall random-effects pooled risk ratio is 1.78 (1.36–2.32).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-952",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 952",
      "pageStart": 952,
      "pageEnd": 952,
      "target": "../sections/page-952.md",
      "narrative": null,
      "content": "Aggerbeck 2018 included children < 5 years old (20%) and 5-17 years old (31%); Ruhwald included children < 5 years old \n(3.5%) and 5-17 years old (8.8%); Other studies included adults. Aggerbeck 2019 and Hoff 2016 included people with TB \nonly. \n \n \n \n \nFigure 8. Vesicle \n \n \n \n \n \n \n \n \nFigure 9. Induration \nProportion of HIV+: Aggerbeck 2018 (25%), Aggerbeck 2019 (20%); Hoff 2016 (39.5%). Other studies included HIV-negative \nindividuals. \nAggerbeck 2018 included children < 5 years old (20%) and 5-17 years old (31%); Ruhwald included children < 5 years old \n(3.5%) and 5-17 years old (8.8%); O",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Population notes for Figures 8 and 9 on vesicles and induration after skin testing.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p953-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 953,
      "pageEnd": 953,
      "target": "../images/img-p953-1.webp",
      "narrative": "draft",
      "content": "A summary forest plot of pooled risk ratios (TBST vs PPD) for any injection site reaction and individual reactions: any ISR 1.09 (6 studies), itching/pruritus 0.87, pain 1.05, rash 0.80, erythema 0.82, swelling/oedema 1.04, vesicle/blister 0.75 and induration 1.78 (1.36–2.32). Only induration has a confidence interval entirely above 1, and only itching one entirely below 1.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p953-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 953,
      "pageEnd": 953,
      "target": "../images/img-p953-2.webp",
      "narrative": "draft",
      "content": "A forest plot of the proportion (%) of participants with any systemic adverse event after TBST: five C-Tb studies pooled at 36.1% (26.8–46.4), ranging from 28.5% to 53.0%, and one C-TST study (Li 2016a) with 0 of 6. The overall random-effects pooled proportion is 35.0% (26.4–44.7) across 3033 participants.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-953",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 953",
      "pageStart": 953,
      "pageEnd": 953,
      "target": "../sections/page-953.md",
      "narrative": null,
      "content": "Figure 10. Pooled estimates of the risk for any ISR and individual ISR.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The caption of Figure 10, pooled estimates of the risk of any and individual injection site reactions.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p954-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 954,
      "pageEnd": 954,
      "target": "../images/img-p954-1.webp",
      "narrative": "draft",
      "content": "A forest plot of the proportion (%) of participants with fever after TBST, grouped by test: C-Tb 2.6% (three studies), C-TST 0.0% (two studies with no events) and DST 2.7% (six studies, 0.0% to 7.0%). The overall random-effects pooled proportion is 2.6% (2.0–3.4) across 2478 participants.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-954",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 954",
      "pageStart": 954,
      "pageEnd": 954,
      "target": "../sections/page-954.md",
      "narrative": null,
      "content": "Figure 11. Frequencies of any systemic adverse events  \nProportion of HIV+: Aggerbeck 2018 (25%), Aggerbeck 2019 (20%); Hoff 2016 (39.5%). Other studies included HIV-negative \nindividuals. \nAggerbeck 2018 included children < 5 years old (20%) and 5-17 years old (31%); Ruhwald included children < 5 years old \n(3.5%) and 5-17 years old (8.8%); Other studies included adults. Aggerbeck 2019 and Hoff 2016 included people with TB \nonly. \nFigure 12. Frequencies of fever \nProportion of HIV+: Aggerbeck 2018 (25%), Aggerbeck 2019 (20%); Hoff 2016 (39.5%). Other studies included HIV-negative individuals.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Figures 11 and 12 show frequencies of any systemic adverse events and of fever in skin test studies, with population notes.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p955-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 955,
      "pageEnd": 955,
      "target": "../images/img-p955-1.webp",
      "narrative": "draft",
      "content": "A forest plot of the proportion (%) of participants with headache after C-Tb in six studies, ranging from 3.8% (Aggerbeck 2013) to 17.2% (Aggerbeck 2019). The random-effects pooled proportion is 11.3% (7.8–16.0) across 2723 participants, with high heterogeneity (I² = 79%).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-955",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 955",
      "pageStart": 955,
      "pageEnd": 955,
      "target": "../sections/page-955.md",
      "narrative": null,
      "content": "Aggerbeck 2018 included children < 5 years old (20%) and 5-17 years old (31%); Ruhwald included children < 5 years old (3.5%) and 5-17 \nyears old (8.8%). \nAksenova 2009, Belushkov 2018, Rutkovsky 2020, and Yarovaya, 2019 included only children and adolescents <18 years old. Aksenova \n2009, Slogotskaya, and Streltsova included people with active TB only. \n \n \nFigure 13. Frequencies of headache \nProportion of HIV+: Aggerbeck 2018 (25%), Aggerbeck 2019 (20%); Hoff 2016 (39.5%). Other studies included HIV-negative individuals.  \nAggerbeck 2018 included children < 5 years old (20%) and 5-17 years o",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Population notes for the systemic adverse event figures, including studies limited to children or people with active TB, and the caption for Figure 13 on headache.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-956",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 956",
      "pageStart": 956,
      "pageEnd": 956,
      "target": "../sections/page-956.md",
      "narrative": null,
      "content": "Web Annex B.22. Cost–effectiveness of Mycobacterium tuberculosis \nantigen-based skin tests: a systematic review \n \nReview team: Lara Goscé1, Finna Parkinson1, Kasim Allel,1,2 Elena Surkova3, Irina Kontsevaya4, Ting Ting Wang1, \nVictoria Liu2, Yohhei Hamada1, Peter White5, Molebogeng X Rangaka1  \n \n1Institute for Global Health, University College London, UK \n2London School of Hygiene and Tropical Medicine, UK \n3Royal Brompton Hospital. Part of Guy's and St Thomas' NHS Foundation Trust London, UK \n4Research Center Borstel, Germany \n5Imperial College, UK \n \nTechnical advisor: Peter White1, Ibrahi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title page of Web Annex B.22, a systematic review by a team mainly from University College London on the cost-effectiveness of Mycobacterium tuberculosis antigen-based skin tests, prepared for a February 2022 WHO guideline meeting.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-957",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 957",
      "pageStart": 957,
      "pageEnd": 957,
      "target": "../sections/page-957.md",
      "narrative": null,
      "content": "Abbreviations  \nBCG: Bacillus Calmette-Guérin \nCBA: Cost-benefit analysis \nCUA: Cost-utility analysis  \nCEA: Cost-effective analysis \nDOT: Directly observed therapy \nHICs: High-income countries \nHIV: Human immunodeficiency virus \nICER: incremental cost-effectiveness ratios \nIGRA: interferon-gamma release assays \nLMICs: Low- and middle-income countries \nPPD: purified protein derivative \nPRISMA – Preferred Reporting Items for Systematic Reviews and Meta-Analyses \nPSA: probabilistic sensitivity analysis \nQALYS: Quality Adjusted Life Years \nQFT: QuantiFERON-TB Gold \nTB: Tuberculosis \nTBI: Latent t",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A list of abbreviations used in the cost-effectiveness review, such as BCG, ICER, IGRA, TBST, TPT and TST.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-958",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 958",
      "pageStart": 958,
      "pageEnd": 958,
      "target": "../sections/page-958.md",
      "narrative": null,
      "content": "BACKGROUND \n \nWith an estimated 1.7 billion people infected with tuberculosis infection (TBI), progression of infection to \nactive TB disease poses a large public health risk1. Prominent risk factors of this transition include clinical risk \nfactors such as human immunodeficiency virus (HIV), diabetes, undernutrition, and contextual or societal risk \nfactors in the most vulnerable and disadvantaged populations including, TB case contacts especially children, \nimmigrants2. WHO recommends TB preventive treatment (TPT) to disrupt disease progression. Testing for TB \ninfection is recommended to gu",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives the background to the review, noting about 1.7 billion people with TB infection and WHO's recommendation of preventive treatment guided by testing where possible. It contrasts the tuberculin skin test, which needs two visits and is less specific after BCG vaccination, with blood-based IGRA tests, and introduces newer specific skin tests.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-959",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 959",
      "pageStart": 959,
      "pageEnd": 959,
      "target": "../sections/page-959.md",
      "narrative": null,
      "content": "sensitivity for TBI, especially in resource-constrained settings and contexts where BCG vaccination is \nimplemented routinely. 8,9. The diagnostic accuracy of the tests alone, however, is not sufficient evidence for \nthe recommendation of their use in TBI testing guidelines. To provide sufficient evidence for the intervention, \ncost-effectiveness must also be considered to ensure that the tests are affordable and feasible to implement.  \n \nThe present report provides a systematic review of the literature on the cost-effectiveness of the new tests \ncompared to currently available tests for TBI.",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains that accuracy alone does not justify recommending the new skin tests, so their cost-effectiveness must be examined. It states the research question and four objectives: two systematic reviews of costs and cost-effectiveness of the new and current tests, and modelling of cost-effectiveness across settings.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-960",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 960",
      "pageStart": 960,
      "pageEnd": 960,
      "target": "../sections/page-960.md",
      "narrative": null,
      "content": "LAYOUT OF THE REPORT \n \nThe report is presented in two parts.  \nSection A: Abstract, methods and results for Objectives 1-2  \nSection B: Abstract, methods and result for Objectives 3-4.  \n \nThis document contains Section A. Section B is submitted as a separate attachment.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains that the report has two parts, with this document containing Section A on the systematic reviews and Section B on the modelling provided separately.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-961",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 961",
      "pageStart": 961,
      "pageEnd": 961,
      "target": "../sections/page-961.md",
      "narrative": null,
      "content": "Systematic literature reviews  \n \nAbstract \nBackground: The tuberculin skin test (TST) and interferon-gamma release assays (IGRA) are the currently used \ntests for identifying individuals with TB infection that should be offered TB preventive treatment, however \nchallenges around access and implementation have limited their use. Novel Mycobacterium tuberculosis \nspecific skin tests (TBST) such as the DiaskinTest, ESAT6-CFP10 (now called C-TST) C-Tb, and DPPD have been \ndeveloped in recent years, these may provide accurate and scalable options. We conducted a systematic \nreview of the economic",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The abstract of the systematic reviews, explaining that novel TB-specific skin tests may offer accurate, scalable alternatives to the tuberculin skin test and IGRA. It reports finding only eight economic studies of the new tests, seven on Diaskintest in Russia and one in Brazil, with low unit costs but generally poor study quality, and calls for more high-quality studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-962",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 962",
      "pageStart": 962,
      "pageEnd": 962,
      "target": "../sections/page-962.md",
      "narrative": null,
      "content": "Methods \n \nWe performed two different literature reviews. \n(a) a primary review looking at the costs and cost-effectiveness of the novel skin tests (registered on \nPROSPERO: CRD42021275585) \n(b) a secondary review for the cost-effectiveness of PPD-TST and IGRAs tests (registered on PROSPERO: \nCRD42021275684) \n \nFirstly, we reviewed the literature on the cost and cost-effectiveness of the novel tests. Due to the anticipated \nlack of evidence, a secondary review of the literature surrounding the cost and cost- effectiveness of the standard \nPPD-TST and IGRA was also conducted to supplement the p",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the methods of two registered literature reviews, one on costs and cost-effectiveness of the new skin tests and a secondary one on the tuberculin skin test and IGRA. It lists the English, Chinese and Russian databases searched, the three concept groups of search terms, and the use of manufacturer data and a WHO public call for data.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-963",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 963",
      "pageStart": 963,
      "pageEnd": 963,
      "target": "../sections/page-963.md",
      "narrative": null,
      "content": "We did not have an exclusion criterion for specific populations, and all articles had to be published in English, \nChinese or Russian and relate to TBI in humans. Papers had to be full original economic evaluations, meaning \nhaving clear measures of costs, outcomes, and sufficient incremental analysis using the healthcare perspective. \nNo time restriction was imposed for the primary review on novel skin tests, whereas we searched those articles \npublished between 2011 to July 2021 for the secondary review on PPD-TST or IGRA tests. \n \nStudy selection & data extraction \nStudies were included if",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Sets out inclusion criteria requiring full economic evaluations from a health care perspective, published in English, Chinese or Russian, and the test products covered. It describes how reviewers screened papers in each language, extracted data, converted costs to 2021 US dollars and assessed quality with the Drummond checklist.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-964",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 964",
      "pageStart": 964,
      "pageEnd": 964,
      "target": "../sections/page-964.md",
      "narrative": null,
      "content": "Results \n \nIdentification, screening, eligibility, and inclusion \n \nFor the primary search (systematic review 1) on cost-effectiveness of novel skin tests, 367 records were \nidentified for full text screening (103 written in English/Chinese, and 264 in Russian) of which only 8 were \nrelevant to the research question (only one of those written in English and 7 in Russian) (See PRISMA flow in \nFigure 1). For the secondary search on cost-effectiveness of PPD-TST or IGRA tests, 56 out of 407 records were \nchosen for full-text screening and 29 papers were selected for analysis (see PRISMA flow in F",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports the search results, with 8 relevant studies on the new skin tests (mostly in Russian) and 29 studies on the tuberculin skin test and IGRA selected for analysis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-vfig-p965",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 965,
      "pageEnd": 965,
      "target": "../figures/vfig-p965.png",
      "narrative": "draft",
      "content": "A PRISMA flowchart for the primary systematic review of novel tests, with two streams. English and Chinese databases (EMBASE 1317, MEDLINE 1424, Chinese Biomedical Literature Database 282, China National Knowledge Infrastructure 656) gave 2271 records after removing duplicates; 2168 did not report new skin tests, 103 went to a second screening and 1 record was included. The e-library search gave 405 records; 141 were not relevant, 264 had full-text review, 257 did not report health economics analysis and 7 were included.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-965",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 965",
      "pageStart": 965,
      "pageEnd": 965,
      "target": "../sections/page-965.md",
      "narrative": null,
      "content": "Figure 1. PRISMA flowchart for primary systematic review (Novel tests) \n \nNote: no eligible study was identified among papers shared by test manufacturers and those identified though \na WHO public call for data. \n \n \n \n \n \n \n \n \n \n \n \n \nEnglish and Chinese languages searches       \nRussian language search      \n405 records in e-library \n \n141 records deemed not relevant \n \n264 records for full text review \n \n257 records do not report health \neconomics analysis \n \n7 records included \n1317 records in EMBASE \n1424 records in MEDLINE \n282 Chinse Biomedical Literature Database \n656 China National K",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Figure 1, a PRISMA flowchart for the primary review of novel skin tests, showing searches in English, Chinese and Russian databases leading to the included studies, with a note that no eligible study came from manufacturers or the WHO call for data.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-vfig-p966",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 966,
      "pageEnd": 966,
      "target": "../figures/vfig-p966.png",
      "narrative": "draft",
      "content": "A PRISMA flowchart for the secondary systematic review (PPD-TST and IGRA), followed by body text: 652 records in EMBASE and 272 in MEDLINE; 407 records published since July 2011 after removing duplicates; 350 excluded after title and abstract screening; 57 reports assessed in full text, of which 28 were excluded for not reporting a clear health system perspective and costs and/or ICERs; 29 studies included in the review.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-966",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 966",
      "pageStart": 966,
      "pageEnd": 966,
      "target": "../sections/page-966.md",
      "narrative": null,
      "content": "Figure 2. PRISMA flowchart for the secondary systematic review (PPD-TST and IGRA) \n \n \nDescription of the articles found in the primary review (specific TBST)  \n \nEight economic evaluations were conducted from middle-income countries (Russia and Brazil), these mostly \ncompared DiaskinTest against PPD-TST (Table A2, supplementary). One study met most of the Drummond’s \nchecklist factors and reported all the required information. Steffen et al (2020) 12 studied the cost-effectiveness \nof two novel TBST (Diaskintest and C-TST) for people living with HIV (PLHIV) in Brazil compared to PPD-TST and",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Figure 2, a PRISMA flowchart for the secondary review of the tuberculin skin test and IGRA, followed by a description of the eight studies on the new tests, mostly Russian studies in children comparing Diaskintest with the tuberculin skin test. It singles out a Brazilian Markov model study in people living with HIV as the only one meeting most quality criteria.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-967",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 967",
      "pageStart": 967,
      "pageEnd": 967,
      "target": "../sections/page-967.md",
      "narrative": null,
      "content": "We evaluated study quality using the Drummond’s checklist for health economic evaluations (supplementary \nmaterial, Table A7). On average, we identified that most studies did not present any sensitivity analysis, and \nthat the information on relevant parameters such as time horizon, type of model used, year of the evaluation, \nand discounting rate were unclear, or not stated. Overall, study quality was concerning. \n \nDescription of the articles in the secondary review (PPD-TST and IGRA)  \n \nOf the 29 papers included in the review, 8 were based in the United Kingdom (UK), 6 in the United States",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes that most studies of the new skin tests lacked sensitivity analyses and clear model details, so their quality was concerning. It then describes the 29 studies on the tuberculin skin test and IGRA, mostly from high-income countries, which used various models and were generally of high quality.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-968",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 968",
      "pageStart": 968,
      "pageEnd": 968,
      "target": "../sections/page-968.md",
      "narrative": null,
      "content": "compared strategies to DiaskinTest only. The dominance of DiaskinTest was very sensitive to unit costs of \nDiaskinTest which is highly uncertain due to using market value and hence varies widely by health system and \ncountry.  The rest of the articles found that the cost of the DiaskinTest kit ranged from $1.29 to $3.49 and that \nit was not very sensitive to unit costs after employing univariate sensitivity analyses (if measured). All these \nstudies, apart from Steffen et al, found that DiaskinTest was cost-effective using a wide-range of methods, \nranging from a cost-effectiveness ratio of 2.",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarizes findings that Diaskintest kits cost about $1.29–3.49 and that almost all studies judged it cost-effective compared with the tuberculin skin test, with results sensitive to unit costs. Table 1 gives unit costs of Diaskintest and C-TST from the primary review in 2021 US dollars.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-969",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 969",
      "pageStart": 969,
      "pageEnd": 969,
      "target": "../sections/page-969.md",
      "narrative": null,
      "content": "Most studies evaluated the costs and cost-effectiveness of PPD-TST and IGRA with wide-ranging study \npopulations including PLHIV, immunocompromised people other than PLHIV, immigrants/migrants, healthcare \nworkers  and different methods were used, especially time horizons. 22 studies were set in low TB burden \ncountries (UK, USA, Canada, Norway, Oman, Spain), 4 in lower moderate (Brazil, Japan, Singapore), 1 in upper-\nmoderate (Hong Kong) and two in a high burden country (South Africa). Four studies were based in low-and-\nmiddle income countries, whereas 25 in high-income countries.  Most stud",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarizes the secondary review, noting that most studies were in low-burden, high-income countries and that testing was more likely to be cost-effective in high-risk groups or high-burden settings, without consensus on which test to use. It reports that studies in people living with HIV generally found IGRA more cost-effective, with one finding Diaskintest most cost-effective, and describes reasons for variation in unit costs.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-970",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 970",
      "pageStart": 970,
      "pageEnd": 970,
      "target": "../sections/page-970.md",
      "narrative": null,
      "content": "Table 2: Unit costs of PPD-TST (2021 USD) \n \n \nPPD-TST COSTS \n \n \nStudy ID \nCountry \nTest \nKit \nStaff \nTime \nConsu\nmable \nOverh\neads \nLabora\ntory \nDisposa\nbles \nUnit Costs \nLinas (2011) 22 \nUSA \n$2.88 \n$23.92 \n$26.81 \ndel Campo, \n(2012) 25 \nSpain \n✓ \n✓ \n✓ \n$60.35 \nEralp, (2012) a 26 UK \n$31.12 \nMandalakas, \n(2013) 27  \nSouth \nAfrica \n✓ \n✓ \n✓ \n$21.92̶   \n99.13 \nKim (2018) 28 \nSouth \nAfrica \n \n$1.75 \n$1.70 \n$0.08 \n \n$4.40 \nPareek (2013) b 29 UK \n$68.34 \nSteffen (2013) 7 \nUSA \n$5.83 \n$3.79 \n$2.84 \n$0.10 \n$12.56 \nSwaminath, \n(2013) \nUSA \n$48.99 \nVerma (2013) 30 \nCanada \n$16.81 \n$26.77 \n$43.58 \nCap",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 2 lists unit costs of the tuberculin skin test from studies in different countries, broken down into kit, staff time, consumables and other components, in 2021 US dollars.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-971",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 971",
      "pageStart": 971,
      "pageEnd": 971,
      "target": "../sections/page-971.md",
      "narrative": null,
      "content": "Notes: A tick mark (✓) stands for those articles that mention they included certain cost components but did not explicitly state the figures \nand only included the total costs per test diagnostic. There were no further details provided but only the overall costs of the test If no tick (✓) \nis observed. \n a Calculated from the National Institute for Health and Clinical Excellence. Clinical diagnosis and management of Tuberculosis, and measures \nfor its prevention and control. NICE Clinical Guidelines 2011 \nb Calculated from the National Collaborating Centre for Chronic Conditions. Tuberculosis:",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes explaining Table 2, followed by part of a table giving unit costs of IGRA tests such as QuantiFERON and T-SPOT from studies in several countries.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-972",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 972",
      "pageStart": 972,
      "pageEnd": 972,
      "target": "../sections/page-972.md",
      "narrative": null,
      "content": "Notes A tick mark (✓) stands for those articles that mention they included certain cost components but did not explicitly state the figures and \nonly included the total costs per diagnostic test. There were no further details provided but only the overall costs of the test If no tick (✓) is \nobserved.     \na Calculated from the National Institute for Health and Clinical Excellence. Clinical diagnosis and management of Tuberculosis, and measures \nfor its prevention and control. NICE Clinical Guidelines 2011 \nb Calculated from the National Collaborating Centre for Chronic Conditions. Tuberculosi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes to the cost tables and the rest of the IGRA unit cost table, covering studies from Canada, Brazil, Singapore and the USA.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-973",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 973",
      "pageStart": 973,
      "pageEnd": 973,
      "target": "../sections/page-973.md",
      "narrative": null,
      "content": "Interpretation \n \nConsidering the same factors (ingredients) for economic costing, costs of novel skin tests (DiaskinTest=$5.07, C-\nTST=$9.9912) were substantially lower than those of IGRA (average cost was $89.33; ranging from $22.1712 to \n$247.1227) and the TST (average cost=$37.84, ranging from $4.4028 to $181.6324). However, in isolation, \nobserved low costs of DiaskinTest and C-TST compared to TST and IGRA are not sufficient evidence for likely \neconomic impact, particularly given the limited number of studies. More high-quality studies are needed (only \none high-quality study found) cons",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Interprets the findings, noting that novel skin tests cost far less than IGRA or the tuberculin skin test but that low costs alone, from few and generally poor-quality studies in Russia and Brazil, do not show likely economic impact. It states that a modelling study was undertaken with results in Part B.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-974",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 974",
      "pageStart": 974,
      "pageEnd": 974,
      "target": "../sections/page-974.md",
      "narrative": null,
      "content": "References \n \n1. \nHouben RM, Dodd PJ. The global burden of latent tuberculosis infection: a re-estimation using \nmathematical modelling. PLoS medicine 2016; 13(10): e1002152. \n2. \nWorld Health Organization. Global tuberculosis report 2020: World Health Organization; 2020. \n3. \nFarhat M, Greenaway C, Pai M, Menzies D. False-positive tuberculin skin tests: what is the absolute \neffect of BCG and non-tuberculous mycobacteria? The International Journal of Tuberculosis and Lung Disease \n2006; 10(11): 1192-204. \n4. \nSAGE. Report on BCG vaccine use for protection against mycobacterial infections incl",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The start of the reference list for the cost-effectiveness review, citing work on the burden of TB infection, WHO reports and economic studies, including some in Russian.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-975",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 975",
      "pageStart": 975,
      "pageEnd": 975,
      "target": "../sections/page-975.md",
      "narrative": null,
      "content": "16. \nСолодун ИЮ, Эва ХМ, Башлакова ЕЕ, Ермолаева ТН, Давыдовская МВ, Евдошенко ЕП. Клинико-\nэкономический анализ применения метода диагностики туберкулезной инфекции у детей и подростков \nс использованием аллергена туберкулезного рекомбинантного. Проблемы стандартизации в \nздравоохранении 2017; (3-4). \n17. \nЧугаев Ю, Камаева Н, Цветков А, Кудлай Д, Черняев И. ИННОВАЦИОННЫЕ РЕКОМБИНАНТНЫЕ \nТЕХНОЛОГИИ ВЫЯВЛЕНИЯ И ДИАГНОСТИКИ ТУБЕРКУЛЕЗА У ДЕТЕЙ И ПОДРОСТКОВ: ДОСТИЖЕНИЯ И \nПРОБЛЕМЫ. Pediatriya named after GN Speransky 2020; 96(6). \n18. \nTasillo A, Salomon JA, Trikalinos TA, Horsburgh CR, Marks SM",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Part of the reference list, citing Russian-language economic studies of Diaskintest and English-language studies on IGRA and tuberculin skin test screening of migrants and other groups.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-976",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 976",
      "pageStart": 976,
      "pageEnd": 976,
      "target": "../sections/page-976.md",
      "narrative": null,
      "content": "30. \nVerma G, Chuck A, Jacobs P. Tuberculosis screening for long-term care: a cost-effectiveness analysis. \nThe International journal of tuberculosis and lung disease 2013; 17(9): 1170-7. \n31. \nLa’Marcus TW, Coleman MS, de la Motte Hurst C, et al. A cost-benefit analysis of a proposed overseas \nrefugee latent tuberculosis infection screening and treatment program. BMC public health 2015; 15(1): 1-14. \n32. \nNijhawan AE, Iroh PA, Brown LS, Winetsky D, Porsa E. Cost analysis of tuberculin skin test and the \nQuantiFERON-TB Gold In-tube test for tuberculosis screening in a correctional setting in D",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The end of the reference list, citing cost-effectiveness studies of TB infection screening in long-term care, refugees, migrants and other populations.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-977",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 977",
      "pageStart": 977,
      "pageEnd": 977,
      "target": "../sections/page-977.md",
      "narrative": null,
      "content": "SUPPLEMENTARY DOCUMENTATION  \n \nSection A:  Systematic literature reviews \n \nI. \nLiterature review \n \nFor the primary systematic literature review (new skin tests), the search included papers from inception until 30 \nJuly 2021. Since the previous review included papers published until 20 October 2020, our updated search \nexcluded studies published before October 2020.  \n \nTable A1.1. Search strategy for the primary systematic literature review (nov \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \nOther \ndatabases: \n e-library (www.e-library.ru) for Russian literature  \nSearch Terms: Диаскинтест* или «Аллерген",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The start of the supplementary documentation, describing the search period for the primary review and showing the search strategy terms for TB, diagnostic tests and cost-effectiveness.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-045",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table A1.1. Search strategy for the primary systematic literature review (nov",
      "pageStart": 977,
      "pageEnd": 977,
      "target": "../tables/table-a1-1.csv",
      "narrative": "draft",
      "content": "Numbered search lines used in the primary systematic literature review of novel skin tests for TB infection, combining tuberculosis, recombinant antigen and skin-test terms with Boolean operators.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-978",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 978",
      "pageStart": 978,
      "pageEnd": 978,
      "target": "../sections/page-978.md",
      "narrative": null,
      "content": "Study Characteristics \nKulikov 2009 14 \nAksenova 2011 13 \nYagudina 2013 15 \nCountry Setting \nRussia \nRussia \nRussia \nYear of Cost Valuation \nNot Stated \n2010 (Assumption) \nNot stated. \nCurrency \nRubles \nRubles \nRubles  \nStudy Population \nChildren and \nAdolescents \nChildren \nChildren and Adolescents \nIndex Diagnostic Test \nStrategies \nDiaskinTest \n  \nDiaskinTest \nDiaskinTest \nTST + DiaskinTest  \nAlternative Provided \nTST \nTST \nTST \nType of economic \nevaluation \nCEA \nCost Analysis \nCEA \nSource of costing \nStudies varying \nDiaskinTest costs in \nrelation to ICER of \nMantoux. \nEmpirical Data \nColle",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table A2 extracts data from Russian economic studies of Diaskintest in children and adolescents, giving cost year, currency, strategies compared and unit costs.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-979",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 979",
      "pageStart": 979,
      "pageEnd": 979,
      "target": "../sections/page-979.md",
      "narrative": null,
      "content": "Continuation of Table A2 \n \nStudy Characteristics \nMoiseeva 2014  \nSolodun 2017 16 \nSinitsyn 2018 \nCountry Setting \nRussia \nRussia \nRussia \nYear of Cost \nValuation \n2013 \nNot stated. \n2015 \nCurrency \nRubles \nRubles \nRubles \nStudy Population \nChildren  \nChildren and \nAdolescents \nPeople living with HIV \nIndex Diagnostic Test \nStrategies \nDiaskinTest \n1. TST+ DiaskinTest \n1) DiaskinTest \n2) TST + DiaskinTest  \nDiaskinTest \nAlternative Provided \nTST \nTST \nDo nothing. \nType of economic \nevaluation \nCEA/CUA \nDepends on \neffectiveness \nmeasure. \nCEA \nCBA \nSource of costing \nIn accordance with \nthe p",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of Table A2 with data from further Russian Diaskintest studies, including one in people living with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-980",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 980",
      "pageStart": 980,
      "pageEnd": 980,
      "target": "../sections/page-980.md",
      "narrative": null,
      "content": "Continuation of Table A2 \n \nStudy Characteristics \nChugaev 2020 17 \nSteffen (2020) 12 \nCountry Setting \nRussia \nBrazil \nYear of Cost Valuation \n2013 for TST \n2020 \n2019 for DiaskinTest \nCurrency \nRubles \nUSD \nStudy Population \nChildren \nAdults living with HIV \nIndex Diagnostic Test \nStrategies \nDiaskinTest \n1.  TST, 2.  QFT-GIT, e. EC test \n  \n  \nAlternative Provided \nTST \nDiaskinTest \nType of economic evaluation \nCost analysis \nCUA \nSource of costing \nRetrospective cohort study. \nMinistry of Health and \nPrimary outcome \nCost of strategy and cost of additional \ndiagnosis of TB. \nICER: $/QALY g",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of Table A2 with data from a Russian study in children and a Brazilian cost-utility study in adults living with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-981",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 981",
      "pageStart": 981,
      "pageEnd": 981,
      "target": "../sections/page-981.md",
      "narrative": null,
      "content": "Table A3: Data extraction results for all the entries from the secondary systematic review \n \nStudy Characteristics \nLinas (2011) 22 \nPareek (2011)38 \ndel Campo (2012) 25 \nCountry Setting \nUSA \nUK \nSpain \nYear of Cost Valuation \n2011 \n2010 \n2012 \nCurrency \nUSD \nGBP \nEuros \nStudy Population \n1) Immigrant/Migrants \n2) Immunocompromised \n3) Vulnerable \nRecently arrived \nimmigrants \n(<16y.o. and 16- \n35y.o.) \nHealthcare workers. \nIndex Diagnostic Test \nStrategies \n1. QFT \n2. TST \nIGRA \n1. TST (10mm) \n2. QFT \n3. TST (5mm) + QFT \nAlternatives Provided \nNo screening. \nNo screening \nTST (5mm) \nType of",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table A3 begins extracting data from the secondary review, covering studies from the USA, UK and Spain on tuberculin skin test and IGRA strategies in migrants, immunocompromised people and health workers.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-046",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table A3: Data extraction results for all the entries from the secondary systematic review",
      "pageStart": 981,
      "pageEnd": 990,
      "target": "../tables/table-a3.csv",
      "narrative": "draft",
      "content": "Data extracted from the economic evaluations of TST and IGRA included in the secondary systematic review. For each study (three per block), rows give the country, year of cost valuation, currency, study population, index test strategies, alternative, type of economic evaluation, source of costing, primary outcome, model type, time horizon, discounting, sensitivity analyses, key variables and willingness-to-pay threshold.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-982",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 982",
      "pageStart": 982,
      "pageEnd": 982,
      "target": "../sections/page-982.md",
      "narrative": null,
      "content": "Continuation of Table A3 \n \nStudy Characteristics \nEralp. (2012) 26 \nShah (2012) 39 \nMandalakas (2013) 27 \nCountry Setting \nUK \nUSA \nSouth Africa \nYear of Cost Valuation \n2011 \n2012 \n2009 \nCurrency \nGBP \nUSD \nUSD \nStudy Population \nHealthcare Workers \nIndividuals at primary \nhealth clinic with \npositive TST. \nChildren (0-2 and 3-\n5y.o.) contacts. \nIndex Diagnostic Test \nStrategies \n1. QFT \n2. TST \nTST + QFT-GIT \n1) TST \n2) TST+QFT. \n3) TST-QFT. \n4) TST+ T-SPOT \n5) TST- T-SPOT 6) QFT. \n7) T-SPOT \nAlternative Provided \nTST + QFT \nTST \nNo Screening \nType of economic \nevaluation \nCEA \nCEA \nCEA \nSo",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of Table A3 with data from UK, US and South African studies of TB infection testing in health workers, clinic attendees and child contacts.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-983",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 983",
      "pageStart": 983,
      "pageEnd": 983,
      "target": "../sections/page-983.md",
      "narrative": null,
      "content": "Continuation of Table A3 \n \nStudy Characteristics \nPareek (2013) 29 \nSteffen (2013) 7 \nSwaminath (2013) 40 \nCountry Setting \nUK \nUSA \nUSA \nYear of Cost Valuation \n2011 \n2012 \nNot stated. \nCurrency \nGBP \nUSD \nUSD \nStudy Population \nRecently arrived \nimmigrants (≤35 y.o.). \n35y.o. close contacts. \nImmunosuppressed with IBD \nIndex Diagnostic Test \nStrategies \n1. T-SPOT \n2. QFT \n3. TST + T-SPOT \n4. TST + QFT \n1. QFT \n2. TST + QFT \nQFT \nAlternative Provided \nTST \nTST \nTST \nType of economic \nevaluation \nCEA \nCEA \nCEA \nSource of costing \nPublished Literature \nPublished Literature. \nPublished Literatu",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of Table A3 with data from UK and US studies on immigrants, close contacts and immunosuppressed patients.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-984",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 984",
      "pageStart": 984,
      "pageEnd": 984,
      "target": "../sections/page-984.md",
      "narrative": null,
      "content": "Continuation of Table A3 \n \n \nStudy Characteristics \nVerma (2013) 30 \nCapocci, (2015) 21 \nWingate (2015) 31 \nCountry Setting \nCanada \nUK \nUSA \nYear of Cost Valuation \n2012 \n2012 \n2012 \nCurrency \nCanadian Dollars \nEuro \nUSD \nStudy Population \n>65y.o. in long term \ncare \nAdults living with HIV \nPre arrival refugees \nto USA. \nIndex Diagnostic Test \nStrategies \nTST \n1. QFT + TST \n2. QFT (for higher risk) \n3. QFT (for all) \nTST \nAlternative Provided \nNo screening \nNo Screening \nNo Screening \nType of economic \nevaluation \nCEA \nCEA/CUA \nCBA \nSource of costing \nPublished Literature \nPublished Literatu",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of Table A3 with data from studies in Canada, the UK and the USA on older people in long-term care, adults with HIV and refugees.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-985",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 985",
      "pageStart": 985,
      "pageEnd": 985,
      "target": "../sections/page-985.md",
      "narrative": null,
      "content": "Continuation of Table A3 \n \nStudy Characteristics \nAuguste. (2016)19 \nNijhawan (2016) 32 \nCampbell (2017) 41 \nCountry Setting \nUK \nUSA \nCanada \nYear of Cost Valuation \n2012 \n2013 \n2016 \nCurrency \nEuro \nUSD \nCanadian Dollars \nStudy Population \n1) Children \n2) Immunocompromised \n3) Recently Arrived \nImmigrants \n4) General Population. \nAdults entering jail. \nPre-arrival \nrefugees to USA. \nIndex Diagnostic Test \nStrategies \n1. IGRA \n2. TST + IGRA \n3. Simultaneous. \nQFT \n1. TST \n2. IGRA \n3. TST + IGRA \nAlternative Provided \nTST \nTST \nNo Screening \nType of economic \nevaluation \nCUA \nCost Analysis.",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of Table A3 with data from UK, US and Canadian studies covering children, immunocompromised people, immigrants, prisoners and refugees.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-986",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 986",
      "pageStart": 986,
      "pageEnd": 986,
      "target": "../sections/page-986.md",
      "narrative": null,
      "content": "Continuation of Table A3 \n \n \nStudy Characteristics \nHaukaas (2017) 33 \nMullie (2017) 34 \nTasillo, (2017) 18 \nCountry Setting \nNorway \nCanada \nUSA \nYear of Cost Valuation \n2013 \n2015 \n2015 \nCurrency \nEuro \nCanadian Dollars \nUSD \nStudy Population \nRecently arrived \nimmigrants <35y.o. \nHCW with negative \nTST at time of \nemployment. \nUS-born or migrants \nliving with or without \ncomorbidities. \nIndex Diagnostic Test \nStrategies \n1. QFT (for those with \nrisk factors) \n2. TST + QFT \n3. QFT (for all) \nQFT-GIT \n1. TST \n2. IGRA \n3. IGRA + TST \n4. IGRA - TST \nAlternative Provided \nNo Screening \nTST \nNo",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of Table A3 with data from Norwegian, Canadian and US studies on immigrants, health workers and people with comorbidities.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-987",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 987",
      "pageStart": 987,
      "pageEnd": 987,
      "target": "../sections/page-987.md",
      "narrative": null,
      "content": "Continuation of Table A3 \n \nStudy Characteristics \nAbubakar (2018) 24 \nLi (2018) 42 \nSohn (2018) 35 \nCountry Setting \nUK \nHong Kong \nJapan \nYear of Cost \nValuation \nNot Stated. \nNot stated. \n2015 \nCurrency \nGBP \nUSD \nUSD \nStudy Population \n1. Recent immigrants. \n2. Contacts \nElderly (>65y.o.) at \nadmission to \nresidential care \nhome. \nAdolescents (13- \n18y.o.) contacts. \nIndex Diagnostic Test \nStrategies \n1. T-SPOT.TB 2. QFT-GIT \n3. TST (varying cut offs) \n 4. Confirm positive or \nnegative T-SPOT or QFT-GIT \nafter TST \nIGRA \n1. TST + QFT \n2. QFT \nAlternative Provided \nNo Screening \nNo screenin",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of Table A3 with data from studies in the UK, Hong Kong and Japan on immigrants, contacts, older people in care homes and adolescents.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-988",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 988",
      "pageStart": 988,
      "pageEnd": 988,
      "target": "../sections/page-988.md",
      "narrative": null,
      "content": "Continuation of Table A3 \n \nStudy Characteristics \nCampbell (2019a) 36 \nCampbell (2019b) 23 \nLoureiro (2019) 37 \nCountry Setting \nCanada \nCanada \nBrazil \nYear of Cost \nValuation \n2016 \n2016 \n2016 \nCurrency \nCanadian Dollars \nCanadian Dollars \nUSD \nStudy Population \nPre-arrival immigrants. \nMigrants with \neither late-stage CKD \nor beginning \ndialysis. \nPrimary HCW. \nIndex Diagnostic Test \nStrategies \n1. QFT-GIT \n2. TST + QFT-GIT \n3. TST (>10mm) \n1. QFT-GIT \n2. TST (>10mm) \n1. TST (>10mm) \n2. TST (>10mm) + QFT \n3. TST (>5mm) + QFT \nQFT \nAlternative Provided \n(Baseline for Incremental \nAnalysis)",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of Table A3 with data from Canadian and Brazilian studies on immigrants, migrants with kidney disease and health workers.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-989",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 989",
      "pageStart": 989,
      "pageEnd": 989,
      "target": "../sections/page-989.md",
      "narrative": null,
      "content": "Continuation of Table A3 \n \nStudy Characteristics \nPng (2019) 43 \nAl Abri (2020) 44 \nJo (2020) 20 \nCountry Setting \nSingapore \nOman \nUSA \nYear of Cost \nValuation \n2016 \n2020 \n2018 \nCurrency \nUSD \nUSD \nUSD \nStudy Population \nHCW. \n20-year-old recent \nimmigrants. \n(1) non–US-born, \n(2) living with diabetes, \n(3) HIV-positive, \n(4) experiencing recent \nhomelessness, \n(5) incarcerated \nIndex Diagnostic \nTest Strategies \nQFT-GIT either annually or \nevery three years for \ncombination of 1) new hires 2) \nhigh risk 3) \ninternational 4) \nuniversal. \n1. QFT \n2. TST \n3. CXR \nAll with varying \ntreatments.",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of Table A3 with data from studies in Singapore, Oman and the USA on health workers, immigrants and various high-risk groups.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-990",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 990",
      "pageStart": 990,
      "pageEnd": 990,
      "target": "../sections/page-990.md",
      "narrative": null,
      "content": "Continuation of Table A3 \n \nStudy Characteristics \nSteffen (2020) 12 \nKim (2018) 28 \nCountry Setting \nBrazil \nSouth Africa \nYear of Cost \nValuation \n2020 \n2016 \nCurrency \nUSD \nUSD \nStudy Population \nAdults living with HIV \nHIV+ pregnant women \nIndex Diagnostic Test \nStrategies \n1. TST \n2. QFT-GIT \n3. EC Test \n1. TST \n \nAlternative Provided \n(Baseline for \nIncremental Analysis) \nDiaskinTest \nQFT-GIT \nType of economic \nevaluation \nCUA \nCEA \nSource of costing \nMinistry of Health and \nmarket value. \nNational Health \nLaboratory service \nPrimary outcome \nICER: $/QALY gained \nICER: $/DALY averted \nTy",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of Table A3 with data from studies in Brazil on adults with HIV and in South Africa on pregnant women with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-991",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 991",
      "pageStart": 991,
      "pageEnd": 991,
      "target": "../sections/page-991.md",
      "narrative": null,
      "content": "128\n \nTable A7. Drummond checklist for studies quality: Cost/Cost-effectiveness analyses for Novel skin tests for diagnosing TBI \n \nDrummond Checklist Questions \nKulikov 2009  \nAksenova 2011 \nYagudina 2013 \n1.     Was a well-defined question posed in \nanswerable form? \n  \n  \n  \n1.1. Did the study examine both costs and effects of \nthe service(s) or programme(s)? \nYes \nNo, only presents costs \nYes \n1.2. Did the study involve a comparison of \nalternatives? \nYes \nYes \nYes \n1.3. Was a viewpoint for the analysis stated and was \nthe study placed in any decision-making context? \nYes, children and \nad",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table A7 begins applying the Drummond quality checklist to economic studies of novel skin tests, here for three Russian Diaskintest studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-047",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table A7. Drummond checklist for studies quality: Cost/Cost-effectiveness analyses for Novel skin tests for diagnosing TBI",
      "pageStart": 991,
      "pageEnd": 1001,
      "target": "../tables/table-a7.csv",
      "narrative": "draft",
      "content": "Drummond checklist assessment of the quality of cost and cost-effectiveness analyses of novel skin tests for diagnosing TB infection. Rows are the checklist questions, grouped under ten main criteria; columns give each study's answer and comments, in two blocks of studies.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-992",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 992",
      "pageStart": 992,
      "pageEnd": 992,
      "target": "../sections/page-992.md",
      "narrative": null,
      "content": "129\n3.2. Was effectiveness established through an \noverview of clinical studies? \nYes. \n \nSensitivity and \nspecificity of tests \nvalues from number \nof clinical studies. \nNo, cost analysis, no CE \nYes \n3.3. Were observational data or assumptions used to \nestablish effectiveness? If so, what are the potential \nbiases in results? \nYes. Observational \ndata, no time \nhorizon/discounting \nprovided. \nN/A \nAssumption on the proportion of \npatients with a dubious and \npositive test with 2TE PPD-L and \nDiaskintest drug® to be equal to \ntwo different existing artiles, \nwhich might lead to biased \nconclu",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of the Drummond checklist table for novel skin test studies, covering how effectiveness was established and which costs were identified.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-993",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 993",
      "pageStart": 993,
      "pageEnd": 993,
      "target": "../sections/page-993.md",
      "narrative": null,
      "content": "130\n4.3. Were the capital costs, as well as operating \ncosts, included? \nTest and treatment \ncosts included, the \nrest is unclear \nOnly research costs, sample \nanalyses costs, preventive \ntreatment, consultation and Xrays \nYes, it included treatment, \ndiagnostic, drugs and \nchemoteraphy costs, registering \nand operating costs, among \nothers \n5.    Were costs and consequences measured \naccurately in appropriate physical units? \n  \n  \n  \n5.1. Were any of the identified items omitted from \nmeasurement? If so, does this mean that they \ncarried no weight in the subsequent analysis? \nOutcome costs i",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of the Drummond checklist for novel skin test studies, covering cost components, measurement and valuation.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-994",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 994",
      "pageStart": 994,
      "pageEnd": 994,
      "target": "../sections/page-994.md",
      "narrative": null,
      "content": "131\n6.4. Was the valuation of consequences appropriate \nfor the question posed? \nYes \nN/A \nYes \n7.    Were costs and consequences adjusted for \ndifferential timing? \n  \n  \n  \n7.1. Were costs and consequences that occur in the \nfuture ‘discounted’ to their present values? \nNo \nN/A \nNot stated \n7.2. Was there any justification given for the \ndiscount rate used? \nDiscounting rate No \nstated \nN/A \nN/A \n8.    Was an incremental analysis of costs and \nconsequences of alternatives performed? \n  \n  \n  \n8.1. Were the additional (incremental) costs \ngenerated by one alternative over another \ncompared to",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of the Drummond checklist for novel skin test studies, covering discounting, incremental analysis and sensitivity analysis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-995",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 995",
      "pageStart": 995,
      "pageEnd": 995,
      "target": "../sections/page-995.md",
      "narrative": null,
      "content": "132\nto changes in key \nparameters with the \ncost of a second \nclinical visit being the \nmost influential to the \ncost-effectiveness \nratios \n10.    Did the presentation and discussion of study \nresults include all issues of concern to users? \n \n \n \n10.1. Were the conclusions of the analysis based on \nsome overall index or ratio of costs to consequences \n(e.g. cost-effectiveness ratio)? \nYes (2.28 rubles \ncompared to 3.42) \n  \nYes \n10.2. Were the results compared with those of \nothers who have investigated the same question? If \nso, were allowances made for potential differences \nin study metho",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of the Drummond checklist for novel skin test studies, covering how results and conclusions were presented and discussed.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-996",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 996",
      "pageStart": 996,
      "pageEnd": 996,
      "target": "../sections/page-996.md",
      "narrative": null,
      "content": "133\n10.5. Did the study discuss issues of implementation, \nsuch as the feasibility of adopting the ‘preferred’ \nprogramme given existing financial or other \nconstraints, and whether any freed resources could \nbe redeployed to other worthwhile programmes? \nYes \nYes \nNo",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The final checklist item for the first three novel skin test studies, on whether implementation feasibility was discussed.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-997",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 997",
      "pageStart": 997,
      "pageEnd": 997,
      "target": "../sections/page-997.md",
      "narrative": null,
      "content": "134\n \nContinuation of Table A7 \n \n \nDrummond Checklist Questions \nMoiseeva 2014 \nSolodun 2017 \nSinitsyn 2018 \nChugaev 2020 \n1.     Was a well-defined question posed in \nanswerable form? \n  \n  \n  \n  \n1.1. Did the study examine both costs and \neffects of the service(s) or programme(s)? \nYes \nYes \nYes \nYes \n1.2. Did the study involve a comparison of \nalternatives? \nYes \nYes \nYes \nYes \n1.3. Was a viewpoint for the analysis \nstated and was the study placed in any \ndecision-making context? \nYes \nYes, children and adolescents  \nYes, HIV patients \nYes, children \n2. Was a comprehensive description of",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of Table A7 applying the Drummond checklist to four more Diaskintest studies, covering study questions, alternatives and effectiveness.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-998",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 998",
      "pageStart": 998,
      "pageEnd": 998,
      "target": "../sections/page-998.md",
      "narrative": null,
      "content": "135\n3.2. Was effectiveness established through \nan overview of clinical studies? \nYes \nYes \nYes, CB analysis \nYes \n3.3. Were observational data or \nassumptions used to establish \neffectiveness? If so, what are the potential \nbiases in results? \nNot clear CE \nmethod used \nYes, it uses observational data \nfrom specific settings \nYes, ICER is not computed and no \nother interventions were employed \nrather than do-nothing and \nDiaskintest \n  \n4. Were all the important and relevant \ncosts and consequences for each \nalternative identified? \n  \n  \n  \n  \n4.1. Was the range wide enough for the \nresearch",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of the Drummond checklist for further novel skin test studies, covering effectiveness evidence and cost components.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-999",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 999",
      "pageStart": 999,
      "pageEnd": 999,
      "target": "../sections/page-999.md",
      "narrative": null,
      "content": "136\n5.1. Were any of the identified items \nomitted from measurement? If so, does \nthis mean that they carried no weight in \nthe subsequent analysis? \nNo \nNo \nNo \nNo \n5.2. Were there any special circumstances \n(e.g., joint use of resources) that made \nmeasurement difficult? Were these \ncircumstances handled appropriately? \nNo \nNo \nNo \nNo \n6.    Were the cost and consequences \nvalued credibly? \n  \n  \n  \n  \n6.1. Were the sources of all values clearly \nidentified? \nUnclear \nYes, most costs coming from \nthe State Treasure Healthcare \ninstitution \nYes \n  \n6.2. Were market values employed for \nchange",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of the Drummond checklist for further novel skin test studies, covering measurement, valuation and discounting.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1000",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1000",
      "pageStart": 1000,
      "pageEnd": 1000,
      "target": "../sections/page-1000.md",
      "narrative": null,
      "content": "137\n8.    Was an incremental analysis of costs \nand consequences of alternatives \nperformed? \n  \n  \n  \n  \n8.1. Were the additional (incremental) \ncosts generated by one alternative over \nanother compared to the additional \neffects, benefits, or utilities generated? \nYes \nYes \nNon ICER computed, comparison \nbetween two alternatives \nYes Diaskin over \nTST, costs saved  \n9.    Was allowance made for uncertainty \nin the estimates of costs and \nconsequences? \n \n \n \n  \n9.1. If data on costs and consequences \nwere stochastic (randomly determined \nsequence of observations), were \nappropriate statistic",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of the Drummond checklist for further novel skin test studies, covering incremental and sensitivity analyses and the basis of conclusions.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1001",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1001",
      "pageStart": 1001,
      "pageEnd": 1001,
      "target": "../sections/page-1001.md",
      "narrative": null,
      "content": "138\n10.2. Were the results compared with \nthose of others who have investigated the \nsame question? If so, were allowances \nmade for potential differences in study \nmethodology? \nNo \nNo \nNo \nNo \n10.3. Did the study discuss the \ngeneralisability of the results to other \nsettings and patient/client groups? \nNo \nNo \nNo \nYes \n10.4. Did the study allude to, or take \naccount of, other important factors in the \nchoice or decision under consideration \n(e.g. distribution of costs and \nconsequences, or relevant ethical issues)? \nNo \nNo \nNo \nNo \n10.5. Did the study discuss issues of \nimplementation, such",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The last Drummond checklist items for further novel skin test studies, on comparison with other work, generalisability and implementation.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1002",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1002",
      "pageStart": 1002,
      "pageEnd": 1002,
      "target": "../sections/page-1002.md",
      "narrative": null,
      "content": "139\nTable A8. Drummond Checklist for studies quality: Cost/Cost-effectiveness analyses for TST or IGRA for diagnosing TBI \n \n \nDrummond Checklist \nQuestions \nLinas (2011) \nPareek (2011) \ndel Campo (2012) \nEralp (2012) \nShah (2012) \n1. Was a well-defined question posed in answerable form? \n1.1. Did the study examine \nboth costs and effects of \nthe service(s) or \nprogramme(s)? \nYes \nYes \nYes \nYes \nYes \n1.2. Did the study involve \na comparison of \nalternatives? \nYes \nYes \nYes \nYes \nYes \n1.3. Was a viewpoint for \nthe analysis stated and \nwas the study placed in \nany decision-making \ncontext? \nYes",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table A8 begins applying the Drummond checklist to economic studies of the tuberculin skin test and IGRA, for five studies from 2011–2012.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-048",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table A8. Drummond Checklist for studies quality: Cost/Cost-effectiveness analyses for TST or IGRA for diagnosing TBI",
      "pageStart": 1002,
      "pageEnd": 1043,
      "target": "../tables/table-a8.csv",
      "narrative": "draft",
      "content": "Drummond checklist assessment of the quality of cost and cost-effectiveness analyses of the tuberculin skin test (TST) or interferon-gamma release assays (IGRA) for diagnosing TB infection. Rows are the checklist questions, grouped under ten main criteria; columns give each study's answer and comments, in six successive blocks of studies.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1003",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1003",
      "pageStart": 1003,
      "pageEnd": 1003,
      "target": "../sections/page-1003.md",
      "narrative": null,
      "content": "140\ncontrolled clinical trial? If \nso, did the trial protocol \nreflect what would \nhappen in regular \npractice? \n3.2. Was effectiveness \nestablished through an \noverview of clinical \nstudies? \nYes. \n \nSensitivity and \nspecificity of tests \nvalues from number \nof clinical studies. \nYes. \n \nSensitivity and \nspecificity of tests \nvalues from number \nof clinical studies. \nYes. \n \nSensitivity and \nspecificity of tests \nvalues from number of \nclinical studies. \nYes. \n \nSensitivity and \nspecificity of tests \nvalues from number of \nclinical studies. \nYes. \n \nSensitivity and \nspecificity of tests \nvalu",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of Table A8, covering effectiveness evidence and the range of costs and consequences for the first five tuberculin skin test and IGRA studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1004",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1004",
      "pageStart": 1004,
      "pageEnd": 1004,
      "target": "../sections/page-1004.md",
      "narrative": null,
      "content": "141\n4.3. Were the capital \ncosts, as well as operating \ncosts, included? \nNo. \n \nState only direct \nmedical costs \nincluded.  \nNo. \n \nNo. \nYes, stated overheads \nincluded but no \ndescription provided. \nYes, operating costs \nsuch as quality \nassurance, specimen \ntransport, supply \ndelivery, and \nestimates for rent and \nutilities included. \n5.    Were costs and consequences measured accurately in appropriate physical units? \n5.1. Were any of the \nidentified items omitted \nfrom measurement? If so, \ndoes this mean that they \ncarried no weight in the \nsubsequent analysis? \nNo \nYes. \n \nExcluded drug",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of Table A8, covering which cost components were included and how costs were valued.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1005",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1005",
      "pageStart": 1005,
      "pageEnd": 1005,
      "target": "../sections/page-1005.md",
      "narrative": null,
      "content": "142\n6.3. Where market values \nwere absent (e.g. \nvolunteer labour), or \nmarket values did not \nreflect actual values (such \nas clinic space donated at \na reduced rate), were \nadjustments made to \napproximate market \nvalues? \nN/A \nN/A \nN/A \nN/A \nN/A \n6.4. Was the valuation of \nconsequences appropriate \nfor the question posed? \nYes \nYes \nYes \nYes \nYes \n7.    Were costs and consequences adjusted for differential timing? \n7.1. Were costs and \nconsequences that occur \nin the future ‘discounted’ \nto their present values? \nYes, 3%. \nYes, 3.5%. \nNo. \nYes, 5%. \nYes, 3%. \n7.2. Was there any \njustificati",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of Table A8, covering valuation of consequences, discounting and incremental analysis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1006",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1006",
      "pageStart": 1006,
      "pageEnd": 1006,
      "target": "../sections/page-1006.md",
      "narrative": null,
      "content": "143\n \n \n \n \n9.    Was allowance made for uncertainty in the estimates of costs and consequences? \n9.1. If data on costs and \nconsequences were \nstochastic (randomly \ndetermined sequence of \nobservations), were \nappropriate statistical \nanalyses performed? \nYes \nYes \nYes \nYes \nYes \n9.2. If sensitivity analysis \nwas employed, was \njustification provided for \nthe range of values (or for \nkey study parameters)? \nYes \nYes \nYes \nYes \nYes \n9.3. Were the study \nresults sensitive to \nchanges in the values? \nYes – many \nconclusions for \ndifferent populations \nsensitive to key \nparameters. \nNo \nNo \nNo \nN",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of Table A8, covering allowance for uncertainty and the basis of study conclusions.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1007",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1007",
      "pageStart": 1007,
      "pageEnd": 1007,
      "target": "../sections/page-1007.md",
      "narrative": null,
      "content": "144\n10.2. Were the results \ncompared with those of \nothers who have \ninvestigated the same \nquestion? If so, were \nallowances made for \npotential differences in \nstudy methodology? \nNo \nYes \nYes \nNo \nNo \n10.3. Did the study discuss \nthe generalisability of the \nresults to other settings \nand patient/client groups? \nNo \nYes \nNo \nNo \nYes \n10.4. Did the study allude \nto, or take account of, \nother important factors in \nthe choice or decision \nunder consideration (e.g. \ndistribution of costs and \nconsequences, or relevant \nethical issues)? \nNo \nYes \nNo \nYes \nNo \n10.5. Did the study discuss \nissues",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of Table A8, covering comparison with other studies, generalisability and implementation issues.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1008",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1008",
      "pageStart": 1008,
      "pageEnd": 1008,
      "target": "../sections/page-1008.md",
      "narrative": null,
      "content": "145\nContinuation of Table A8 \n \nDrummond Checklist \nQuestions \nMandalakas (2013) \nPareek (2013) \nSteffen (2013) \nSwaminath (2013) \nVerma (2013) \n1. Was a well-defined question posed in answerable form? \n1.1. Did the study examine \nboth costs and effects of \nthe service(s) or \nprogramme(s)? \nYes \nYes \nYes \nYes \nYes \n1.2. Did the study involve \na comparison of \nalternatives? \nYes \nYes \nYes \nYes \nYes \n1.3. Was a viewpoint for \nthe analysis stated and \nwas the study placed in \nany decision-making \ncontext? \nYes \n \nChildren (0-2 and 3-\n5y.o.) contacts in \nSouth Africa. \nYes \n \nRecently arrived \nimm",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of Table A8 for five studies from 2013, covering study questions, alternatives and effectiveness.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1009",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1009",
      "pageStart": 1009,
      "pageEnd": 1009,
      "target": "../sections/page-1009.md",
      "narrative": null,
      "content": "146\nhappen in regular \npractice? \n3.2. Was effectiveness \nestablished through an \noverview of clinical \nstudies? \nYes. \n \nSensitivity and \nspecificity of tests \nvalues from number \nof clinical studies. \nYes. \n \nSensitivity and \nspecificity of tests \nvalues from number \nof clinical studies. \nYes. \nSensitivity and specificity \nof tests values from \nnumber of clinical \nstudies. \nYes. \n \nSensitivity and \nspecificity of tests \nvalues from number of \nclinical studies. \nYes \n \nSensitivity and \nspecificity of tests \nvalues from number \nof clinical studies. \n3.3. Were observational \ndata or assumptions",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of the Drummond checklist for the 2013 studies, covering effectiveness evidence and the range of costs and consequences.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1010",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1010",
      "pageStart": 1010,
      "pageEnd": 1010,
      "target": "../sections/page-1010.md",
      "narrative": null,
      "content": "147\n4.3. Were the capital \ncosts, as well as operating \ncosts, included? \nUnclear. \nOutpatient \nhospitalisation costs \nincluded but no \ndescription provided.  \nNo, state only direct \nmedical costs \nincluded and clearly \nstate breakdown of \nhospitalisation costs. \nNo. \n \n \n \n \n \nNo, authors use Linas \n2011 hospitalisation \ncosts. \nNo. \nHospitalisation costs \nbreakdown do not \nstate any \noverheads/operating \ncosts. \n5.    Were costs and consequences measured accurately in appropriate physical units? \n5.1. Were any of the \nidentified items omitted \nfrom measurement? If so, \ndoes this mean that th",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of the Drummond checklist for the 2013 studies, covering cost components and the sources of values.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1011",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1011",
      "pageStart": 1011,
      "pageEnd": 1011,
      "target": "../sections/page-1011.md",
      "narrative": null,
      "content": "148\n6.2. Were market values \nemployed for changes \ninvolving resources gained \nor depleted? \nYes \nYes \nYes \nYes \nYes \n6.3. Where market values \nwere absent (e.g. \nvolunteer labour), or \nmarket values did not \nreflect actual values (such \nas clinic space donated at \na reduced rate), were \nadjustments made to \napproximate market \nvalues? \nN/A \nN/A \nN/A \nN/A \nN/A \n6.4. Was the valuation of \nconsequences appropriate \nfor the question posed? \n \n \nYes \nYes \nYes \nYes \nYes \n7.    Were costs and consequences adjusted for differential timing? \n7.1. Were costs and \nconsequences that occur \nin the future",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of the Drummond checklist for the 2013 studies, covering market values, discounting and the discount rate used.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1012",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1012",
      "pageStart": 1012,
      "pageEnd": 1012,
      "target": "../sections/page-1012.md",
      "narrative": null,
      "content": "149\n8.1. Were the additional \n(incremental) costs \ngenerated by one \nalternative over another \ncompared to the \nadditional effects, \nbenefits, or utilities \ngenerated? \n \n \n \nYes \nYes \nYes \nYes \nYes \n9.    Was allowance made for uncertainty in the estimates of costs and consequences? \n9.1. If data on costs and \nconsequences were \nstochastic (randomly \ndetermined sequence of \nobservations), were \nappropriate statistical \nanalyses performed? \nYes \nYes \nYes \nYes \nYes \n9.2. If sensitivity analysis \nwas employed, was \njustification provided for \nthe range of values (or for \nkey study parameters)?",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of the Drummond checklist for the 2013 studies, covering incremental analysis and sensitivity of results to key parameters.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1013",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1013",
      "pageStart": 1013,
      "pageEnd": 1013,
      "target": "../sections/page-1013.md",
      "narrative": null,
      "content": "150\n10.1. Were the \nconclusions of the analysis \nbased on some overall \nindex or ratio of costs to \nconsequences (e.g. cost-\neffectiveness ratio)? \nYes \nYes \nYes \nNo \n \nNo cost-effectiveness \nratio was used. \nConclusions were \nbased on “highest \nbenefits” and “lowest \ncost” strategy.  \nYes \n10.2. Were the results \ncompared with those of \nothers who have \ninvestigated the same \nquestion? If so, were \nallowances made for \npotential differences in \nstudy methodology? \nNo \nYes \nYes \nNo \nNo \n10.3. Did the study discuss \nthe generalisability of the \nresults to other settings \nand patient/client grou",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A continuation of the Drummond checklist for the 2013 studies, covering how conclusions were reached and other factors considered.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1014",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1014",
      "pageStart": 1014,
      "pageEnd": 1014,
      "target": "../sections/page-1014.md",
      "narrative": null,
      "content": "151\n10.5. Did the study discuss \nissues of implementation, \nsuch as the feasibility of \nadopting the ‘preferred’ \nprogramme given existing \nfinancial or other \nconstraints, and whether \nany freed resources could \nbe redeployed to other \nworthwhile programmes? \nYes \nYes \nYes \nNo \nNo",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The final Drummond checklist item for the 2013 studies, on discussion of implementation feasibility.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1015",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1015",
      "pageStart": 1015,
      "pageEnd": 1015,
      "target": "../sections/page-1015.md",
      "narrative": null,
      "content": "152\nContinuation of Table A8",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A near-blank page carrying only a page number and the heading for a continuation of Table A8.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1016",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1016",
      "pageStart": 1016,
      "pageEnd": 1016,
      "target": "../sections/page-1016.md",
      "narrative": null,
      "content": "153\nDrummond Checklist \nQuestions \nCapocci, (2015) \nWingate (2015) \nAuguste. (2016) \nNijhawan (2016) \nCampbell (2017) \n2. Was a well-defined question posed in answerable form? \n1.1. Did the study examine \nboth costs and effects of \nthe service(s) or \nprogramme(s)? \nYes \nYes \nYes \nYes \nYes \n1.2. Did the study involve \na comparison of \nalternatives? \nYes \nYes \nYes \nYes \nYes \n1.3. Was a viewpoint for \nthe analysis stated and \nwas the study placed in \nany decision-making \ncontext? \nYes \n \nAdults living with HIV \nin the UK. \nYes \n \nPre arrival refugees to \nUSA.  \nYes \nChildren, \nimmunocompromised",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Begins a table applying the Drummond checklist for economic evaluations to five cost-effectiveness studies of tuberculosis infection testing published in 2015–2017. All five compared costs and effects of alternatives for groups such as people with HIV, refugees, immigrants and people entering jail, and most included a do-nothing option.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1017",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1017",
      "pageStart": 1017,
      "pageEnd": 1017,
      "target": "../sections/page-1017.md",
      "narrative": null,
      "content": "154\n3.1. Was this done \nthrough a randomised, \ncontrolled clinical trial? If \nso, did the trial protocol \nreflect what would \nhappen in regular \npractice? \nNo \nNo \nNo \nNo \nNo \n3.2. Was effectiveness \nestablished through an \noverview of clinical \nstudies? \nSomewhat. \n \nSensitivity of IGRA \nused from one clinical \nstudy.  \n \nYes \n \nSensitivity and \nspecificity of TST \nvalues from number \nof clinical studies \nYes \n \nSystematic review \ncarried out to establish \nsensitivity and specificity \nof tests.  \nYes \n \nSensitivity and \nspecificity of test values \nfrom clinical studies \nand systematic review.",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the Drummond checklist table for the same five studies, on how effectiveness was established. None used randomized trials; most drew test sensitivity and specificity from clinical studies or reviews, and the table notes how well their assumptions were referenced.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1018",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1018",
      "pageStart": 1018,
      "pageEnd": 1018,
      "target": "../sections/page-1018.md",
      "narrative": null,
      "content": "155\n4.1. Was the range wide \nenough for the research \nquestion at hand? \nYes \nYes \nYes \nYes \nYes \n4.2. Did it cover all \nrelevant viewpoints? \nYes \nYes \nYes \nYes \nYes \n4.3. Were the capital \ncosts, as well as operating \ncosts, included? \nUnclear breakdown of \ncosts provided. \nNo.  \n \nIncludes labour but no \nindication of other \noperating/capital \ncosts.  \nNo.  \n \nClear breakdown \nprovided and \ncapital/overheads not \nincluded. \nYes. \n \nFor QFT-GIT laboratory \noperations included. \nNo. \n \n5.    Were costs and consequences measured accurately in appropriate physical units? \n5.1. Were any of the",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the checklist table on whether the studies identified and measured all relevant costs and consequences. It records gaps such as unclear cost breakdowns, missing capital costs and one study omitting treatment and outcome costs.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1019",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1019",
      "pageStart": 1019,
      "pageEnd": 1019,
      "target": "../sections/page-1019.md",
      "narrative": null,
      "content": "156\n6.2. Were market values \nemployed for changes \ninvolving resources gained \nor depleted? \nYes \nYes \nYes \nYes \nYes \n6.3. Where market values \nwere absent (e.g. \nvolunteer labour), or \nmarket values did not \nreflect actual values (such \nas clinic space donated at \na reduced rate), were \nadjustments made to \napproximate market \nvalues? \nN/A \nN/A \nN/A \nN/A \nN/A \n6.4. Was the valuation of \nconsequences appropriate \nfor the question posed? \nYes \nYes \nYes \nYes \nYes \n7.    Were costs and consequences adjusted for differential timing? \n7.1. Were costs and \nconsequences that occur \nin the future ‘dis",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the checklist on valuation and discounting, noting that future costs were mostly discounted at 1.5–3.5% with rates justified by national guidance, while one study did not discount.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1020",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1020",
      "pageStart": 1020,
      "pageEnd": 1020,
      "target": "../sections/page-1020.md",
      "narrative": null,
      "content": "157\n8.1. Were the additional \n(incremental) costs \ngenerated by one \nalternative over another \ncompared to the \nadditional effects, \nbenefits, or utilities \ngenerated? \nYes \nYes \nYes \nYes \nYes \n9.    Was allowance made for uncertainty in the estimates of costs and consequences? \n9.1. If data on costs and \nconsequences were \nstochastic (randomly \ndetermined sequence of \nobservations), were \nappropriate statistical \nanalyses performed? \nYes \nYes  \nYes \nYes \nYes  \n9.2. If sensitivity analysis \nwas employed, was \njustification provided for \nthe range of values (or for \nkey study parameters)? \nYes",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Records that all five studies did incremental analyses and allowed for uncertainty, with only one finding its results sensitive to the treatments received.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1021",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1021",
      "pageStart": 1021,
      "pageEnd": 1021,
      "target": "../sections/page-1021.md",
      "narrative": null,
      "content": "158\n10.1. Were the \nconclusions of the analysis \nbased on some overall \nindex or ratio of costs to \nconsequences (e.g. cost-\neffectiveness ratio)? \nYes \nYes  \nYes \nNo \n \nYes \n10.2. Were the results \ncompared with those of \nothers who have \ninvestigated the same \nquestion? If so, were \nallowances made for \npotential differences in \nstudy methodology? \nNo \nYes \nYes  \nYes \nYes \n10.3. Did the study discuss \nthe generalisability of the \nresults to other settings \nand patient/client groups? \nYes \nNo \nYes \nYes  \nYes \n10.4. Did the study allude \nto, or take account of, \nother important factors in \nthe",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Records whether each study reported cost-effectiveness ratios, compared results with other work, and discussed generalizability and other decision factors.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1022",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1022",
      "pageStart": 1022,
      "pageEnd": 1022,
      "target": "../sections/page-1022.md",
      "narrative": null,
      "content": "159\n10.5. Did the study discuss \nissues of implementation, \nsuch as the feasibility of \nadopting the ‘preferred’ \nprogramme given existing \nfinancial or other \nconstraints, and whether \nany freed resources could \nbe redeployed to other \nworthwhile programmes? \nYes \nNo \nYes \nYes  \nYes",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Records whether each study discussed implementation issues such as feasibility and redeployment of freed resources.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1023",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1023",
      "pageStart": 1023,
      "pageEnd": 1023,
      "target": "../sections/page-1023.md",
      "narrative": null,
      "content": "160\n \nContinuation of Table A8",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A page marking the continuation of Table A8.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1024",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1024",
      "pageStart": 1024,
      "pageEnd": 1024,
      "target": "../sections/page-1024.md",
      "narrative": null,
      "content": "161\nDrummond Checklist \nQuestions \nHaukaas (2017) \nMullie (2017) \nTasillo, (2017) \nAbubakar (2018) \nLi (2018) \n1. Was a well-defined question posed in answerable form? \n1.1. Did the study examine \nboth costs and effects of \nthe service(s) or \nprogramme(s)? \nYes \nYes \nYes \nYes \nYes \n1.2. Did the study involve a \ncomparison of alternatives? \nYes \nYes \nYes \nYes \nYes \n1.3. Was a viewpoint for the \nanalysis stated and was the \nstudy placed in any \ndecision-making context? \nYes \n \nRecently arrived \nimmigrants <35y.o to \nNorway \nYes \n \nHCW with negative TST \nat time of employment.  \nYes \n \nUS-born or",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Applies the Drummond checklist to a second set of five cost-effectiveness studies of tuberculosis infection testing from 2017–2018. Their target groups include young immigrants to Norway, health workers, US-born people and migrants, recent UK immigrants and elderly care home residents in Hong Kong.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1025",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1025",
      "pageStart": 1025,
      "pageEnd": 1025,
      "target": "../sections/page-1025.md",
      "narrative": null,
      "content": "162\n3.2. Was effectiveness \nestablished through an \noverview of clinical studies? \nSomewhat. \n \nSensitivity and specificity \nof TST values from only \none study. \nYes \n \nSensitivity and specificity \nof tests values from \nnumber of clinical \nstudies. \nYes \n \nSensitivity and specificity \nof tests values from \nnumber of clinical \nstudies. \nNo \nYes \n \nSensitivity and \nspecificity of \ntests values \nfrom number of \nclinical studies. \n3.3. Were observational \ndata or assumptions used \nto establish effectiveness? \nIf so, what are the potential \nbiases in results? \nYes \n \nExpert opinion and \nestimation",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Shows how the second set of studies established effectiveness, mainly from clinical studies of test accuracy, and notes key assumptions such as no quality-of-life loss from uncomplicated treatment. It also records that all covered a wide enough range of costs and viewpoints.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1026",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1026",
      "pageStart": 1026,
      "pageEnd": 1026,
      "target": "../sections/page-1026.md",
      "narrative": null,
      "content": "163\n4.3. Were the capital costs, \nas well as operating costs, \nincluded? \nUnclear. \n \nHospitalisation costs \nincluded but no \nbreakdown provided. \nUnclear. \n \nHospitalisation costs \nincluded but no \nbreakdown provided. \nUnclear. \n \nTreatment costs \nincluded but no \nbreakdown provided. \nNo. \n \nClear breakdown \nprovided and \ncapital/overheads \nnot included \nNo. \n \nClear \nbreakdown \nprovided and \ncapital/overhea\nds not included \n5.    Were costs and consequences measured accurately in appropriate physical units? \n5.1. Were any of the \nidentified items omitted \nfrom measurement? If so, \ndoes this",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Records for the second set of studies whether capital and operating costs were included and whether any items were left out of measurement. Notes include unclear hospital cost breakdowns and omitted costs that may underestimate totals.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1027",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1027",
      "pageStart": 1027,
      "pageEnd": 1027,
      "target": "../sections/page-1027.md",
      "narrative": null,
      "content": "164\n6.3. Where market values \nwere absent (e.g. volunteer \nlabour), or market values \ndid not reflect actual values \n(such as clinic space \ndonated at a reduced rate), \nwere adjustments made to \napproximate market \nvalues? \nN/A \nN/A \nN/A \nN/A \nN/A \n6.4. Was the valuation of \nconsequences appropriate \nfor the question posed? \nYes \nYes \nYes \nYes \nYes \n7.    Were costs and consequences adjusted for differential timing? \n7.1. Were costs and \nconsequences that occur in \nthe future ‘discounted’ to \ntheir present values? \nYes, 4% \n \nOnly costs discounted in \nbase case analysis. \nYes, 3% \nYes, 3% \nYes",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Records discount rates of 3–5% and their justification for the second set of studies, all of which performed incremental analyses.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1028",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1028",
      "pageStart": 1028,
      "pageEnd": 1028,
      "target": "../sections/page-1028.md",
      "narrative": null,
      "content": "165\n9.1. If data on costs and \nconsequences were \nstochastic (randomly \ndetermined sequence of \nobservations), were \nappropriate statistical \nanalyses performed? \nYes \nYes \nYes  \nYes \nYes \n9.2. If sensitivity analysis \nwas employed, was \njustification provided for \nthe range of values (or for \nkey study parameters)? \nNo \nYes \nYes  \nYes \nYes \n9.3. Were the study results \nsensitive to changes in the \nvalues? \nYes. \nSensitive to IGRA cost.  \nNo \nYes \n \nSensitive to LTBI rate \nand TST sensitivity. \nNo \nNo. \n10.    Did the presentation and discussion of study results include all issues of concern t",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Records statistical and sensitivity analyses for the second set of studies, noting results sensitive to IGRA cost, infection rate and TST sensitivity, and whether results were compared with other studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1029",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1029",
      "pageStart": 1029,
      "pageEnd": 1029,
      "target": "../sections/page-1029.md",
      "narrative": null,
      "content": "166\n10.3. Did the study discuss \nthe generalisability of the \nresults to other settings and \npatient/client groups? \nNo \nYes \nNo \nYes \nYes \n10.4. Did the study allude \nto, or take account of, other \nimportant factors in the \nchoice or decision under \nconsideration (e.g. \ndistribution of costs and \nconsequences, or relevant \nethical issues)? \nYes \nYes \nYes  \nYes \nYes \n10.5. Did the study discuss \nissues of implementation, \nsuch as the feasibility of \nadopting the ‘preferred’ \nprogramme given existing \nfinancial or other \nconstraints, and whether \nany freed resources could \nbe redeployed to othe",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Records whether the second set of studies discussed generalizability, other decision factors and implementation.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1030",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1030",
      "pageStart": 1030,
      "pageEnd": 1030,
      "target": "../sections/page-1030.md",
      "narrative": null,
      "content": "167\n \nContinuation of Table A8",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A page marking the continuation of Table A8.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1031",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1031",
      "pageStart": 1031,
      "pageEnd": 1031,
      "target": "../sections/page-1031.md",
      "narrative": null,
      "content": "168\nDrummond Checklist \nQuestions \nSohn (2018). \nCampbell (2019a). \nCampbell (2019b). \nLoureiro (2019). \nPng (2019). \n1. Was a well-defined question posed in answerable form? \n1.1. Did the study examine \nboth costs and effects of \nthe service(s) or \nprogramme(s)? \nYes \nYes \nYes \nYes \nYes \n1.2. Did the study involve a \ncomparison of alternatives? \nYes \nYes \nYes \nYes \nYes \n1.3. Was a viewpoint for the \nanalysis stated and was the \nstudy placed in any \ndecision-making context? \nYes \n \nAdolescents (13-18y.o.) \ncontacts in Japan. \nYes \n \nPre-arrival immigrants to \nCanada  \nYes \n \nMigrants with eith",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Applies the Drummond checklist to a third set of five studies from 2018–2019, covering adolescent contacts in Japan, immigrants to Canada, migrants with kidney disease, and health workers in Brazil and Singapore. All compared costs and effects of alternatives, and none used randomized trials.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1032",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1032",
      "pageStart": 1032,
      "pageEnd": 1032,
      "target": "../sections/page-1032.md",
      "narrative": null,
      "content": "169\n3.2. Was effectiveness \nestablished through an \noverview of clinical studies? \nYes \n \nSensitivity and specificity \nof tests values from \nnumber of clinical \nstudies and systematic \nreviews. \nYes \n \nSensitivity and specificity \nof tests values from \nnumber of clinical \nstudies. \nYes \n \nSensitivity and specificity \nof tests values from \nnumber of clinical \nstudies. \nYes \n \nSensitivity and \nspecificity of tests \nvalues from number \nof clinical studies. \nYes \n3.3. Were observational \ndata or assumptions used \nto establish effectiveness? \nIf so, what are the potential \nbiases in results? \nYes",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Shows how the third set of studies established effectiveness from clinical studies and reviews and comments on their key assumptions. One study assumed perfect test accuracy, while another gave no references for its assumptions.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1033",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1033",
      "pageStart": 1033,
      "pageEnd": 1033,
      "target": "../sections/page-1033.md",
      "narrative": null,
      "content": "170\n4.3. Were the capital costs, \nas well as operating costs, \nincluded? \nYes. \n \nIncludes overheads and \nvehicle operations. \nUnclear. \n \nHospitalisation costs \nincluded but no \nbreakdown provided. \nUnclear. \n \nHospitalisation costs \nincluded but no \nbreakdown provided. \nYes. \n \nNo capital costs \nstated but \nequipment use costs \nclearly included. \nYes. \n \nState that \nhospitalisation \nand testing costs \ninclude labour \nand overhead \ncosts. \n5.    Were costs and consequences measured accurately in appropriate physical units? \n5.1. Were any of the \nidentified items omitted \nfrom measurement?",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Records cost coverage for the third set of studies, such as whether overheads and equipment were included. It also notes omitted items judged to have minimal impact and whether value sources were stated.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1034",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1034",
      "pageStart": 1034,
      "pageEnd": 1034,
      "target": "../sections/page-1034.md",
      "narrative": null,
      "content": "171\n6.3. Where market values \nwere absent (e.g. volunteer \nlabour), or market values \ndid not reflect actual values \n(such as clinic space \ndonated at a reduced rate), \nwere adjustments made to \napproximate market \nvalues? \nN/A \nN/A \nN/A \nN/A \nN/A \n6.4. Was the valuation of \nconsequences appropriate \nfor the question posed? \nYes \nYes \nYes \nYes \nYes \n7.    Were costs and consequences adjusted for differential timing? \n7.1. Were costs and \nconsequences that occur in \nthe future ‘discounted’ to \ntheir present values? \nNo \nYes, 3% \nYes, 1.5% \nNo \nYes, 3% \n7.2. Was there any \njustification given fo",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Records discounting choices for the third set of studies, with some not discounting because of short time horizons, and confirms that all did incremental analyses.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1035",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1035",
      "pageStart": 1035,
      "pageEnd": 1035,
      "target": "../sections/page-1035.md",
      "narrative": null,
      "content": "172\n9.    Was allowance made for uncertainty in the estimates of costs and consequences? \n9.1. If data on costs and \nconsequences were \nstochastic (randomly \ndetermined sequence of \nobservations), were \nappropriate statistical \nanalyses performed? \nYes \nYes \nYes \nYes \nYes \n9.2. If sensitivity analysis \nwas employed, was \njustification provided for \nthe range of values (or for \nkey study parameters)? \nYes \nYes \nYes \nYes \nYes \n9.3. Were the study results \nsensitive to changes in the \nvalues? \nYes \n \nSensitive to QFT-GIT \ncost. \n \nYes  \n \nFor migrants from low \nincidence conclusion \nonly cost-eff",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Records sensitivity analyses for the third set of studies, including sensitivity to QFT-GIT cost and to migrants' origin, and whether conclusions used cost-effectiveness ratios.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1036",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1036",
      "pageStart": 1036,
      "pageEnd": 1036,
      "target": "../sections/page-1036.md",
      "narrative": null,
      "content": "173\n10.2. Were the results \ncompared with those of \nothers who have \ninvestigated the same \nquestion? If so, were \nallowances made for \npotential differences in \nstudy methodology? \nYes \nYes \nYes \nYes \nYes \n10.3. Did the study discuss \nthe generalisability of the \nresults to other settings and \npatient/client groups? \nYes \nYes \nYes \nNo \nYes \n10.4. Did the study allude \nto, or take account of, other \nimportant factors in the \nchoice or decision under \nconsideration (e.g. \ndistribution of costs and \nconsequences, or relevant \nethical issues)? \nYes \nYes \nYes \nYes \nNo \n10.5. Did the study discuss",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Records whether the third set of studies compared results with others and discussed generalizability, decision factors and implementation.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1037",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1037",
      "pageStart": 1037,
      "pageEnd": 1037,
      "target": "../sections/page-1037.md",
      "narrative": null,
      "content": "174",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A page containing only a page number.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1038",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1038",
      "pageStart": 1038,
      "pageEnd": 1038,
      "target": "../sections/page-1038.md",
      "narrative": null,
      "content": "175\nContinuation of Table A8",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A page marking the continuation of Table A8.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1039",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1039",
      "pageStart": 1039,
      "pageEnd": 1039,
      "target": "../sections/page-1039.md",
      "narrative": null,
      "content": "176\nDrummond Checklist Questions \nAl Abri (2020). \nJo (2020). \nSteffen (2020) \nKim (2018) \n1.Was a well-defined question posed in answerable form? \n \n1.1. Did the study examine both costs \nand effects of the service(s) or \nprogramme(s)? \nYes \nYes \nYes \nYes \n1.2. Did the study involve a comparison \nof alternatives? \nYes \nYes \nYes \nYes \n1.3. Was a viewpoint for the analysis \nstated and was the study placed in any \ndecision-making context? \nYes \n \n20-year-old recent \nimmigrants in Oman.  \nYes \nGeneral population \nwith vulnerable sub-\npopulations.   \nYes \n \nAdults living with HIV in \nBrazil. \nYes",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Applies the Drummond checklist to a final four studies from 2018–2020, covering recent immigrants in Oman, a general population, adults with HIV in Brazil and pregnant women with HIV. None included a do-nothing option or used trials, and all drew test accuracy from reviews or clinical studies.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1040",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1040",
      "pageStart": 1040,
      "pageEnd": 1040,
      "target": "../sections/page-1040.md",
      "narrative": null,
      "content": "177\n3.3. Were observational data or \nassumptions used to establish \neffectiveness? \nIf so, what are the potential biases in \nresults? \nYes \n \nNo clear assumptions \nstated. \n \nTB incidence and the \nLTBI rate, BCG \nvaccination rate, (MDR-\nTB) rate, TB mortality, \nand rate of adherence \nto treatment were \nimputed from Omani \ndata. \nYes  \n \nMost assumptions for \neffectiveness all have \nadequate \nreferences/reasoning. \n \nYes \n \nAll assumptions for \neffectiveness all have \nadequate \nreferences/reasoning. \n \n \nYes \n \nAll assumptions for \neffectiveness all have \nadequate \nreferences/reasoning. \n \n4. W",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Records the effectiveness assumptions of the final four studies, noting one that imputed rates from Omani data without clear assumptions. It also records how fully each covered costs, including overheads and equipment.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1041",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1041",
      "pageStart": 1041,
      "pageEnd": 1041,
      "target": "../sections/page-1041.md",
      "narrative": null,
      "content": "178\n5.2. Were there any special circumstances \n(e.g., joint use of resources) that made \nmeasurement difficult? Were these \ncircumstances handled appropriately? \nNo \nNo \nNo \nNo \n6.    Were the cost and consequences valued credibly? \n \n6.1. Were the sources of all values clearly \nidentified? \nNo \nYes \nYes \nYes \n6.2. Were market values employed for \nchanges involving resources gained or \ndepleted? \nNo \nYes \nYes \nYes \n6.3. Where market values were absent (e.g. \nvolunteer labour), or market values did not \nreflect actual values (such as clinic space \ndonated at a reduced rate), were \nadjustments m",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Records how credibly the final four studies valued costs and how they discounted, with one study lacking clear value sources, market values and discounting. All performed incremental analyses.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1042",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1042",
      "pageStart": 1042,
      "pageEnd": 1042,
      "target": "../sections/page-1042.md",
      "narrative": null,
      "content": "179\n9.1. If data on costs and consequences were \nstochastic (randomly determined sequence \nof observations), were appropriate statistical \nanalyses performed? \nYes \nYes \nYes \nYes \n9.2. If sensitivity analysis was employed, was \njustification provided for the range of values \n(or for key study parameters)? \nNo \nYes \nYes \nYes \n9.3. Were the study results sensitive to \nchanges in the values? \nYes \n \nSensitive to \ntreatments \nreceived. \nYes \nSpecificity of QFT-GIT. \nYes \nCost of DiaskinTest. \nYes \nSensitive to the \nprobability of \ndeveloping TB and LTBI \nprevalence \n10.    Did the presentation and",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Records sensitivity results for the final four studies, such as sensitivity to QFT-GIT specificity, DiaskinTest cost and TB risk, and whether they discussed generalizability and other factors.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1043",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1043",
      "pageStart": 1043,
      "pageEnd": 1043,
      "target": "../sections/page-1043.md",
      "narrative": null,
      "content": "180\n10.5. Did the study discuss issues of \nimplementation, such as the feasibility of \nadopting the ‘preferred’ programme given \nexisting financial or other constraints, and \nwhether any freed resources could be \nredeployed to other worthwhile \nprogrammes? \nNo \nYes \nYes \nYes",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Records whether the final four studies discussed implementation issues.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1044",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1044",
      "pageStart": 1044,
      "pageEnd": 1044,
      "target": "../sections/page-1044.md",
      "narrative": null,
      "content": "181\nTable A9. Summary of the proportion of articles accomplishing each of the Drummond’s criteria",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title of Table A9, summarizing the proportion of articles meeting each Drummond criterion.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-049",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table A9. Summary of the proportion of articles accomplishing each of the Drummond’s criteria",
      "pageStart": 1044,
      "pageEnd": 1046,
      "target": "../tables/table-a9.csv",
      "narrative": "draft",
      "content": "Percentage of the reviewed economic evaluations that met each Drummond checklist criterion, for all articles, for DiaskinTest-related articles and for TST or IGRA articles. Rows are the checklist questions, grouped under ten main criteria.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1045",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1045",
      "pageStart": 1045,
      "pageEnd": 1045,
      "target": "../sections/page-1045.md",
      "narrative": null,
      "content": "182\nDrummond Checklist Questions \n% Of papers accomplishing each \ncriterion \nAll articles \nDiaskinTest-\nrelated \nTST or \nIGRA \n1.     Was a well-defined question posed in answerable form? \n1.1. Did the study examine both costs and effects of the service(s) or programme(s)? \n97.1% \n87.5% \n100.0% \n1.2. Did the study involve a comparison of alternatives? \n100.0% \n100.0% \n100.0% \n1.3. Was a viewpoint for the analysis stated and was the study placed in any decision-making context? \n100.0% \n100.0% \n100.0% \n2. Was a comprehensive description of the competing alternatives given? \n2.1. Were there any i",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table A9 gives the share of all reviewed economic studies, DiaskinTest studies and TST or IGRA studies meeting each Drummond criterion. Almost all compared alternatives and drew effectiveness from clinical studies, none used randomized trials, and few included capital costs, while 60% discounted future costs.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1046",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1046",
      "pageStart": 1046,
      "pageEnd": 1046,
      "target": "../sections/page-1046.md",
      "narrative": null,
      "content": "183\n7.2. Was there any justification given for the discount rate used? \n60.0% \n12.5% \n75.0% \n8.    Was an incremental analysis of costs and consequences of alternatives performed? \n8.1. Were the additional (incremental) costs generated by one alternative over another compared to the additional effects, benefits, or \nutilities generated? \n94.3% \n75.0% \n100.0% \n9.    Was allowance made for uncertainty in the estimates of costs and consequences? \n9.1. If data on costs and consequences were stochastic (randomly determined sequence of observations), were appropriate statistical \nanalyses performed?",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues Table A9 with the share of reviewed studies meeting the remaining Drummond criteria, from discount rate justification to discussion of implementation. DiaskinTest-related studies met many criteria less often than TST or IGRA studies, for example in sensitivity analysis and comparison with other work.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1047",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1047",
      "pageStart": 1047,
      "pageEnd": 1047,
      "target": "../sections/page-1047.md",
      "narrative": null,
      "content": "_(no text on this page)_",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A blank page with no extracted text.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1048",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1048",
      "pageStart": 1048,
      "pageEnd": 1048,
      "target": "../sections/page-1048.md",
      "narrative": null,
      "content": "Web Annex B.23. Modelling for economic evidence for the use of \nMycobacterium tuberculosis antigen-based skin tests \n \nReview team: Lara Goscé1, Finna Parkinson1, Kasim Allel,1,2 Elena Surkova3, Irina Kontsevaya4, Ting Ting Wang1, \nVictoria Liu2, Yohhei Hamada1, Peter J White5, Molebogeng X Rangaka1  \n \n1Institute for Global Health, University College London, UK \n2London School of Hygiene and Tropical Medicine, UK \n3Royal Brompton Hospital. Part of Guy's and St Thomas' NHS Foundation Trust London, UK \n4Research Center Borstel, Germany \n5Faculty of Medicine, School of Public Health, Imperial Co",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title page of Web Annex B.23, a modelling study of the economic evidence for tuberculosis antigen-based skin tests, naming the review team and stating it was prepared for a February 2022 WHO guideline meeting.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1049",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1049",
      "pageStart": 1049,
      "pageEnd": 1049,
      "target": "../sections/page-1049.md",
      "narrative": null,
      "content": "Abbreviations  \n \nBCG: Bacillus Calmette-Guérin \nCBA: Cost-benefit analysis \nCUA: Cost-utility analysis  \nCEA: Cost-effective analysis \nDOT: Directly observed therapy \nHICs: High-income countries \nHIV: Human immunodeficiency virus \nICER: incremental cost-effectiveness ratios \nIGRA: interferon-gamma release assays \nLMICs: Low- and middle-income countries \nPPD: purified protein derivative \nPRISMA – Preferred Reporting Items for Systematic Reviews and Meta-Analyses \nPSA: probabilistic sensitivity analysis \nQALYS: Quality Adjusted Life Years \nQFT: QuantiFERON-TB Gold \nTB: Tuberculosis \nTBI: Latent",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A list of abbreviations used in the economic modelling annex on tuberculosis skin tests.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1050",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1050",
      "pageStart": 1050,
      "pageEnd": 1050,
      "target": "../sections/page-1050.md",
      "narrative": null,
      "content": "Extended Summary \n \nBackground: Our systematic review on the economic impact of novel Mycobacterium tuberculosis specific skin \ntests for tuberculosis infection (TBST) conducted for this WHO meeting (Section A, Economic impact report) \nshows that studies are few and limited by study quality. However, our report does provide the most updated \nevidence on the cost-effectiveness of the use of TBST compared to currently used tests for tuberculosis \ninfection. We aim to use this evidence to parameterise an economic modelling study to assess whether TBST \nare affordable and feasible to implement und",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The extended summary explains that a Markov model was built to compare the cost-effectiveness of new antigen-based skin tests with the PPD tuberculin skin test and IGRA in Brazil, South Africa and the UK, focusing on DiaskinTest because only its maker supplied costs. Costs are from a health system perspective, and sensitivity analyses explore test prices and uncertainty in clinical and utility parameters.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1051",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1051",
      "pageStart": 1051,
      "pageEnd": 1051,
      "target": "../sections/page-1051.md",
      "narrative": null,
      "content": "Results: In Brazil, strategy 1 is cost saving compared to strategy 2 and 3. Compared to PPD-TST, DiaskinTest is \ncost-saving at $5.6 with an incremental gain of 0.024 QALYs per patient. Compared to IGRA, DiaskinTest is cost \nsaving at $8.4 with an incremental gain of 0.01 QALYs. In South Africa, DST is more cost saving than IGRA and \nPPD-TST. Compared to PPD-TST, DiaskinTest is cost-saving at $4.39 with an incremental gain of 0.0152 QALYs. \nCompared to IGRA, DiaskinTest is cost saving at $64.41 with an incremental gain of 0.0065 QALYs. In the UK, \nstrategy 1 is cost saving compared to strategy",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that DiaskinTest was cost saving with small QALY gains against both comparators in Brazil and South Africa, and against the skin test but not IGRA in the UK, where it remained cost-effective, and sets out how far test prices could rise and stay cost saving. The interpretation stresses that results depend on country-specific costing and limited evidence and that other new skin tests could not be modelled for lack of cost data.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1052",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1052",
      "pageStart": 1052,
      "pageEnd": 1052,
      "target": "../sections/page-1052.md",
      "narrative": null,
      "content": "factor likely to influence cost-effectiveness will be unit costs. Our sensitivity analysis suggests, regardless of \ntype, TBST could be both cost saving and cost-effective or simply cost-effective depending on the specific \ncountry context despite some variations of the unit costs.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Concludes that since accuracy appears similar across new skin tests, unit cost will mainly drive cost-effectiveness, which may vary by country.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1053",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1053",
      "pageStart": 1053,
      "pageEnd": 1053,
      "target": "../sections/page-1053.md",
      "narrative": null,
      "content": "Research Question \n \nAre Mycobacterium tuberculosis specific antigen-based skin tests for the detection of tuberculosis infection \n(TBST) cost-effective compared to currently available in vitro IGRA or the PPD-TST tests in reducing TB in \nBrazil, South Africa, and the UK? \n \nObjectives \n \n5) \nTo model the possible cost-effectiveness of index tests vs current tests by using unit costs of \nindex tests and their accuracy (sensitivity and specificity), to inform a Markov model simulating a \ncohort of individuals being offered TBI testing. Includes unit costs of index tests as supplied by the \nmanu",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "States the research question and objectives of the modelling: comparing antigen-based skin tests with IGRA and PPD skin tests in Brazil, South Africa and the UK using a Markov model and sensitivity analyses. The outcomes are cost-effectiveness ratios or net benefit per QALY and test costs per patient.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1054",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1054",
      "pageStart": 1054,
      "pageEnd": 1054,
      "target": "../sections/page-1054.md",
      "narrative": null,
      "content": "PRESENTATION OF THE ECONOMIC IMPACT REPORT \n \nThe final report is presented in two parts.  \nSection A: Abstract, methods and results for Objectives 1-2  \nSection B: Abstract, methods and result for Objectives 3-4.  \n \nThis document contains Section B. Section A is submitted as a separate attachment.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains that the economic report has two parts and that this document is Section B, with Section A submitted separately.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1055",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1055",
      "pageStart": 1055,
      "pageEnd": 1055,
      "target": "../sections/page-1055.md",
      "narrative": null,
      "content": "Economic evaluation of the interventions and cost-effectiveness analysis  \n \n \n \nModel scheme and characteristics \n \nWe created a Markov model simulating a cohort of individuals transitioning among different states, the cohort \nbeing comprised at baseline of individuals without active TB. Following the cascade of care, these individuals \nare tested for TB infection (TBI), if positive, they may initiate and eventually complete or interrupt treatment. \nConsequently, the model’s states are: (1) no tuberculosis infection, tested; (2) tuberculosis infection, tested; (3) \nno tuberculosis infection,",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the Markov model, which follows a cohort without active TB through testing for infection, treatment, completion or interruption, active TB, recovery or death, across eleven states. It is parametrized for the UK, Brazil and South Africa, and the modelling objectives are restated.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p1056-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 1056,
      "pageEnd": 1056,
      "target": "../images/img-p1056-1.webp",
      "narrative": "draft",
      "content": "Model diagram (Figure 1) of the TB infection care cascade. From a starting node 'Tested for TB infection', arrows branch to 'No TBI tested' (leading to 'No TBI treated', 'No TBI untreated' and 'No TBI Treatment interrupted') and 'TBI tested' (leading to 'TBI treated', 'TBI untreated' and 'TBI Treatment interrupted'). 'TBI treated' and 'TBI Treatment interrupted' both point to 'TBI untreated', which leads by a dashed arrow to 'Active TB'; each end state ends in a green triangle terminal node.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1056",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1056",
      "pageStart": 1056,
      "pageEnd": 1056,
      "target": "../sections/page-1056.md",
      "narrative": null,
      "content": "Figure 1. Model diagram. \nThe model describes the following states and steps in the TB infection care cascade: (1) no tuberculosis infection, tested; (2) \ntuberculosis infection, tested; (3) no tuberculosis infection, treated; (4) no tuberculosis infection, untreated; (5) no \ntuberculosis infection, treatment started but interrupted; (6) tuberculosis infection, treated; (7) tuberculosis infection, \nuntreated; (8) tuberculosis infection, treatment started but interrupted; (9) active TB; (10) no TB; (11) death. We start with \na cohort of individuals without TB. \n \n \nModel parameters \n \nThe proba",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the model diagram and parameters, starting with 100,000 people tested and followed over 20 annual cycles. Test accuracy comes from the authors' systematic review, the skin test's specificity is assumed equal to IGRA's, and the model covers the general population without separating by HIV status.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1057",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1057",
      "pageStart": 1057,
      "pageEnd": 1057,
      "target": "../sections/page-1057.md",
      "narrative": null,
      "content": "Table 1. Model parameters \nParameter \nValue \nSource \nPrevalence of TB infection in \nTB-negative individuals, \npercentage \nBrazil= 13.27% [95%CI= 9.4 – 18.1] \nUK= 1.76% \n[95%CI= 1.3 – 2.7] \nSouth Africa= 31.53% \n[95%CI= 28.8 – 36.1] \nLiterature 1 (see Table A1 \nsupplementary) \nPeople completing treatment \nafter initiation following a \npositive TBI result, \npercentage \nBrazil= 33.96% \nUK= 44.2% \nSouth Africa= 12.06% \nBrazil 2 \nUK, NHS 3 \nSouth Africa 4 \n \nPeople not initiating \ntreatment after testing \npositive for TBI, percentage \nBrazil= 39.62% \nUK=36% \nSouth Africa= 27.8% \nBrazil 2 \nUK, NHS 3",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 1 lists model parameters by country, including the prevalence of TB infection, treatment initiation, completion and interruption, progression to active TB, treatment efficacy, TB treatment coverage and recovery rates.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1058",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1058",
      "pageStart": 1058,
      "pageEnd": 1058,
      "target": "../sections/page-1058.md",
      "narrative": null,
      "content": "Death from Active TB \n(treated+untreated) \nBrazil = 11.1% [95%CI= 9.1–12.7] \nUK = 14.1% [95%CI= 10.3–16.4] \nSouth Africa = 9.4% [95%CI= 8.0–10.6] \n9,10 \nProbability of a true positive \ntest result if the patient has \nTB infection (sensitivity) \n \nTST: 88.24 (78.20–94.01) \nDiaskintest: 91·18 (81.72–95.98) \nIGRA test (QFT): 89.66 (78.83–95.28) \nIGRA test (T-spot): 90.91 (79.95–96.16) \nC-TB test (skin): 86·06 (82.39–89.07) \nTBST systematic review11 \nProbability of a true negative \ntest result if the patient does \nnot have TB infection \n(specificity) \nTST: 93.31 (90.22–95.48) \nDiaskintest: 99.15 (",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the parameters with TB death rates and the sensitivity and specificity of the skin tests and IGRA. It explains that DiaskinTest unit costs combine the manufacturer's price with standard skin test costs, and that the largest observed country costs were chosen as a conservative assumption.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1059",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1059",
      "pageStart": 1059,
      "pageEnd": 1059,
      "target": "../sections/page-1059.md",
      "narrative": null,
      "content": "account for the full cost of each visit and follow-up. The study assumes that test strategies involving TST \nrequire two clinic visits, whereas those involving only IGRAs (T-SPOT.TB or QFT-GIT or both) only require one \nclinic visit. The TST visits represent the day of PPD placement and a further visit for assessment of results any \nskin induration. The IGRA visit represents the phlebotomy visit. Testing strategies in the UK are in line with \nNICE 13. For South Africa, we consider the full screening strategy described in Mandalakas et al. (2012)14 for a \ncohort of child case contacts, which in",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes how testing and treatment costs were drawn from published studies for each country, and explains the QALY assumptions, 20-year horizon and country discount rates. It lists the sensitivity and scenario analyses, including a worst case in which one vial is used for a single patient.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1060",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1060",
      "pageStart": 1060,
      "pageEnd": 1060,
      "target": "../sections/page-1060.md",
      "narrative": null,
      "content": "approximatively $18.55, meaning an additional $16.93 (16.88 –16.98) on the DiaskinTest unit cost in \nthe three countries analysed. \nStrategies \nThree testing strategies were considered in this analysis: (i) DiaskinTest; (ii) PPD-TST test; (iii) IGRA, \nQuantiFERON-TB test (either Gold In-Tube or Gold-Plus). The model was parametrised for each of the three \nselected countries and ran considering these three testing strategies. Results from the DiaskinTest strategy \n(strategy 1) were compared against the other two, PPD-TST (strategy 2) and IGRA (strategy 3). We used \nDrummond’s checklist to asses",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives the extra cost of the single-use vial scenario and describes the three testing strategies compared, with results reported as cost-effectiveness ratios or net benefit against country thresholds. Table 2 lists unit costs of each test by country and utility values.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1061",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1061",
      "pageStart": 1061,
      "pageEnd": 1061,
      "target": "../sections/page-1061.md",
      "narrative": null,
      "content": "Utility loss due to active TB treatment adverse effects \nUtility loss due to TBI treatment \n0.17 \n0.2 \n \n \nNotes: Costs provided in 2021 USDs ($). \naUnit cost calculated by summing DiaskinTest costs as provided by the manufacturer (Tables B4−5) and PPD-TST associated \ncosts (excluding test cost) from literature review 12,14,17.  \nbThese costs represent the whole screening strategy including the costs of the tests (disposables, administration, reading, \nlaboratory technicians), two clinic visits and one chest radiograph. Screening strategies that include an IGRA also include \nthe cost of one ou",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the utility values and notes on test costs, then gives Table 3 on the cost of TB infection treatment regimens by country. Costs are shown as drug and staff costs in 2021 US dollars.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1062",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1062",
      "pageStart": 1062,
      "pageEnd": 1062,
      "target": "../sections/page-1062.md",
      "narrative": null,
      "content": "(supplementary material). \n \n \n \nTable 4 Costs per active DOT TB treatment (drug-sensitive), by country (inflated, expressed in 2021 USDs) \nCosts per active TB \ntreatment (6 months \ntreatment) \n \nDrug costs \nStaff costs (medical \nand nurse follow-\nup consultations) a \nChest X-ray, \nculture tests, and \nliver function \ntests costs h \n \nTotal cost per \ncase \nBrazil \n \n2 months RHZE \n$16.7 20 \n \n$332.3 b \n \n$41.93 e \n \n4 months RH \n$19.09 20 \n$410.52 \n6 months treatment \n$35.79 \n \nSouth Africa \n \n \n \n$64.68 c \n \n \n2 months RHZE \n$20.72 21 \n \n$221.53 f \n \n$332.83 \n4 months RH \n$25.90 21 \n6 months t",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table 4 gives the costs of treating drug-sensitive active TB under directly observed therapy in each country, with notes on visit and test costs. The results section begins with Brazil, where DiaskinTest was cost saving compared with both other tests.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p1063-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 1063,
      "pageEnd": 1063,
      "target": "../images/img-p1063-1.webp",
      "narrative": "draft",
      "content": "Incremental cost-effectiveness plane for Brazil (part of Figure 2): a scatter plot of Monte Carlo simulations with Incremental QALYs on the x-axis (−2 to 2) and Incremental Costs on the y-axis (about −20 to 20), comparing DiaskinTest versus PPD-TST (blue) and DiaskinTest versus IGRA (green). Both clouds overlap and are centred near zero QALYs with mostly small positive incremental costs.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p1063-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 1063,
      "pageEnd": 1063,
      "target": "../images/img-p1063-2.webp",
      "narrative": "draft",
      "content": "Incremental cost-effectiveness plane for South Africa (part of Figure 2): a scatter plot of Monte Carlo simulations with Incremental QALYs on the x-axis (−2 to 2) and Incremental Costs on the y-axis (about −50 to 100), comparing DiaskinTest versus PPD-TST (blue) and DiaskinTest versus IGRA (green). The green points (versus IGRA) lie mostly above zero cost, around 50, while the blue points (versus PPD-TST) are centred near zero cost.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1063",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1063",
      "pageStart": 1063,
      "pageEnd": 1063,
      "target": "../sections/page-1063.md",
      "narrative": null,
      "content": "saving at $8.4 with an incremental gain of 0.01 QALYs. As the intervention was cost saving, we calculated \nincremental net benefit (INB) rather than incremental cost-effective ratio (Table 5). Sensitivity analysis on unit \ncosts of TBST shows that, assuming same sensitivity and specificity values as DiaskinTest, a testing strategy \nusing any TBST would remain cost saving compared to PPD-TST for an increase in TBST unit cost of up to ~2-3 \ntimes the unit cost of DiaskinTest; a TBST unit cost of $16.06 (more than triple DiaskinTest’s unit cost of 5.33$) \nwould lead to an average expected cost pe",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the Brazil results, showing that any antigen skin test would stay cost saving against the PPD skin test up to about three times DiaskinTest's price. In the worst-case vial scenario it would no longer save costs but would remain cost-effective.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p1064-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 1064,
      "pageEnd": 1064,
      "target": "../images/img-p1064-1.webp",
      "narrative": "draft",
      "content": "Incremental cost-effectiveness plane for the UK (part of Figure 2): a scatter plot of Monte Carlo simulations with Incremental QALYs on the x-axis (−2 to 2) and Incremental Costs on the y-axis (about −400 to 400), comparing DiaskinTest versus PPD-TST (blue) and DiaskinTest versus IGRA (green). The blue points (versus PPD-TST) sit mostly at or above zero cost, while the green points (versus IGRA) sit mostly at or below zero cost.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1064",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1064",
      "pageStart": 1064,
      "pageEnd": 1064,
      "target": "../sections/page-1064.md",
      "narrative": null,
      "content": "Figure 2. Incremental cost-effectiveness planes of DiaskinTest versus PPD-TST (blue), and DiaskinTest versus IGRA (green), \nin Brazil, South Africa and UK. These figures show scatter plots of 1,000 Monte Carlo simulations for the mean incremental \ncosts and incremental effectiveness in quality-adjusted life years. \n \nSouth Africa \nStrategy 1 was cost saving compared to both strategy 2 and 3 (Figure 2). Compared to PPD-TST, DiaskinTest \nwas cost-saving at $4.39 with an incremental gain of 0.0152 QALYs. Compared to IGRA, DiaskinTest was cost \nsaving at $64.41 with an incremental gain of 0.0065 Q",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents cost-effectiveness planes and results for South Africa, where DiaskinTest was cost saving against both comparators and would remain so up to a modestly higher price. In the worst-case vial scenario it would remain cost-effective, and the UK results begin.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1065",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1065",
      "pageStart": 1065,
      "pageEnd": 1065,
      "target": "../sections/page-1065.md",
      "narrative": null,
      "content": "DiaskinTest was cost saving at $73.33 with an incremental gain of 0.0351 QALYs. Compared to IGRA, \nDiaskinTest showed an incremental gain of $15.80 and an incremental gain of 0.0266 QALYs, or $521.45 (95% \nCI 500.94 ̶ 545.07) per QALY, below the willingness-to-pay threshold of $20.223 per QALY16. Sensitivity analyses \non cost of the test show that, with same sensitivity and specificity values as DiaskinTest, a TBST strategy would \nbecome cost saving compared to IGRA for a decrease in the unit cost of the test to approximatively the same \nunit cost as IGRA. Meaning, a test unit cost of $149.0 w",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports UK results, where DiaskinTest was cost saving compared with the skin test and cost-effective compared with IGRA, staying cost-effective at prices up to about 2.35 times its own cost. Table 5 sets out modelled costs, QALYs and incremental net benefit for each strategy in the three countries.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1066",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1066",
      "pageStart": 1066,
      "pageEnd": 1066,
      "target": "../sections/page-1066.md",
      "narrative": null,
      "content": "IGRA \n624.91  \n12.8044  ̶\n 15.80 ̶\n 0.0303 ̶\n 628.56 \nNote. Positive incremental costs indicate the amount of money saved on average, when using DiaskinTest instead of PPD-\nTST or IGRA. Vice versa, negative incremental costs indicate the amount of money lost when using DiaskinTest. Similarly, \nnegative incremental QALYs indicate that strategies using PPD-TST or IGRA comport a loss in QALYs compared to \nDiaskinTest strategy. \na Mean per patient (discounted). \n \nInterpretation \n \nOur modelling results show that DiaskinTest potentially dominates both PPD-TST and IGRA in Brazil and South \nAfrica,",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Interprets the results, noting that DiaskinTest dominated the other tests in Brazil and South Africa but cost more than IGRA in the UK. It discusses limits such as manufacturer costs that vary with order size, wasted doses from opened vials, the provider-only perspective and differences between study and real-world costing.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1067",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1067",
      "pageStart": 1067,
      "pageEnd": 1067,
      "target": "../sections/page-1067.md",
      "narrative": null,
      "content": "effects. For this reason, unit costs and, consequently, final expected costs, can vary vastly according to the \nspecific activities chosen to define the service. The unit costs used in our analysis (Table 2) show large \ndifferences between countries because of this. UK’s unit costs are based on NHS tariffs which do not \ndifferentiate between activities but between staff members involved, thus UK costs are heavily affected by \nlocal salaries, resulting much larger than, for example, Brazil’s unit costs, which looked at the specific activity \nand the time required to complete it. South Africa’s",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains that unit costs vary widely between countries because of different costing methods and assumptions about clinic visits, and that other new skin tests could not be modelled because they are not yet marketed. It suggests future work on at-risk groups such as people with HIV, children, contacts and migrants.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1068",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1068",
      "pageStart": 1068,
      "pageEnd": 1068,
      "target": "../sections/page-1068.md",
      "narrative": null,
      "content": "References \n \n1. \nHouben RM, Dodd PJ. The global burden of latent tuberculosis infection: a re-estimation using \nmathematical modelling. PLoS medicine 2016; 13(10): e1002152. \n2. \nWysocki AD, Villa TCS, Arakawa T, et al. Latent tuberculosis infection diagnostic and treatment cascade \namong contacts in primary health care in a city of Sao Paulo State, Brazil: cross-sectional study. PloS one 2016; \n11(6): e0155348. \n3. \nNational Health System. National latent TB infection testing and treatment programme. 2021. \nhttps://www.england.nhs.uk/tuberculosis-programme/area-for-action-8-national-latent-t",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The first part of the reference list for the economic modelling annex, citing studies on latent TB burden, cascades of care, test accuracy and cost-effectiveness.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1069",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1069",
      "pageStart": 1069,
      "pageEnd": 1069,
      "target": "../sections/page-1069.md",
      "narrative": null,
      "content": "17. \nSteffen RE, Caetano R, Pinto M, et al. Cost-effectiveness of Quantiferon®-TB Gold-in-Tube versus \ntuberculin skin testing for contact screening and treatment of latent tuberculosis infection in Brazil. PloS one \n2013; 8(4): e59546. \n18. \nKim H, Hanrahan C, Martinson N, Golub J, Dowdy D. Cost-effectiveness of universal isoniazid preventive \ntherapy among HIV-infected pregnant women in South Africa. The International Journal of Tuberculosis and \nLung Disease 2018; 22(12): 1435-42. \n19. \nWhite P, Jit M. What is the Cost-Effectiveness of Latent Tuberculosis Infection (LTBI) Treatment with \nDi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the reference list, citing cost-effectiveness and treatment cost studies from Brazil, South Africa and the UK and WHO TB treatment guidance.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1070",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1070",
      "pageStart": 1070,
      "pageEnd": 1070,
      "target": "../sections/page-1070.md",
      "narrative": null,
      "content": "SUPPLEMENTARY DOCUMENTATION \n \nEconomic evaluation of the interventions and cost-effectiveness analysis \n \nModel parameters additional information \n \nTable B1. Calculation of the TBI progression \nISO \ncode  \n \nLTBI estimates 2014 1 \n  \nPopulation estimates 2014 9 \n \n \nLTBI/population \n(%) \nBRA \n26,900,000  \n[19,100,000 - 36,700,000] \n202,763,744 \n13.27 \n[9.4 – 18.1] \nGBR \n1,150,000  \n[ 867,000 - 1,790,000] \n65,423,048 \n1.76 \n[1.3 – 2.7] \nZAF \n17,200,000  \n[15,700,000 - 19,700,000] \n54,544,184 \n31.53 \n[28.8 – 36.1] \n \n \nTable B2. Progression from LTBI to active TB: population estimates \nProgres",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Supplementary tables give the calculated prevalence of TB infection in Brazil, the UK and South Africa and estimated rates of progression from infection to active TB.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-050",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table B1. Calculation of the TBI progression",
      "pageStart": 1070,
      "pageEnd": 1070,
      "target": "../tables/table-b1.csv",
      "narrative": "draft",
      "content": "Estimated number of people with latent TB infection (LTBI) in 2014, total population in 2014 and the resulting LTBI prevalence (%) with uncertainty ranges, for Brazil (BRA), the United Kingdom (GBR) and South Africa (ZAF).",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/table-051",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table B2. Progression from LTBI to active TB: population estimates",
      "pageStart": 1070,
      "pageEnd": 1071,
      "target": "../tables/table-b2.csv",
      "narrative": "draft",
      "content": "Probabilities (with 95% CI) of progression from latent TB infection to active TB used as model parameters, for recent and remote infection in the general population and by HIV status, with the reference sources.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1071",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1071",
      "pageStart": 1071,
      "pageEnd": 1071,
      "target": "../sections/page-1071.md",
      "narrative": null,
      "content": "Proportion of TST or IGRA positive, untreated, HIV \npositive individuals progressing to active TB (post‐\nexposure TB) over 20 years \n0.40 (0.20‐0.80) \n22,24 \nNotes:  All individuals in the model are assumed to belong to the general population with recent TBI infection. Differentiation \nbetween HIV status or duration of infection are not included in the model but these values are reported in this table to show \nhow the model could be adapted  to study different scenarios by modifying its parametrisation. \n \n \nTable B3. Costs per TBI treatment, by country (not inflated, different currencies) \nCo",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives progression rates for people with HIV and explains that the model could be adapted for such groups. Table B3 shows uninflated TB infection treatment costs by regimen and country.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-052",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table B3. Costs per TBI treatment, by country (not inflated, different currencies)",
      "pageStart": 1071,
      "pageEnd": 1072,
      "target": "../tables/table-b3.csv",
      "narrative": "draft",
      "content": "Drug costs and staff (medical and nurse follow-up consultation) costs of isoniazid preventive treatment regimens of 3, 6, 9 and 12 months (3H–12H) in Brazil, South Africa and the United Kingdom, in local currency and cost year, with the source and unit consultation costs.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1072",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1072",
      "pageStart": 1072,
      "pageEnd": 1072,
      "target": "../sections/page-1072.md",
      "narrative": null,
      "content": "treatment with isoniazid, 12= 12 months treatment with isoniazid, 2. a Calculated considering one medical check plus the \nnumber of follow-up checks administered by nurses depending upon the regimen, it is as follows: 3H: 2 nurse consultations; 6H: \n5; 9H: 8; 12H: 11. \n*Drug costs were obtained from the NHS drug tariff (2014) (for H and R) and British National Formulary (2013) (for Pz); \nquantities of drugs used for each regimen were supplied by NICE. Staff costs were calculated from the amounts of staff time \nrequired for administration of LTBI treatment, based on GDG advice, and the cost of",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains how treatment staff costs were calculated, then Tables B4 and B5 give DiaskinTest vial prices by order volume and delivery costs by country. They work these out as a cost per patient, using manufacturer data.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-053",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table B4. Cost of the Diaskin Test per vial for different delivery volumes and considering that 1 vial is used for 15 patients",
      "pageStart": 1072,
      "pageEnd": 1072,
      "target": "../tables/table-b4.csv",
      "narrative": "draft",
      "content": "Price per vial of the Diaskin Test by order volume (from 667 to 150,000 or more packs), and delivery costs to Brazil, the UK and South Africa for the two smallest volumes.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/table-054",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table B5. Full costs including delivery, and cost per patient by country",
      "pageStart": 1072,
      "pageEnd": 1072,
      "target": "../tables/table-b5.csv",
      "narrative": "draft",
      "content": "Full Diaskin Test costs for 667-pack and 6,667-pack orders, excluding delivery and including delivery to Brazil, the UK and South Africa, with the number of patients treated and the resulting price per patient.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1073",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1073",
      "pageStart": 1073,
      "pageEnd": 1073,
      "target": "../sections/page-1073.md",
      "narrative": null,
      "content": "210\nTable B6: Drummond checklist for the quality of our study: Cost-effectiveness analyses for Novel skin tests for diagnosing \nTBI \n \nDrummond Checklist Questions \nOur study \n1.Was a well-defined question posed in answerable form? \n  \n1.1. Did the study examine both costs and effects of the service(s) or programme(s)? \nYes \n1.2. Did the study involve a comparison of alternatives? \nYes \n1.3. Was a viewpoint for the analysis stated and was the study placed in any decision-making context? \nYes \n  \nAdult population \n2. Was a comprehensive description of the competing alternatives given? \n \n2.1. W",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Applies the Drummond checklist to the authors' own modelling study, which compared alternatives and drew test accuracy from clinical studies and reviews.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-055",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table B6: Drummond checklist for the quality of our study: Cost-effectiveness analyses for Novel skin tests for diagnosing TBI",
      "pageStart": 1073,
      "pageEnd": 1075,
      "target": "../tables/table-b6.csv",
      "narrative": "draft",
      "content": "Self-assessment of the review's own cost-effectiveness analysis of novel skin tests for diagnosing TB infection against the Drummond checklist. Rows are the checklist questions, grouped under ten main criteria; the single column gives the answer and comments for this study.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1074",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1074",
      "pageStart": 1074,
      "pageEnd": 1074,
      "target": "../sections/page-1074.md",
      "narrative": null,
      "content": "211\n4.2. Did it cover all relevant viewpoints? \nYes \n4.3. Were the capital costs, as well as operating costs, included? \nDiagnostic, treatment, \nstaff, laboratory, \nmaterial, and operating \ncosts included \n  \n \n  \n \n5.    Were costs and consequences measured accurately in appropriate physical units? \n \n5.1. Were any of the identified items omitted from measurement?  \nNo \n  \n \nIf so, does this mean that they carried no weight in the subsequent analysis? \n \n5.2. Were there any special circumstances (e.g., joint use of resources) that made measurement difficult? Were \nthese circumstances handled",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the self-assessment, noting that diagnostic, treatment, staff and laboratory costs were included and that costs were discounted with justification.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1075",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1075",
      "pageStart": 1075,
      "pageEnd": 1075,
      "target": "../sections/page-1075.md",
      "narrative": null,
      "content": "212\n9.1. If data on costs and consequences were stochastic (randomly determined sequence of observations), were \nappropriate statistical analyses performed? \nYes \n9.2. If sensitivity analysis was employed, was justification provided for the range of values (or for key study \nparameters)? \nYes \n9.3. Were the study results sensitive to changes in the values? \nYes \n  \n \n  \nSensitive to treatments \nreceived. \n10.    Did the presentation and discussion of study results include all issues of concern to users? \n \n10.1. Were the conclusions of the analysis based on some overall index or ratio of costs",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Completes the self-assessment, reporting appropriate statistical and sensitivity analyses and discussion of generalizability and implementation.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1076",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1076",
      "pageStart": 1076,
      "pageEnd": 1076,
      "target": "../sections/page-1076.md",
      "narrative": null,
      "content": "Web Annex F \nQualitative evidence for the use of Mycobacterium \ntuberculosis antigen-based skin tests \n \n \nReview team: \nAmy Hai Yan Chan1,2 , Kate Loveys1, Matthew Quaife3, Amy Clarke2, Liliya \nEugenevna Ziganshina4, Irina Kontsevaya5, Chloe Tuck6, Elena Surkova7, Ting \nTing Wang2, Liu (Victoria) Wan-Hsin3, Yohhei Hamada2, Molebogeng X \nRangaka2  \n \n1. The University of Auckland, Auckland, New Zealand \n2. University College London, London, United Kingdom \n3. London School of Hygiene and Tropical Medicine, United Kingdom \n4. Cochrane Russia, Russian Federation \n5. Research Center Borstel, Germ",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "The title page of Web Annex F, a review of qualitative evidence on tuberculosis antigen-based skin tests, naming the review team prepared for the February 2022 WHO guideline meeting.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1077",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1077",
      "pageStart": 1077,
      "pageEnd": 1077,
      "target": "../sections/page-1077.md",
      "narrative": null,
      "content": "0",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A page containing only a stray character.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1078",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1078",
      "pageStart": 1078,
      "pageEnd": 1078,
      "target": "../sections/page-1078.md",
      "narrative": null,
      "content": "1 \n \nAbbreviations \nBCG – Bacillus Clamette-Guérin  \nDCE – Discrete Choice Experiment \nGRADE-CERQual – Confidence in the Evidence from Reviews of Qualitative Research \nHIC – High Income Country \nHIV – Human Immunodeficiency Virus \nIGRA – Interferon gamma release assays  \nLMIC – Low-Middle Income Country \nPPD-TST – Purified protein derivative tuberculin skin tests \nPRISMA – Preferred Reporting Items for Systematic Reviews and Meta-Analyses \nQFT-gold – QuantiFERON-TB Gold   \nTBI – Tuberculosis infection \nTBST – Mycobacterium tuberculosis specific antigen-based skin tests  \nWHO – World Health Org",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists abbreviations and opens the executive summary, explaining that end-user values, acceptability, feasibility and equity effects of new TB skin tests are unknown. It describes three work packages: systematic reviews of qualitative evidence, 20 interviews with providers and people with TB infection, and an online discrete choice experiment. Confidence in evidence was rated with GRADE-CERQual.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1079",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1079",
      "pageStart": 1079,
      "pageEnd": 1079,
      "target": "../sections/page-1079.md",
      "narrative": null,
      "content": "2 \n \nwere developed for consumers and providers to represent the different potential determinants of \nchoice to either, and a minimum sample size of 45 was deemed sufficient to draw statistical \ninference in both designs. Choice data were analysed using multinomial logistic regression models, \nand willingness-to-accept analyses used to compare results within and across designs.  \nResults:  \nCharacteristics of data included in each work packages \nFor work package A, four studies were identified that met the inclusion criteria for both systematic \nreviews. From Review A on specific TBSTs, only o",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that the reviews found only four relevant studies, of low to moderate confidence, while the interviews covered 20 people and the choice experiment 234 respondents, mainly from India, the United States and South Africa. Providers and consumers valued accuracy, convenience, a positive experience, low cost and low resource needs. They preferred tests that avoid follow-up visits and specialist staff and use existing infrastructure.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1080",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1080",
      "pageStart": 1080,
      "pageEnd": 1080,
      "target": "../sections/page-1080.md",
      "narrative": null,
      "content": "3 \n \nAcceptability \nQualitative data from systematic reviews and end-user interviews suggests TBST were perceived to \nhave greater specificity and sensitivity than the TST. Having greater test accuracy was deemed \ndesirable to avoid the negative consequences of false positives or negatives. However, the TBST was \nexpected to have many of the same limitations as skin tests in terms of patient experience (e.g., a \nreturn visit required, discomfort, a welt on the arm, stigma) compared to the IGRA. The IGRA was \ndeemed the preferred test option in countries which already have the IGRA in use, as t",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that new skin tests were seen as more accurate than the tuberculin skin test but sharing its drawbacks compared with IGRA, which remained preferred where already in use. They appear feasible and low-cost where the tuberculin skin test is already used and are unlikely to create new equity problems. The conclusion finds them more acceptable than and as feasible as the tuberculin skin test, with likely equity benefits.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1081",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1081",
      "pageStart": 1081,
      "pageEnd": 1081,
      "target": "../sections/page-1081.md",
      "narrative": null,
      "content": "4 \n \npreferrable for use in more resource-limited settings compared to the TST due to the higher \naccuracy whilst retaining portability and ability to administer in a community / primary care setting. \nTest preferences were influenced by end-user values such as test accuracy, convenience, cost, \npositive patient experience, and resource requirements. IGRA remain the preferred test option in \nhigh-income countries with existing infrastructure to support their use; it is unlikely that widespread \nuse of the IGRA would change in these settings where IGRA is mostly the standard of care, \nparticula",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Sets out the background to a study of end-user values and preferences on tests for TB infection, comparing the tuberculin skin test (TST) and interferon-gamma release assays (IGRA). It notes that IGRAs are preferred in high-income settings partly because they need only one visit. It then explains that IGRAs cost more and need laboratories, phlebotomy and trained technicians, and are often run in batches.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1082",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1082",
      "pageStart": 1082,
      "pageEnd": 1082,
      "target": "../sections/page-1082.md",
      "narrative": null,
      "content": "5 \n \nsystem costs; thus, it is not necessarily true that results would be immediate. IGRA are also more \ninvasive, requiring venepuncture, which may impair acceptability.  \n \nOver the past decade, Mycobacterium tuberculosis specific antigen-based skin tests (TBST) have been \ndeveloped with the aim of improving diagnostic performance while using the same platform as the \nPPD-TST. These antigen-based skin tests include the C-Tb (Staten Serum Institut, Denmark), the \nDiaskintest (Generium, Russia), and the ESAT6-CFP10 test (Anhui Zhifei Longcom, China). Emerging \nevidence suggests that specific a",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the drawbacks of IGRAs and introduces newer TB antigen-based skin tests (TBST) such as C-Tb, Diaskintest and the ESAT6-CFP10 test, which may match IGRA accuracy at lower cost. It then states the project's aim of exploring end-user preferences and trade-offs through a systematic review, interviews and a discrete choice experiment.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1083",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1083",
      "pageStart": 1083,
      "pageEnd": 1083,
      "target": "../sections/page-1083.md",
      "narrative": null,
      "content": "6 \n \nthe relevant section describing each of the following work packages: \nA. Systematic review of qualitative evidence \nB. End-user interviews \nC. Discrete choice experiment (DCE)  \n \nAn expert technical advisory panel comprised of multidisciplinary experts with skills in global health, \ninfectious diseases and behavioural medicine contributed to the design of all three studies.",
      "truncated": false,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists the three work packages of the TB skin test preference study (qualitative evidence review, end-user interviews and a discrete choice experiment) and notes the role of a multidisciplinary expert advisory panel.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1084",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1084",
      "pageStart": 1084,
      "pageEnd": 1084,
      "target": "../sections/page-1084.md",
      "narrative": null,
      "content": "7 \n \n3.1 Work Package A-Systematic review \nThis activity intends to answer the following objectives: \nObjective A: To conduct a systematic review that summarizes the evidence on end-user \nvalues and preferences regarding the acceptability and feasibility of antigen-based skin \ntests for TB infection, compared to current tests of infection (IGRA, TST). \nObjective C: To evaluate any potential impacts of TBST on health equity \n \nTwo systematic reviews were conducted that synthesized the qualitative research evidence to date \non end-user values and preferences for the use of TBST for latent TB inf",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces the first work package, two systematic reviews of qualitative evidence on end-user views of antigen-based skin tests compared with IGRA and TST, and their possible equity effects. It describes the databases searched, including Russian and Chinese sources, plus manual searches and a WHO public call for data.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1085",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1085",
      "pageStart": 1085,
      "pageEnd": 1085,
      "target": "../sections/page-1085.md",
      "narrative": null,
      "content": "8 \n \nantigens). \nGrey literature \nGrey literature was not searched as part of the review.   \n \n3.1.2 Eligibility criteria \nStudy design \nIncluded studies were primary research studies that used qualitative methodologies or study \ndesigns. Studies were required to use qualitative methods for data collection (e.g., interviews, focus \ngroups, open-ended survey questions) and analysis (e.g., thematic analysis, grounded theory). Mixed \nmethods studies that used qualitative methods for data collection and analysis were included. \nSurvey studies that presented qualitative free text that was collected",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Sets out the eligibility criteria for the qualitative reviews, which included studies using qualitative data collection and analysis and excluded those analysing qualitative data quantitatively. It also covers the populations, tests and settings considered and how two reviewers screened the results in several languages.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1086",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1086",
      "pageStart": 1086,
      "pageEnd": 1086,
      "target": "../sections/page-1086.md",
      "narrative": null,
      "content": "9 \n \nQualitative data pertaining to test acceptability, feasibility, and health equity were extracted from \nthe included papers independently in duplicate by two raters (AC, KL) using Covidence software. \nData were also extracted on study characteristics, including publication language, country, type of \nfacility (including public/private, rural/urban), study objectives, methods of data collection and \nanalysis, study duration, conceptual framework, conflicts of interest, study participant type and \nmean age, percentage male, ethnicity, diagnostic tests involved, programmatic features of the",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes how qualitative data on acceptability, feasibility and equity, plus study characteristics, were extracted in duplicate. It explains the thematic synthesis approach and the use of GRADE-CERQual to judge confidence in each review finding.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p1087-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 1087,
      "pageEnd": 1087,
      "target": "../images/img-p1087-1.webp",
      "narrative": "draft",
      "content": "PRISMA flow chart (Figure A1, Review A) in two panels. English/Chinese literature: 1317 records in EMBASE, 1424 in Medline, 282 in the Chinese Biomedical Literature Database and 565 in China National Knowledge Infrastructure databases give 2271 records after removing duplicates; 2168 that do not report new skin tests are excluded, leaving 103 records for second screening and 0 records eligible. Russian literature: 405 records in e-library, 141 deemed not relevant, 264 for full-text review, all 264 excluded, giving 0 records potentially eligible; separately, 1 paper from a WHO public call for data yields 1 record eligible.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p1087-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 1087,
      "pageEnd": 1087,
      "target": "../images/img-p1087-2.webp",
      "narrative": "draft",
      "content": "PRISMA flow chart (Figure A1, Review B): 123 studies identified from MEDLINE, EMBASE, CINAHL and PsychInfo; 88 records after duplicates removed, of which 69 were deemed irrelevant; 19 full-text studies assessed for eligibility, with 16 excluded (11 wrong study design, 4 wrong intervention, 1 wrong publication type); 3 studies in final inclusion.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1087",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1087",
      "pageStart": 1087,
      "pageEnd": 1087,
      "target": "../sections/page-1087.md",
      "narrative": null,
      "content": "10 \n \n3.1.4 Results from Systematic Review A and B \n3.1.4.1 Results of the Search \nFour studies were identified that met the inclusion criteria (see Figure A1 for Preferred Reporting \nItems for Systematic Reviews and Meta-Analyses (PRISMA) flow charts). Review A included papers \nfocused on specific TBST compared to the IGRA and/or the TST, while Review B focused on papers on \nthe IGRA and TST compared to each other or without a comparator. From Review A for TBSTs, only \none study was identified18, which came from a WHO public call for data relating to the feasibility and \nacceptability of anti",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that four studies met inclusion criteria: one on the Diaskintest from a WHO call for data and three on IGRA and TST, published between 2014 and 2021, with PRISMA flow charts.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1088",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1088",
      "pageStart": 1088,
      "pageEnd": 1088,
      "target": "../sections/page-1088.md",
      "narrative": null,
      "content": "11 \n \n3.1.4.2 Description of the Studies \nA table summarizing the Characteristics of the included studies identified from Review A and B are \npresented in Table A1.  \nTest \nFrom Review A, only one study reported on a TBST which was the Diaskintest18, which was from the \nWHO public call for data. The Diaskintest was evaluated as part of a large-scale screening program \nfor children and adolescents in the Penza region of the Russian Federation. From Review B,  there \nwere three published studies reported on the IGRA and the TST19,20,21. Two studies focused on the \nuse of the IGRA and TST as part",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the four included studies, from a mass child screening programme in Russia's Penza region to HIV clinic services and immigrant screening in the Netherlands. It summarises their methodological limitations, such as missing participant quotes and unreported ethics. It begins the GRADE-CERQual assessment of confidence in the findings.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1089",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1089",
      "pageStart": 1089,
      "pageEnd": 1089,
      "target": "../sections/page-1089.md",
      "narrative": null,
      "content": "12 \n \nhealth equity), 3 findings (end user values and preferences, acceptability, feasibility) were rated with \nmoderate confidence based on the findings from Review A and B. Health equity received a low \nconfidence rating. Appendix B provides a detailed summary table of the GRADE-CERQual analysis. \nThere were serious concerns about data adequacy for the feasibility and health equity findings due \nto the small number of studies supporting each finding and the low data richness (i.e., small amount \nof relevant data) in 3 of 4 studies (1 from Review A18, and 2 from Review B20, 21). Moreover, the",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports moderate confidence in the review findings on values, acceptability and feasibility and low confidence for health equity, citing small numbers of studies, thin data and focus on narrow subgroups.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1090",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1090",
      "pageStart": 1090,
      "pageEnd": 1090,
      "target": "../sections/page-1090.md",
      "narrative": null,
      "content": "0 \n \nTable A1: Systematic Reviews A and B- Characteristics of Included Studies \nStudy ID \nPublication \ntype \nTBI Tests \nCountry \nFacility Type \nPublic/ \nPrivate \nRural/ \nUrban \nParticipants \nProgrammatic features \nComparato\nrs \nData \ncollection \nStudy \nduration \nMethod of data \nanalysis \nConceptual \nframework \nConflicts of Interest \nReview A findings on TBST \nAksenova \n2021 \n(Review A) \nReport \n(WHO call \nfor data) \n \nDiaskintest \nRussian \nFederation \nNot reported \nNot \nreported \nNot \nreported \nParents of \nchildren and \nadolescents \nundergoing TB \nscreening. \nDemographic \ndata not \nreported.",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "A table of characteristics of the four studies included in the qualitative reviews, covering test type, country, setting, participants, data collection, analysis, frameworks and conflicts of interest.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-056",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table A1: Systematic Reviews A and B- Characteristics of Included Studies",
      "pageStart": 1090,
      "pageEnd": 1090,
      "target": "../tables/table-a1.csv",
      "narrative": "draft",
      "content": "Characteristics of the four qualitative studies included in Systematic Review A (TBST) and Systematic Review B (TST and IGRA): publication type, tests, country, facility type and setting, participants, programmatic features, comparators, data collection, study duration, method of analysis, conceptual framework and conflicts of interest.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p1091-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 1091,
      "pageEnd": 1091,
      "target": "../images/img-p1091-1.webp",
      "narrative": "draft",
      "content": "Diagram of themes and subthemes (Figure A2) arranged in three columns of labelled circles with different shading. Acceptability: Consumer demographics, Consumer experience, Accuracy / certainty of results, Suitability in specific subgroups. Feasibility: Consumer experience, Cost, Resource requirements, Health system setting. Equity: Access and transport issues, Scalability for large scale screening / suitability for subgroups.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p1091-2",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 1091,
      "pageEnd": 1091,
      "target": "../images/img-p1091-2.webp",
      "narrative": "draft",
      "content": "Diagram titled 'End user values and preferences' (part of Figure A2) with five labelled circles: Accuracy / certainty of results, Convenience, Consumer experience, Cost and Resource requirements.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1091",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1091",
      "pageStart": 1091,
      "pageEnd": 1091,
      "target": "../sections/page-1091.md",
      "narrative": null,
      "content": "0 \n \n3.1.4.5 Review Findings \nOur qualitative evidence synthesis from Work Package A draws on the findings from both Systematic \nReviews A and B. Overall, there were limited data available hence the decision to combine findings \nfrom both Reviews, with Review A on TBST retrieving only one relevant data report, and Review B on \nTST and IGRA retrieving only three papers.  \nThe findings identified factors contributing to end-user values/preferences, acceptability, feasibility, \nand health equity impacts of TBST, IGRA and TST, corresponding to our four objectives. We present \neach finding under ea",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces the combined qualitative findings on test values, acceptability, feasibility and equity, noting how limited the data were. It reports that accuracy shaped preferences, with IGRA often favoured over TST for being less subjective, although some clinic staff doubted IGRA accuracy in people with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1092",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1092",
      "pageStart": 1092,
      "pageEnd": 1092,
      "target": "../sections/page-1092.md",
      "narrative": null,
      "content": "1 \n \nthat there was a high frequency of indeterminate results from the IGRA, although this was not \nsupported in the clinic data where only 1% of indeterminate results were received20.  \n“Sometimes we’re not sure if they measure it correctly, or if the TST was done correctly, did \nthey take enough, was it inserted sub-dermally? There can come human errors as well. With \nblood, I think the result is the result.” (South Africa) \nConvenience was found to affect test preferences in all 4 studies18,19,20,21. Although limited data was \nprovided, the Diaskintest was reported to be convenient for use",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports how convenience factors such as clinic waits, transport and crowding affected preferences for IGRA and TST, while the Diaskintest was seen as convenient for mass screening. It then discusses people-related factors in test acceptability, including discomfort from the TST, illustrated with quotes from South Africa.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1093",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1093",
      "pageStart": 1093,
      "pageEnd": 1093,
      "target": "../sections/page-1093.md",
      "narrative": null,
      "content": "2 \n \nTest Factors \nAcceptability was affected by test-related factors, including the accuracy of the test, the certainty of \nresults, and the suitability of the test for use in particular subgroups (e.g., children and adolescents, \npersons with HIV, needle-phobic individuals). \nTBST and IGRA were perceived as having greater specificity than the TST18,21. The TST was described \nas having a high rate of false positives; thus the IGRA was deemed preferable in terms of \nspecificity21. From Review A, the high specificity of the TBST was found to improve the motivation of \nparents to perform TB test",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes test-related factors affecting acceptability, with TBST and IGRA seen as more specific than TST and the Diaskintest's specificity encouraging parents to test their children. It notes no contextual factors affected acceptability. It then opens the feasibility findings, which came only from IGRA and TST studies and were rated with moderate confidence.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1094",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1094",
      "pageStart": 1094,
      "pageEnd": 1094,
      "target": "../sections/page-1094.md",
      "narrative": null,
      "content": "3 \n \nAspects of the individual experience were found to affect the feasibility of the IGRA and TST. A \ncommon feasibility concern for the TST, based on findings from Review B, were that consumers were \nrequired to visit the clinic a second time within 72 hours to read the test result19,20,21. Kerrigan et al. \ndescribed how in the North West Province of South Africa, many consumers experienced difficulties \nwith visiting a clinic due to the long journey and wait time involved, as well as issues with accessing \ntransportation19. Similar concerns were echoed in HIV clinics in the United States, a",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains that the need for a second clinic visit to read the TST was a key feasibility concern, especially where travel and waiting times are long, while IGRA needs only one visit. It reports provider views from South Africa that using routine blood draws could ease follow-up problems, along with a patient's worry about delays.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1095",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1095",
      "pageStart": 1095,
      "pageEnd": 1095,
      "target": "../sections/page-1095.md",
      "narrative": null,
      "content": "4 \n \nthe low frequency with which they would typically have a blood draw: \n“While there was general consensus on this issue, one patient did relay a concern about \ntiming, as many patients may only have a CD4 blood draw 1–2 times a year. So while this \npatient thought screening the CD4 blood draw for TB was a good idea, he wouldn’t want that \nto delay finding out he had TB, as he would want to know that information as soon as \npossible. This issue was also raised by clinical providers, and a few in turn mentioned that \nscreening the blood for TB could potentially complement, as opposed to repl",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the feasibility findings with concerns about infrequent blood draws delaying detection, and notes that public funding and supply of the TST helped feasibility in South Africa. It then summarises health equity issues, including access and transport barriers, staffing, scalability for mass screening and suitability for children, people with HIV and needle-phobic people.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1096",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1096",
      "pageStart": 1096,
      "pageEnd": 1096,
      "target": "../sections/page-1096.md",
      "narrative": null,
      "content": "5 \n \nIn terms of access and transport issues, the return visit required to read the TST result was focused \non as a major accessibility issue in 3 studies19,20,21. This included at-risk groups in both middle- and \nhigh-income countries (South Africa, United States). Kerrigan et al., which focused on an HIV clinic in \na middle-income country, documented how many consumers experienced challenges with \ntransportation and wait times in understaffed clinics, serving as a barrier for return visits19. \nChallenges with an adequate number of specialized staff to administer the TST and IGRA were \ndocume",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Expands on equity issues, highlighting the return visit for TST reading as a barrier and shortages of trained staff for TST and IGRA. It notes the Diaskintest was reported as suitable for mass screening of children, while some US clinicians doubted IGRA results in people with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1097",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1097",
      "pageStart": 1097,
      "pageEnd": 1097,
      "target": "../sections/page-1097.md",
      "narrative": null,
      "content": "6 \n \n3.1.5 Summary of Work Package A findings from the Systematic \nReviews \nOverall, there were limited data from the reviews on TBSTs, with only 4 studies identified. \nImportantly, these studies were conducted in very specific healthcare settings and populations, \nlimiting our confidence in the data quality and the generalisability of these findings. Most of the data \navailable related to TST and IGRA. Despite the limited data available, we identified key factors \nrelating to our four objectives of end user values/preferences, acceptability, feasibility and health \nequity. Important end-user",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarises the systematic review findings: data on antigen-based skin tests were scarce and setting-specific, and key values included accuracy, convenience, patient experience, cost and resources. It explains how these findings shaped the interview guide and discrete choice experiment.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1098",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1098",
      "pageStart": 1098,
      "pageEnd": 1098,
      "target": "../sections/page-1098.md",
      "narrative": null,
      "content": "7 \n \nWork Package B-Qualitative user interviews \nThis activity intends to answer following objectives: \nA qualitative interview study was conducted with 20 participants who were individuals living with TB \nor providers involved in TB testing and /or care of individuals living with TB. Participants were \nrecruited from the research team’s global networks across high and low-middle income countries, \nwith support from WHO. Participants were approached through organizations/groups as below: \n• \nTB Treatment Action Group \n• \nTB Civil Societies \n• \nMembers of the WHO Guideline Development Group for",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces the second work package, qualitative interviews with 20 people with TB and providers recruited through global networks and organisations such as TB civil society groups and the Stop TB Partnership. It describes the email invitation and the interview topics, tailored to consumers and providers.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1099",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1099",
      "pageStart": 1099,
      "pageEnd": 1099,
      "target": "../sections/page-1099.md",
      "narrative": null,
      "content": "8 \n \n1. Current approach to using skin tests as part of the TB diagnostic process; \n2. End-user values and preferences with regard to TB testing, from their previous experience \nwith skin-based tests – focusing on ease of use and process for TBST, IGRA, and TST; \n3. End-user values and preferences considering the challenges and benefits of skin-based \ntests, particularly TBST, which test(s) they would prefer and reasons for the preference; \n4. Feasibility of implementing TBST in current practice; \n5. Overall acceptability of TBST- burden, attitude, understanding, test efficacy; \n6. Potential i",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists the interview topics, from current skin test use to preferences, feasibility, acceptability and equity, and explains that interviews were transcribed and coded in NVivo. A table and figure describe the 13 providers and 7 consumers from high- and low/middle-income countries.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p1100-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 1100,
      "pageEnd": 1100,
      "target": "../images/img-p1100-1.webp",
      "narrative": "draft",
      "content": "Diagram of end-user values and preferences (Figure B2) shown as five labelled circles: Accuracy / certainty of results, Convenience, Consumer experience, Cost and Resource requirements.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1100",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1100",
      "pageStart": 1100,
      "pageEnd": 1100,
      "target": "../sections/page-1100.md",
      "narrative": null,
      "content": "9 \n \n \n1.2.3 Key Findings  \nThe findings from Work Package A were used to inform the coding framework and analyses \nundertaken in Work Package B – as such, the same subthemes of people, test, and contextual factors \nwere used to explore acceptability and feasibility of the tests, with health equity and end-user values \nand preferences as separate subthemes. The full coding framework used can be found in Appendix \nE. As before, the same shading scheme will be used in the figures to illustrate themes and \nsubthemes for acceptability and feasibility: Dark solid grey refers to people factors; squa",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains that the interview analysis reused the people, test and contextual factor framework from the review. It reports that participants valued accuracy and saw TBST and IGRA as more accurate than TST, with downstream benefits such as avoiding unnecessary treatment and stress.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1101",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1101",
      "pageStart": 1101,
      "pageEnd": 1101,
      "target": "../sections/page-1101.md",
      "narrative": null,
      "content": "10 \n \nperceived to have lower accuracy and were associated with downstream challenges by end-users. \nThese included greater expense and psychological stress.  \n“The big issue I think is the false positives with the current tests that we use and that then \nleads to a whole chain of events which I think is very expensive and stressful for people \nbecause if you’re told you might have TB or you probably have got TB and then you’ve got a \nwhole set of tests, if the novel test is more accurate and it comes back and it says no you \ndidn’t, that saves a lot of stress and all the hospital resources an",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports interviewees' concerns that false-positive TST results cause costly and stressful follow-up. It describes how IGRA was seen as more convenient because it needs one visit. It also notes that IGRAs demand more resources, such as laboratories, phlebotomy and trained staff, than skin tests.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p1102-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 1102,
      "pageEnd": 1102,
      "target": "../images/img-p1102-1.webp",
      "narrative": "draft",
      "content": "Diagram of people-related factors in test acceptability (Figure B3) shown as seven grey circles: Consumer demographics, Consumer experience, Perceptions of skin tests, Knowledge, Familiarity, Usefulness and Provider-consumer communication.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1102",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1102",
      "pageStart": 1102,
      "pageEnd": 1102,
      "target": "../sections/page-1102.md",
      "narrative": null,
      "content": "11 \n \ntest equipment are still required, and clinics and patient transport are needed in some instances.  \n“I feel like the good thing about [the TBSTs] is you don’t need much lab infrastructure, I feel \nlike the need to draw blood is a barrier for IGRAs.” (ID017, HIC, Provider) \n \n3.2.3.2. Acceptability \nPeople-related factors \nThere were 7 subthemes relating to people factors for acceptability of TBI tests (Figure B3). These \nincluded consumer experience, consumer demographics, perceptions of skin tests, knowledge, \nfamiliarity, and provider-consumer communication, which were discussed by bo",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes a provider's view that TBSTs need little laboratory infrastructure. It then presents people-related factors in acceptability, such as consumer experience and perceptions of skin tests, with mixed views on whether new tests would cause anxiety.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1103",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1103",
      "pageStart": 1103,
      "pageEnd": 1103,
      "target": "../sections/page-1103.md",
      "narrative": null,
      "content": "12 \n \n‘live’ reading that took place in front of them, an experience that is avoided with the IGRA. \n“We also found perspective from the patient that maybe people at the beginning, they  \nworry that it’s painful and they scared of injection, but once they (indistinct), they like it very \nmuch because you can see physically the test result on your arm… at the beginning, many \npeople worry about it, they worry that it would be unsafe, severe adverse reactions, and \npeople would die if you injected PPD. Of course it’s not the case, there’s so many evidence in \nthe world (indistinct) but they stil",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents interview quotes on fears about skin test injections and live readings, and on blood draws as a barrier to IGRA, including cultural beliefs about blood. It shows that initial worries about injections often eased with experience. It also contrasts the invasiveness of skin and blood tests.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1104",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1104",
      "pageStart": 1104,
      "pageEnd": 1104,
      "target": "../sections/page-1104.md",
      "narrative": null,
      "content": "13 \n \nLastly, the complicated process involved with skin tests was perceived to reduce acceptability in \nconsumers with low health literacy.  \n“First of all the test is a very complex thing, though to the provider it might be easy but from  \nthe patient’s perspective it is difficult. Like as of my understanding, patient should not have, \nshould not rub on the injected areas and all those certain precautions need to be taken up. So \nto what extent patient keeps those things into mind is one end and this might require two \nvisits to the test clinic. So from the patient perspective, I feel like i",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that the complexity of skin test procedures may reduce acceptability among people with low health literacy. It discusses demographic concerns about using IGRA and TBST in some groups. It also includes providers' hopes for less invasive formats and one view that skin testing should be abandoned altogether.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1105",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1105",
      "pageStart": 1105,
      "pageEnd": 1105,
      "target": "../sections/page-1105.md",
      "narrative": null,
      "content": "14 \n \nsomething that makes much more sense… This whole conversation is surreal to me, why \npeople are doing this with a 100 year old technology… If it’s better than the TB skin test then \nthe issue is replacing that test with this test, that’s, I totally understand that and that’s a \nperfectly reasonable thing. We are trying to get away from skin testing period.” (ID014, HIC, \nProvider) \nKnowledge \nMost participants expressed having limited knowledge about the TBST. One participant reported \nthat being able to see the evidence around the TBST would help to improve her understanding and \npercep",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues with participants' views that skin tests are outdated technology and that most knew little about TBST, calling for public awareness and evidence. It reports that clear explanation and endorsement by providers and WHO were expected to improve acceptability.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p1106-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 1106,
      "pageEnd": 1106,
      "target": "../images/img-p1106-1.webp",
      "narrative": "draft",
      "content": "Diagram of test-related factors in test acceptability (Figure B4) shown as five labelled circles: Accuracy, Reliability, Time to receive result, Suitability in specific subgroups and Ability to differentiate latent and active TB.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1106",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1106",
      "pageStart": 1106,
      "pageEnd": 1106,
      "target": "../sections/page-1106.md",
      "narrative": null,
      "content": "15 \n \nthat it’s a good one, it’s worth carrying on with.” (ID001, HIC, Consumer) \nUsefulness \nA couple of consumers questioned how useful the TBST would be. Interestingly, this did not come up \nas a finding from any healthcare provider participant. One consumer reported needing evidence that \nthe TBST are better than the TST and IGRA to think that they are useful. Another consumer felt that \nhow useful the TBI tests are in practice depends on the care pathway that follows a positive result. \n“If it’s not broken, why are we trying to fix it? What’s the evidence that’s saying what we  \nwere doin",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that some consumers doubted the usefulness of TBST without evidence that it beats existing tests, or questioned what care follows a positive result. Providers expected the new skin tests to match IGRA specificity in skin-test form and would prefer them over the TST.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1107",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1107",
      "pageStart": 1107,
      "pageEnd": 1107,
      "target": "../sections/page-1107.md",
      "narrative": null,
      "content": "16 \n \nspecificity really.” (ID019, HIC, Provider) \n \n“The measurement involves a ballpoint pen with free flowing ink and you sort of make the \nmark on the person’s arm until you start to feel a little bit of resistance and then you stop \nand then you do that on the other side and you stop, and then you measure the distance \nbetween those two stops and that’s probably the most accurate way to interpret the \ninduration. But as you can imagine, I mean thank about doing that and think about the issue, \nhow hard that would be and the sort of saying well if it’s 9mms they’re infected, and if it’s \n1",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents provider criticism of how subjective TST readings are and lists perceived benefits of TBST's greater accuracy, including use in people who inject drugs. It also notes that none of the tests can distinguish latent from active TB.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p1108-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 1108,
      "pageEnd": 1108,
      "target": "../images/img-p1108-1.webp",
      "narrative": "draft",
      "content": "Diagram of contextual factors in test acceptability (Figure B5) shown as three labelled circles: Test environment, Provider perceptions of the test and Infrastructure for test follow-up.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1108",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1108",
      "pageStart": 1108,
      "pageEnd": 1108,
      "target": "../sections/page-1108.md",
      "narrative": null,
      "content": "17 \n \ndeveloping a test to achieve that aim: \n“No the challenges were that it was not really possible to differentiate by the results of the  \nPPD test who is really infected and who really needs preventive treatment… currently all \nresearchers in the field all over the world are looking for some biomarker that could help \ndifferentiate latent TB infection from the active TB infection right, yes, but it is very difficult \nto develop such a biomarker, such a test, because TB infection is very evasive, it can be latent \nand then it can be active.” (ID015, HIC, Provider) \nContextual Factors \nThre",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports views on the difficulty of developing a biomarker that separates latent from active TB. It describes contextual factors in acceptability, such as provider perceptions and follow-up infrastructure, including consequences like deportation after positive tests in travellers. It notes that hospital settings caused anxiety for some consumers.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p1109-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 1109,
      "pageEnd": 1109,
      "target": "../images/img-p1109-1.webp",
      "narrative": "draft",
      "content": "Diagram of people factors in test feasibility (Figure B6): a central grey circle 'Consumer experience' surrounded by four boxed labels: Transport, Healthcare setting, Return visits and Timeliness of results.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1109",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1109",
      "pageStart": 1109,
      "pageEnd": 1109,
      "target": "../sections/page-1109.md",
      "narrative": null,
      "content": "18 \n \nan earlier section (see Acceptability > Consumer experience). This issue could be applicable to all 3 \nTBI tests, although a hospital appointment is most often involved when receiving the IGRA compared \nto the skin tests.  \n3.2.3.3 Feasibility \nPeople Factors \nOnly one subtheme described a people factor that impacted feasibility, which was consumer \nexperience. This subtheme was discussed by both healthcare providers and consumers (Figure B6). \nWhilst provider communication and infrastructure for test follow-up are also potential factors that \nmay be considered as relevant for feasibilit",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Opens the interview findings on feasibility, noting that consumer experience was the only people factor raised. It reports that IGRA's single-visit requirement was seen as a clear advantage that lets providers focus follow-up on positive cases.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p1110-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 1110,
      "pageEnd": 1110,
      "target": "../images/img-p1110-1.webp",
      "narrative": "draft",
      "content": "Diagram of test factors in test feasibility (Figure B7) shown as four labelled circles: Cost, Time to receive result, Portability and Ease of administration.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1110",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1110",
      "pageStart": 1110,
      "pageEnd": 1110,
      "target": "../sections/page-1110.md",
      "narrative": null,
      "content": "19 \n \n“The good thing about the T-SPOT or the IGRA is that they, once they’ve had the blood test  \ndone that’s it and actually if they’re not sort of then we can put effort knowing that it’s \npositive or negative. If it’s positive then obviously we will be following up and putting more \neffort into trying to find somebody whereas often with a Mantoux, if they don’t come back \nwithin the timeframe then it has to be repeated.” (ID012, LMIC, Provider) \nTransportation was also a part of the consumer experience for receiving the IGRA, with consumers \nneeding to visit a clinic or hospital to receive",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents quotes on the single-visit benefit of IGRA versus repeat TST visits and on transport burdens, especially in lower-income countries. It introduces four test factors affecting feasibility: cost, time to result, portability and ease of administration, starting with the lower cost of skin tests.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1111",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1111",
      "pageStart": 1111,
      "pageEnd": 1111,
      "target": "../sections/page-1111.md",
      "narrative": null,
      "content": "20 \n \ncosts in their current practice with TBST, while others acknowledged that the low cost may increase \nthe willingness of governments to fund the TBST. A couple of providers discussed how the low cost \naffected their usage intentions towards the TBST. Quotes from providers discussing these aspects \nare presented below:  \n“I think in my current practice, since we are now dealing with TB patients as well, it will be  \nan added advantage. I’m saving in costs because we are looking forward to do testing of TB \nlike in children whereby I know it’s very difficult for them to submit a sputum and",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports providers' views that the lower cost of TBST could save money and make government funding more likely, including for testing children. It discusses time to result and the portability of skin tests for community outreach. It notes that the TST was considered fairly easy to administer with trained staff.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p1112-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 1112,
      "pageEnd": 1112,
      "target": "../images/img-p1112-1.webp",
      "narrative": "draft",
      "content": "Diagram of contextual factors in test feasibility (Figure B8) shown as three labelled circles: Resource requirements, COVID-19 and Health system setting.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1112",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1112",
      "pageStart": 1112,
      "pageEnd": 1112,
      "target": "../sections/page-1112.md",
      "narrative": null,
      "content": "21 \n \ntricky at first, but ease of use came with practice and appropriate training: \n“I mean, they’re fairly easy to use. I mean they’re a bit fiddly but they’re not, once you get  \nused to it it’s just getting the technique… But it’s yeah, it’s not difficult, once you get the \nhang of it.” (ID012, HIC, Provider) \n \n“We found that we have to retrain our staff at least every six months of the year… with the \nresult sometimes it’s incorrect so every, for example in the community they accepting every \ntime they go to a village, I have to spend the first two or three times intensively to check \nev",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents quotes on the TST technique being fiddly at first but manageable with training, and on the need for frequent retraining in community settings. It also includes a wish for simpler tests such as fingerprick assays. It then introduces contextual factors in feasibility: resource needs, the COVID-19 pandemic and the health system setting.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1113",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1113",
      "pageStart": 1113,
      "pageEnd": 1113,
      "target": "../sections/page-1113.md",
      "narrative": null,
      "content": "22 \n \nFigure describing the contextual factors influencing test feasibility \nResource Requirements \nThe TBST, TST, and IGRA all require access to specific test equipment and trained staff – as such, the \navailability of a trained workforce was considered key for test feasibility. However, the IGRA also \nrequires access to a laboratory for test analysis and a hospital or clinic for the blood draw. As a \nresult, there is also a mandatory transport requirement for the IGRA on the part of the consumer to \nthe hospital or clinic. Patient transport may be a requirement of the TBST and TST in some se",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes contextual factors affecting the feasibility of tests for TB infection, noting that skin tests (TBST and TST) and IGRA all need trained staff and equipment, while IGRA also needs a laboratory, a blood draw and travel to a facility. It introduces equity factors, with tests needing less infrastructure seen as more accessible in resource-limited settings.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p1114-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 1114,
      "pageEnd": 1114,
      "target": "../images/img-p1114-1.webp",
      "narrative": "draft",
      "content": "Diagram of equity factors (Figure B9) shown as five grey circles: Access and transport issues, Suitability to low-income country setting, Scalability for large scale screening programmes, Affected by the infrastructure required and Suitability to subgroups.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1114",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1114",
      "pageStart": 1114,
      "pageEnd": 1114,
      "target": "../sections/page-1114.md",
      "narrative": null,
      "content": "23 \n \n \nFigure describing the factors influencing test health equity \n \nAccess and Transport Issues \nThe IGRA typically requires a visit to a healthcare facility for test administration, whereas the TBST \nand the TST are more portable with less resource requirements, thus a visit to a healthcare facility \nmight not always be required. This is a distinct advantage of the skin tests in low-resource settings, \nas the tests can be made accessible to consumers living a great distance from a healthcare facility, to \nhelp them avoid any of the challenges associated with attending a hospital appointme",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Discusses how tests for TB infection affect health equity, explaining that skin tests are more portable and may not need a facility visit, which helps people living far from clinics. It notes that IGRA was seen as less suited to low-income settings because of its cost and resource needs.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1115",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1115",
      "pageStart": 1115,
      "pageEnd": 1115,
      "target": "../sections/page-1115.md",
      "narrative": null,
      "content": "24 \n \nrequired for implementation compared to the TST and the TBST.  \n“I guess the issue of making it more affordable, its bringing the potential of a different  \ntechnology in a more affordable way to people who have a high burden of the problem.” \n(ID014, HIC, Provider) \n \n“We’ve got access to phlebotomy and can get the IGRA done and it’s quite easy, it sort of \nfeels to me like well what would be the benefit changing. But I see that actually in areas \nwhere actually maybe there isn’t phlebotomy, you haven’t got that sort of access, then \nhaving something quite simple to do and actually you",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents interview quotes on affordability and the value of simple skin tests where phlebotomy is unavailable, and notes that skin tests may suit groups in whom finding veins is hard. It concludes that equity effects of these tests may differ by population.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1116",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1116",
      "pageStart": 1116,
      "pageEnd": 1116,
      "target": "../sections/page-1116.md",
      "narrative": null,
      "content": "25 \n \n3.2.4 Summary of Work Package B findings from User Interviews \nEnd user values and preferences \nHealthcare providers and consumers were found to value TBI test accuracy, convenience, a positive \npatient experience, low cost and resource requirements, and these factors affected their \npreferences for TBI tests. The novel antigen-based skin tests (TBSTs) were discussed as having \ncomparable accuracy to the IGRA, however with better affordability and lower resource \nrequirements. In comparison to the TSTs, the TBSTs were viewed to have better accuracy and \ncomparable affordability, resource",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarizes the user interview findings: providers and consumers valued accuracy, convenience, a good patient experience, low cost and low resource needs. The new antigen-based skin tests were seen as matching IGRA accuracy at lower cost and resource needs, and as more accurate than the TST at similar cost. Interviewees cautioned that in low-income settings the TBST alone would not fix wider access barriers such as stock, staff training and equipment.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1117",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1117",
      "pageStart": 1117,
      "pageEnd": 1117,
      "target": "../sections/page-1117.md",
      "narrative": null,
      "content": "26 \n \nConclusion \nIn conclusion, TBST were deemed to be more acceptable and equally feasible to the TST, with \ngreater benefits for health equity driven by higher sensitivity and specificity. TBST may be more \npreferable for use in more resource-limited settings compared to the TST. IGRA appears the \npreferred test option in high-income countries with the infrastructure to support their use; it is \nunlikely that widespread use of the IGRA would change in these settings. End-users were found to \nvalue attributes such as test accuracy, convenience, a positive patient experience, cost, and resour",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Concludes that antigen-based skin tests were judged more acceptable than and as feasible as the TST, with equity benefits, and possibly preferable in resource-limited settings, while IGRA remains favoured in high-income countries; these findings shaped the discrete choice experiment.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1118",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1118",
      "pageStart": 1118,
      "pageEnd": 1118,
      "target": "../sections/page-1118.md",
      "narrative": null,
      "content": "27 \n \nWork Package C - Discrete choice experiment (DCE) \nThis activity intends to answer following objective: \nObjective D: To explore end-user values and preferences for TB testing by investigating the trade-\noffs end users are willing to make between test attributes derived from systematic reviews and in-\ndepth user interviews, and determine the order of preference of different test attributes using a \ndiscrete choice experiment. \n3.3.1 Aims \nThe aim of this work package was to conduct a discrete choice experiment (DCE) in order to explore \nend-user (consumer and provider) values and prefere",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces Work Package C, a discrete choice experiment to measure the trade-offs consumers and providers are willing to make between attributes of TB infection tests. It explains how attributes were drawn from the reviews and interviews, covering acceptability, accuracy, feasibility and health equity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1119",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1119",
      "pageStart": 1119,
      "pageEnd": 1119,
      "target": "../sections/page-1119.md",
      "narrative": null,
      "content": "28 \n \nA choice task is where a participant is given a choice between different scenarios based on different \nattribute and level combinations. We constructed a DCE with choice tasks that included two options \nof hypothetical TBI test alternatives and an opt-out alternative of choosing neither, which was \nincluded to estimate preferences more relevant to real-world choices.   \nSeparate DCE choice tasks were created for respondents who were providers of TB services and for \nrespondents who were consumers. The attributes included in each DCE differed slightly, though \nsome were common across both",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains how the choice tasks were built, each offering two hypothetical tests plus an opt-out, with separate versions for providers and consumers, and how a pilot informed the final design hosted online.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1120",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1120",
      "pageStart": 1120,
      "pageEnd": 1120,
      "target": "../sections/page-1120.md",
      "narrative": null,
      "content": "1 \n \nTable C1: List of attributes and levels included in both provider and consumer DCEs  \n \nProvider DCE \nConsumer DCE \nAttribute \nLevels \nDescription  \nLevels \nCost to healthcare practice \nUSD 100 \nTime required for test results to \ncome back \n1 day \n  \nUSD 50 \n  \n3 days \n  \nUSD 25 \n  \n7 days \nIf a follow-up appointment is \nneeded to get test results \nIn-person follow-up appointment needed \nIf a follow-up appointment is \nneeded to get test results \nIn-person follow-up appointment needed to get \nresults  \n  \nNo follow-up appointment required \n  \nNo follow-up appointment required \nChance of th",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Tabulates the attributes and levels used in the provider and consumer choice experiments, such as cost, time to results, need for follow-up visits, false positive and negative rates, testing location and how the test is done.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/table-057",
      "types": [
        "folio-assistant-core:Table",
        "doco:Table"
      ],
      "kind": "table",
      "title": "Table C1: List of attributes and levels included in both provider and consumer DCEs",
      "pageStart": 1120,
      "pageEnd": 1120,
      "target": "../tables/table-c1.csv",
      "narrative": "draft",
      "content": "Attributes and their levels used in the discrete choice experiments (DCEs) for providers and for consumers, side by side; attributes common to both DCEs are shaded.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1121",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1121",
      "pageStart": 1121,
      "pageEnd": 1121,
      "target": "../sections/page-1121.md",
      "narrative": null,
      "content": "0 \n \n3.3.2.3 Recruitment  \nParticipants for the DCE were recruited from the same research networks used to recruit the users \nfor the interviews for Work Package B e.g. from a clinical, research and community setting: \n• \nTB Treatment Action Group \n• \nTB Civil Societies \n• \nMembers of the WHO Guideline Development Group for TBST \n• \nGlobal Coalition of TB Activists (GCTA) \n• \nStop TB Partnership \n• \nWHO TBI Task Force \nPotential participants were provided with an online link, which led them to the study information \nsheet. At the bottom of the information sheet webpage, participants were infor",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes recruitment of choice experiment participants through TB networks and WHO groups, the online consent process and analysis methods. It summarizes respondents, most of whom were based in India, the United States, South Africa, Pakistan and Zimbabwe.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/figure-img-p1122-1",
      "types": [
        "folio-assistant-core:Figure",
        "doco:Figure"
      ],
      "kind": "figure",
      "title": "",
      "pageStart": 1122,
      "pageEnd": 1122,
      "target": "../images/img-p1122-1.webp",
      "narrative": "draft",
      "content": "World map of DCE respondent locations (Figure C1), with countries shaded by number of respondents and labelled with counts. The darkest shading is the United States of America (38); India (26), South Africa (20) and a neighbouring country (17), and Pakistan (18) are also shaded more darkly, while many other countries are labelled with 1 to 10 respondents and grey countries have none.",
      "truncated": false,
      "provenance": "ingested",
      "summary": null
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1122",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1122",
      "pageStart": 1122,
      "pageEnd": 1122,
      "target": "../sections/page-1122.md",
      "narrative": null,
      "content": "1 \n \n \nProvider \n(n=186) \nConsumer (n=48) \nGender \n \n \n    Female \n112 (60%) \n26 (54%) \n    Male \n73 (39%) \n22 (46%) \n    Other \n1 (1%) \n \nAge Category \n \n \n    18-35 \n32 (17%) \n13 (27%) \n    36-55 \n106 (57%) \n24 (50%) \n    56-75 \n44 (24%) \n11 (23%) \n    75+ \n4 (2%) \n0 (0%) \n \nFigure C1: Location of DCE respondents  \n \n3.3.3.2 Choice analysis \nTable 4 shows choice model results, analysing DCE data for providers and consumers using \nmultinomial logistic regression models. Where coefficients are significantly greater than zero, an \nattribute can be interpreted as having a relatively positive imp",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Tabulates characteristics of the 186 provider and 48 consumer respondents and begins the choice analysis, noting that consumers strongly preferred any test over no test.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1123",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1123",
      "pageStart": 1123,
      "pageEnd": 1123,
      "target": "../sections/page-1123.md",
      "narrative": null,
      "content": "2 \n \nAmong providers, after test accuracy the next most influential attribute was not requiring an in-\nperson follow-up appointment (β=0.151, p=0.02), followed by not requiring specialist staff or \nequipment to administer or interpret the test (β=0.144, p=0.01). There was weak evidence that \nproviders preferred more expensive tests, though test cost was the least important determinant of \nchoice (β=0.133, p=0.05). Cost was not an attribute in the consumer DCE. \nTable 5: Multinomial logit choice model results \n \n  \n(1) \nConsumer \n(2) \nProvider \n  \nCoefficient \nSE \nCoefficient \nSE \nNo in person",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that for providers, after accuracy, the most influential attributes were avoiding an in-person follow-up visit and not needing specialist staff or equipment, with cost least important. It presents the model results and notes that providers valued avoiding follow-up far more than consumers did.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1124",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1124",
      "pageStart": 1124,
      "pageEnd": 1124,
      "target": "../sections/page-1124.md",
      "narrative": null,
      "content": "3 \n \nslightly more negatively than the same increase in likelihood of a false positive; the opposite was \ntrue among providers.  \n \nFigure C2: Willingness to accept change in false positivity rate for DCE \nattributes.  \n \n \n \nPanel A shows results for consumers, and panel B results for providers. A positive value indicates \nthat respondents are willing to accept a lower likelihood of a false-positive result for that attribute.  \nPanel A, example interpretation: Consumers would be willing to accept an 8.4 percentage point \ngreater false positivity rate to avoid a seven-day wait for test results",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Shows how much higher false-positive rate consumers and providers would accept in exchange for other test attributes, such as avoiding a seven-day wait for results.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1125",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1125",
      "pageStart": 1125,
      "pageEnd": 1125,
      "target": "../sections/page-1125.md",
      "narrative": null,
      "content": "4 \n \nconsumers’ strongest preferences were for the location of testing, with testing in primary care \nstrongly preferred to hospital locations. For providers, the next most important attribute was the \nlack of requiring an in-person follow-up appointment, followed by not needing specialist staff or \nequipment to interpret or administer the test. Willingness-to-accept analyses indicated that both \nconsumers and providers were willing to bear a substantial reduction in test sensitivity to reduce \nwaiting times for results (8.4 percentage point reduction to avoid a 7-day wait), and to avoid in-\np",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarizes the choice experiment: consumers most preferred testing in primary care rather than hospitals, and both groups would accept lower accuracy to get faster results and avoid follow-up visits. It notes that too few responses were collected to compare preferences across regions.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1126",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1126",
      "pageStart": 1126,
      "pageEnd": 1126,
      "target": "../sections/page-1126.md",
      "narrative": null,
      "content": "5 \n \nIntroduction  \nTuberculosis (TB) is an airborne transmissible infectious disease caused by Mycobacterium \ntuberculosis. TB is a major cause of health loss (i.e., healthy life years lost to death, illness, or disability), \naffecting approximately one quarter of the global population and ranking as one of the top ten causes \nof death1. Approximately one quarter to one-third of the global population is estimated to have latent \nTB infection (TBI) 2, 3. TBI may progress to active disease, especially in those with pre-disposing factors \nsuch as HIV infection, diabetes, and immunosuppressive co",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces the report on user perspectives on TB infection testing, describing TB's global burden and that about a quarter to a third of people are estimated to have TB infection, which can progress to disease. It describes current tests (TST and IGRA) and the newer antigen-based skin tests, which are portable and expected to be low cost but need two clinic visits.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1127",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1127",
      "pageStart": 1127,
      "pageEnd": 1127,
      "target": "../sections/page-1127.md",
      "narrative": null,
      "content": "6 \n \nacceptable or feasible for use in routine practice for end users (e.g., individuals and healthcare \nproviders), and whether use of these tests would impact health equity. \n \nTo address this knowledge gap, World Health Organization (WHO) commissioned systematic reviews \ninto the perspectives, preferences, and experiences of users of TBST and explore health equity. The \nreviews aim to qualitatively explore what the end user perspectives, preferences and experiences are \nof TBST, particularly regarding their acceptability and feasibility, and any impacts on health equity. \nAnticipating a pau",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains that WHO commissioned reviews of user perspectives on the new skin tests because their acceptability, feasibility and equity effects were unknown. It outlines the three work packages, a systematic review, user interviews and a discrete choice experiment, guided by an expert advisory panel.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1128",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1128",
      "pageStart": 1128,
      "pageEnd": 1128,
      "target": "../sections/page-1128.md",
      "narrative": null,
      "content": "7 \n \nWork Package A-Systematic review \nThis activity intends to answer the following objectives: \nObjective A: To conduct a systematic review that summarizes the evidence on end-user \nvalues and preferences regarding the acceptability and feasibility of antigen-based skin \ntests for TB infection, compared to current tests of infection (IGRA, TST). \nObjective C: To evaluate any potential impacts of TBST on health equity \n \nTwo systematic reviews were conducted that synthesized the qualitative research evidence to date \non end-user values and preferences for the use of TBST for latent TB infecti",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces Work Package A, two systematic reviews of qualitative evidence on end users' views of antigen-based skin tests compared with IGRA and TST, and on their equity effects. It describes the database searches, including Russian and Chinese databases and a WHO public call for data.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1129",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1129",
      "pageStart": 1129,
      "pageEnd": 1129,
      "target": "../sections/page-1129.md",
      "narrative": null,
      "content": "8 \n \nantigens). \nGrey literature \nGrey literature was not searched as part of the review.   \n \n3.1.2 Eligibility criteria \nStudy design \nIncluded studies were primary research studies that used qualitative methodologies or study \ndesigns. Studies were required to use qualitative methods for data collection (e.g., interviews, focus \ngroups, open-ended survey questions) and analysis (e.g., thematic analysis, grounded theory). Mixed \nmethods studies that used qualitative methods for data collection and analysis were included. \nSurvey studies that presented qualitative free text that was collected",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Sets out the review eligibility criteria, which required qualitative data collection and analysis, and describes screening in duplicate, including by native speakers for Russian and Chinese database results.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1130",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1130",
      "pageStart": 1130,
      "pageEnd": 1130,
      "target": "../sections/page-1130.md",
      "narrative": null,
      "content": "9 \n \nQualitative data pertaining to test acceptability, feasibility, and health equity were extracted from \nthe included papers independently in duplicate by two raters (AC, KL) using Covidence software. \nData were also extracted on study characteristics, including publication language, country, type of \nfacility (including public/private, rural/urban), study objectives, methods of data collection and \nanalysis, study duration, conceptual framework, conflicts of interest, study participant type and \nmean age, percentage male, ethnicity, diagnostic tests involved, programmatic features of the",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes data extraction in duplicate using Covidence, the study characteristics recorded, and the use of GRADE-CERQual to rate confidence in each review finding.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1131",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1131",
      "pageStart": 1131,
      "pageEnd": 1131,
      "target": "../sections/page-1131.md",
      "narrative": null,
      "content": "10 \n \n3.1.4 Results from Systematic Review A and B \n3.1.4.1 Results of the Search \nFour studies were identified that met the inclusion criteria (see Figure A1 for Preferred Reporting \nItems for Systematic Reviews and Meta-Analyses (PRISMA) flow charts). Review A included papers \nfocused on specific TBST compared to the IGRA and/or the TST, while Review B focused on papers on \nthe IGRA and TST compared to each other or without a comparator. From Review A for TBSTs, only \none study was identified18, which came from a WHO public call for data relating to the feasibility and \nacceptability of anti",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that four studies met the inclusion criteria: one on a skin test from WHO's call for data and three on IGRA and TST, published between 2014 and 2021, with PRISMA flow charts.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1132",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1132",
      "pageStart": 1132,
      "pageEnd": 1132,
      "target": "../sections/page-1132.md",
      "narrative": null,
      "content": "11 \n \n3.1.4.2 Description of the Studies \nA table summarizing the Characteristics of the included studies identified from Review A and B are \npresented in Table A1.  \nTest \nFrom Review A, only one study reported on a TBST which was the Diaskintest18, which was from the \nWHO public call for data. The Diaskintest was evaluated as part of a large-scale screening program \nfor children and adolescents in the Penza region of the Russian Federation. From Review B,  there \nwere three published studies reported on the IGRA and the TST19,20,21. Two studies focused on the \nuse of the IGRA and TST as part",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes the included studies, including the only skin test study on Diaskintest in a mass screening programme for children in the Russian Federation, and three studies on IGRA and TST. It begins describing the confidence ratings for the review findings.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1133",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1133",
      "pageStart": 1133,
      "pageEnd": 1133,
      "target": "../sections/page-1133.md",
      "narrative": null,
      "content": "12 \n \nhealth equity), 3 findings (end user values and preferences, acceptability, feasibility) were rated with \nmoderate confidence based on the findings from Review A and B. Health equity received a low \nconfidence rating. Appendix B provides a detailed summary table of the GRADE-CERQual analysis. \nThere were serious concerns about data adequacy for the feasibility and health equity findings due \nto the small number of studies supporting each finding and the low data richness (i.e., small amount \nof relevant data) in 3 of 4 studies (1 from Review A18, and 2 from Review B20, 21). Moreover, the",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that findings on values, acceptability and feasibility were rated moderate confidence and health equity low, owing to few studies, thin data and focus on specific subgroups.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1134",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1134",
      "pageStart": 1134,
      "pageEnd": 1134,
      "target": "../sections/page-1134.md",
      "narrative": null,
      "content": "0 \n \nTable A1: Systematic Reviews A and B- Characteristics of Included Studies \nStudy ID \nPublication \ntype \nTBI Tests \nCountry \nFacility Type \nPublic/ \nPrivate \nRural/ \nUrban \nParticipants \nProgrammatic features \nComparato\nrs \nData \ncollection \nStudy \nduration \nMethod of data \nanalysis \nConceptual \nframework \nConflicts of Interest \nReview A findings on TBST \nAksenova \n2021 \n(Review A) \nReport \n(WHO call \nfor data) \n \nDiaskintest \nRussian \nFederation \nNot reported \nNot \nreported \nNot \nreported \nParents of \nchildren and \nadolescents \nundergoing TB \nscreening. \nDemographic \ndata not \nreported.",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Tabulates the characteristics of the four included studies, including test, country, setting, participants, methods, conceptual framework and conflicts of interest.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1135",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1135",
      "pageStart": 1135,
      "pageEnd": 1135,
      "target": "../sections/page-1135.md",
      "narrative": null,
      "content": "0 \n \n3.1.4.5 Review Findings \nOur qualitative evidence synthesis from Work Package A draws on the findings from both Systematic \nReviews A and B. Overall, there were limited data available hence the decision to combine findings \nfrom both Reviews, with Review A on TBST retrieving only one relevant data report, and Review B on \nTST and IGRA retrieving only three papers.  \nThe findings identified factors contributing to end-user values/preferences, acceptability, feasibility, \nand health equity impacts of TBST, IGRA and TST, corresponding to our four objectives. We present \neach finding under ea",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents review findings on end-user values and acceptability, combining both reviews because data were scarce. It notes the perceived accuracy of the tests and that clinic staff had concerns about IGRA accuracy in adults with HIV.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1136",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1136",
      "pageStart": 1136,
      "pageEnd": 1136,
      "target": "../sections/page-1136.md",
      "narrative": null,
      "content": "1 \n \nthat there was a high frequency of indeterminate results from the IGRA, although this was not \nsupported in the clinic data where only 1% of indeterminate results were received20.  \n“Sometimes we’re not sure if they measure it correctly, or if the TST was done correctly, did \nthey take enough, was it inserted sub-dermally? There can come human errors as well. With \nblood, I think the result is the result.” (South Africa) \nConvenience was found to affect test preferences in all 4 studies18,19,20,21. Although limited data was \nprovided, the Diaskintest was reported to be convenient for use",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues the findings on acceptability, noting that perceived IGRA indeterminate results were not borne out by clinic data, and that convenience shaped test preferences in all four studies. Quotes illustrate views on human error in skin tests and discomfort from the TST.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1137",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1137",
      "pageStart": 1137,
      "pageEnd": 1137,
      "target": "../sections/page-1137.md",
      "narrative": null,
      "content": "2 \n \nTest Factors \nAcceptability was affected by test-related factors, including the accuracy of the test, the certainty of \nresults, and the suitability of the test for use in particular subgroups (e.g., children and adolescents, \npersons with HIV, needle-phobic individuals). \nTBST and IGRA were perceived as having greater specificity than the TST18,21. The TST was described \nas having a high rate of false positives; thus the IGRA was deemed preferable in terms of \nspecificity21. From Review A, the high specificity of the TBST was found to improve the motivation of \nparents to perform TB test",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes test-related factors in acceptability, with skin tests and IGRA seen as more specific than the TST, which had many false positives. It notes that the skin test's specificity encouraged parents to have children tested, and that no study assessed TBST feasibility, with moderate confidence in the evidence.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1138",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1138",
      "pageStart": 1138,
      "pageEnd": 1138,
      "target": "../sections/page-1138.md",
      "narrative": null,
      "content": "3 \n \nAspects of the individual experience were found to affect the feasibility of the IGRA and TST. A \ncommon feasibility concern for the TST, based on findings from Review B, were that consumers were \nrequired to visit the clinic a second time within 72 hours to read the test result19,20,21. Kerrigan et al. \ndescribed how in the North West Province of South Africa, many consumers experienced difficulties \nwith visiting a clinic due to the long journey and wait time involved, as well as issues with accessing \ntransportation19. Similar concerns were echoed in HIV clinics in the United States, a",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes individual factors in feasibility, especially the TST's need for a return visit within 72 hours, which was hard for people facing long journeys, waits and transport problems in South Africa and the United States. It reports that using routine blood draws for IGRA was seen as easing logistics.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1139",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1139",
      "pageStart": 1139,
      "pageEnd": 1139,
      "target": "../sections/page-1139.md",
      "narrative": null,
      "content": "4 \n \nthe low frequency with which they would typically have a blood draw: \n“While there was general consensus on this issue, one patient did relay a concern about \ntiming, as many patients may only have a CD4 blood draw 1–2 times a year. So while this \npatient thought screening the CD4 blood draw for TB was a good idea, he wouldn’t want that \nto delay finding out he had TB, as he would want to know that information as soon as \npossible. This issue was also raised by clinical providers, and a few in turn mentioned that \nscreening the blood for TB could potentially complement, as opposed to repl",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports concern that linking TB testing to infrequent CD4 blood draws might delay diagnosis, and providers' view that it could complement rather than replace other testing. It summarizes contextual feasibility factors such as transport, waiting times, scalability and suitability for subgroups.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1140",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1140",
      "pageStart": 1140,
      "pageEnd": 1140,
      "target": "../sections/page-1140.md",
      "narrative": null,
      "content": "5 \n \nIn terms of access and transport issues, the return visit required to read the TST result was focused \non as a major accessibility issue in 3 studies19,20,21. This included at-risk groups in both middle- and \nhigh-income countries (South Africa, United States). Kerrigan et al., which focused on an HIV clinic in \na middle-income country, documented how many consumers experienced challenges with \ntransportation and wait times in understaffed clinics, serving as a barrier for return visits19. \nChallenges with an adequate number of specialized staff to administer the TST and IGRA were \ndocume",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes access and transport as the main equity issue, with the TST return visit a barrier in several studies, alongside shortages of trained staff. It notes reports that Diaskintest worked well for mass screening of children.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1141",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1141",
      "pageStart": 1141,
      "pageEnd": 1141,
      "target": "../sections/page-1141.md",
      "narrative": null,
      "content": "6 \n \n3.1.5 Summary of Work Package A findings from the Systematic \nReviews \nOverall, there were limited data from the reviews on TBSTs, with only 4 studies identified. \nImportantly, these studies were conducted in very specific healthcare settings and populations, \nlimiting our confidence in the data quality and the generalisability of these findings. Most of the data \navailable related to TST and IGRA. Despite the limited data available, we identified key factors \nrelating to our four objectives of end user values/preferences, acceptability, feasibility and health \nequity. Important end-user",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarizes Work Package A: only four studies were found, mostly on TST and IGRA in specific settings, limiting confidence, but key factors for values, acceptability, feasibility and equity were identified and used to shape the interviews and choice experiment.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1142",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1142",
      "pageStart": 1142,
      "pageEnd": 1142,
      "target": "../sections/page-1142.md",
      "narrative": null,
      "content": "7 \n \nWork Package B-Qualitative user interviews \nThis activity intends to answer following objectives: \nA qualitative interview study was conducted with 20 participants who were individuals living with TB \nor providers involved in TB testing and /or care of individuals living with TB. Participants were \nrecruited from the research team’s global networks across high and low-middle income countries, \nwith support from WHO. Participants were approached through organizations/groups as below: \n• \nTB Treatment Action Group \n• \nTB Civil Societies \n• \nMembers of the WHO Guideline Development Group for",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces Work Package B, a qualitative interview study with 20 people living with TB or providers involved in TB testing and care, recruited through global networks with WHO support. It lists the recruiting organizations and the objectives on views of skin tests and equity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1143",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1143",
      "pageStart": 1143,
      "pageEnd": 1143,
      "target": "../sections/page-1143.md",
      "narrative": null,
      "content": "8 \n \n1. Current approach to using skin tests as part of the TB diagnostic process; \n2. End-user values and preferences with regard to TB testing, from their previous experience \nwith skin-based tests – focusing on ease of use and process for TBST, IGRA, and TST; \n3. End-user values and preferences considering the challenges and benefits of skin-based \ntests, particularly TBST, which test(s) they would prefer and reasons for the preference; \n4. Feasibility of implementing TBST in current practice; \n5. Overall acceptability of TBST- burden, attitude, understanding, test efficacy; \n6. Potential i",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Lists the topics covered in the interviews, from current use of skin tests to user preferences, feasibility, acceptability and equity, and summarizes participants' roles and countries.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1144",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1144",
      "pageStart": 1144,
      "pageEnd": 1144,
      "target": "../sections/page-1144.md",
      "narrative": null,
      "content": "9 \n \n \n1.2.4 Key Findings  \nThe findings from Work Package A were used to inform the coding framework and analyses \nundertaken in Work Package B – as such, the same subthemes of people, test, and contextual factors \nwere used to explore acceptability and feasibility of the tests, with health equity and end-user values \nand preferences as separate subthemes. The full coding framework used can be found in Appendix \nE. As before, the same shading scheme will be used in the figures to illustrate themes and \nsubthemes for acceptability and feasibility: Dark solid grey refers to people factors; squa",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains that the interview analysis used the same people, test and context subthemes as the reviews, and begins reporting findings on end-user values, such as the hope that more accurate tests would reduce unnecessary treatment and stress.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1145",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1145",
      "pageStart": 1145,
      "pageEnd": 1145,
      "target": "../sections/page-1145.md",
      "narrative": null,
      "content": "10 \n \nperceived to have lower accuracy and were associated with downstream challenges by end-users. \nThese included greater expense and psychological stress.  \n“The big issue I think is the false positives with the current tests that we use and that then \nleads to a whole chain of events which I think is very expensive and stressful for people \nbecause if you’re told you might have TB or you probably have got TB and then you’ve got a \nwhole set of tests, if the novel test is more accurate and it comes back and it says no you \ndidn’t, that saves a lot of stress and all the hospital resources an",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that interviewees saw the TST as less accurate, leading to costly and stressful follow-up after false positives. It also describes the resources each test requires, with IGRA needing laboratories, phlebotomy and trained staff.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1146",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1146",
      "pageStart": 1146,
      "pageEnd": 1146,
      "target": "../sections/page-1146.md",
      "narrative": null,
      "content": "11 \n \ntest equipment are still required, and clinics and patient transport are needed in some instances.  \n“I feel like the good thing about [the TBSTs] is you don’t need much lab infrastructure, I feel \nlike the need to draw blood is a barrier for IGRAs.” (ID017, HIC, Provider) \n \n3.2.3.2. Acceptability \nPeople-related factors \nThere were 7 subthemes relating to people factors for acceptability of TBI tests (Figure B3). These \nincluded consumer experience, consumer demographics, perceptions of skin tests, knowledge, \nfamiliarity, and provider-consumer communication, which were discussed by bo",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Notes that skin tests need less infrastructure than IGRA, then sets out seven people-related factors affecting acceptability, such as consumer experience, perceptions, knowledge and provider communication. It describes anxiety linked to new tests, injections and uncertain skin reactions.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1147",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1147",
      "pageStart": 1147,
      "pageEnd": 1147,
      "target": "../sections/page-1147.md",
      "narrative": null,
      "content": "12 \n \n‘live’ reading that took place in front of them, an experience that is avoided with the IGRA. \n“We also found perspective from the patient that maybe people at the beginning, they  \nworry that it’s painful and they scared of injection, but once they (indistinct), they like it very \nmuch because you can see physically the test result on your arm… at the beginning, many \npeople worry about it, they worry that it would be unsafe, severe adverse reactions, and \npeople would die if you injected PPD. Of course it’s not the case, there’s so many evidence in \nthe world (indistinct) but they stil",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents interview findings on the patient experience of tuberculosis infection tests, with quotes on initial fear of injections and the reassurance of seeing the result on one's own arm. A provider describes how long it took to convince people that skin testing was safe. Other participants compare skin tests with blood draws, one noting local myths about taking blood.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1148",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1148",
      "pageStart": 1148,
      "pageEnd": 1148,
      "target": "../sections/page-1148.md",
      "narrative": null,
      "content": "13 \n \nLastly, the complicated process involved with skin tests was perceived to reduce acceptability in \nconsumers with low health literacy.  \n“First of all the test is a very complex thing, though to the provider it might be easy but from  \nthe patient’s perspective it is difficult. Like as of my understanding, patient should not have, \nshould not rub on the injected areas and all those certain precautions need to be taken up. So \nto what extent patient keeps those things into mind is one end and this might require two \nvisits to the test clinic. So from the patient perspective, I feel like i",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that the complex steps of skin testing, such as avoiding rubbing the site and returning for a reading, may reduce acceptability for people with low health literacy. It notes concerns about using blood tests and skin tests in people with darker skin, where veins and reactions can be harder to see. Some providers questioned continuing with skin tests at all.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1149",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1149",
      "pageStart": 1149,
      "pageEnd": 1149,
      "target": "../sections/page-1149.md",
      "narrative": null,
      "content": "14 \n \nsomething that makes much more sense… This whole conversation is surreal to me, why \npeople are doing this with a 100 year old technology… If it’s better than the TB skin test then \nthe issue is replacing that test with this test, that’s, I totally understand that and that’s a \nperfectly reasonable thing. We are trying to get away from skin testing period.” (ID014, HIC, \nProvider) \nKnowledge \nMost participants expressed having limited knowledge about the TBST. One participant reported \nthat being able to see the evidence around the TBST would help to improve her understanding and \npercep",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Continues with a provider's view that skin testing is outdated technology, then reports that most participants knew little about the novel skin tests and wanted evidence and public awareness. Familiarity with existing skin tests and endorsement by bodies such as WHO were expected to improve acceptance.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1150",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1150",
      "pageStart": 1150,
      "pageEnd": 1150,
      "target": "../sections/page-1150.md",
      "narrative": null,
      "content": "15 \n \nthat it’s a good one, it’s worth carrying on with.” (ID001, HIC, Consumer) \nUsefulness \nA couple of consumers questioned how useful the TBST would be. Interestingly, this did not come up \nas a finding from any healthcare provider participant. One consumer reported needing evidence that \nthe TBST are better than the TST and IGRA to think that they are useful. Another consumer felt that \nhow useful the TBI tests are in practice depends on the care pathway that follows a positive result. \n“If it’s not broken, why are we trying to fix it? What’s the evidence that’s saying what we  \nwere doin",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that some consumers doubted the usefulness of the novel skin tests without proof they beat existing tests or without follow-up care after a positive result. It introduces test factors affecting acceptability, with providers favouring novel skin tests over the tuberculin skin test for better specificity.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1151",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1151",
      "pageStart": 1151,
      "pageEnd": 1151,
      "target": "../sections/page-1151.md",
      "narrative": null,
      "content": "16 \n \nspecificity really.” (ID019, HIC, Provider) \n \n“The measurement involves a ballpoint pen with free flowing ink and you sort of make the \nmark on the person’s arm until you start to feel a little bit of resistance and then you stop \nand then you do that on the other side and you stop, and then you measure the distance \nbetween those two stops and that’s probably the most accurate way to interpret the \ninduration. But as you can imagine, I mean thank about doing that and think about the issue, \nhow hard that would be and the sort of saying well if it’s 9mms they’re infected, and if it’s \n1",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Includes a provider's criticism of how hard it is to measure induration precisely, then lists perceived downstream benefits of more accurate novel tests, such as fewer false positives, better diagnosis and avoiding unnecessary treatment. It notes that none of the three test types can distinguish latent from active tuberculosis.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1152",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1152",
      "pageStart": 1152,
      "pageEnd": 1152,
      "target": "../sections/page-1152.md",
      "narrative": null,
      "content": "17 \n \ndeveloping a test to achieve that aim: \n“No the challenges were that it was not really possible to differentiate by the results of the  \nPPD test who is really infected and who really needs preventive treatment… currently all \nresearchers in the field all over the world are looking for some biomarker that could help \ndifferentiate latent TB infection from the active TB infection right, yes, but it is very difficult \nto develop such a biomarker, such a test, because TB infection is very evasive, it can be latent \nand then it can be active.” (ID015, HIC, Provider) \nContextual Factors \nThre",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Quotes a provider on the difficulty of finding a biomarker to separate latent and active tuberculosis, then describes contextual factors in acceptability: provider perceptions, follow-up infrastructure and the test environment. Only consumers raised the test environment as a factor.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1153",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1153",
      "pageStart": 1153,
      "pageEnd": 1153,
      "target": "../sections/page-1153.md",
      "narrative": null,
      "content": "18 \n \nan earlier section (see Acceptability > Consumer experience). This issue could be applicable to all 3 \nTBI tests, although a hospital appointment is most often involved when receiving the IGRA compared \nto the skin tests.  \n3.2.3.3 Feasibility \nPeople Factors \nOnly one subtheme described a people factor that impacted feasibility, which was consumer \nexperience. This subtheme was discussed by both healthcare providers and consumers (Figure B6). \nWhilst provider communication and infrastructure for test follow-up are also potential factors that \nmay be considered as relevant for feasibilit",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Begins the feasibility findings, noting that consumer experience was the only people-related factor discussed and that IGRA's single visit was seen by some providers as an advantage over skin tests needing a return visit.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1154",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1154",
      "pageStart": 1154,
      "pageEnd": 1154,
      "target": "../sections/page-1154.md",
      "narrative": null,
      "content": "19 \n \n“The good thing about the T-SPOT or the IGRA is that they, once they’ve had the blood test  \ndone that’s it and actually if they’re not sort of then we can put effort knowing that it’s \npositive or negative. If it’s positive then obviously we will be following up and putting more \neffort into trying to find somebody whereas often with a Mantoux, if they don’t come back \nwithin the timeframe then it has to be repeated.” (ID012, LMIC, Provider) \nTransportation was also a part of the consumer experience for receiving the IGRA, with consumers \nneeding to visit a clinic or hospital to receive",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Quotes providers on the advantage of IGRA needing no return visit and describes the transport and parking burden of attending hospital for a blood test, likely greater in lower-income countries. It then notes that skin tests were seen as cheaper than IGRA.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1155",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1155",
      "pageStart": 1155,
      "pageEnd": 1155,
      "target": "../sections/page-1155.md",
      "narrative": null,
      "content": "20 \n \ncosts in their current practice with TBST, while others acknowledged that the low cost may increase \nthe willingness of governments to fund the TBST. A couple of providers discussed how the low cost \naffected their usage intentions towards the TBST. Quotes from providers discussing these aspects \nare presented below:  \n“I think in my current practice, since we are now dealing with TB patients as well, it will be  \nan added advantage. I’m saving in costs because we are looking forward to do testing of TB \nlike in children whereby I know it’s very difficult for them to submit a sputum and",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Reports that providers saw cost savings from novel skin tests and thought low cost might make governments more willing to fund them, especially where sputum collection from children is difficult. It then discusses ease of administration, with the tuberculin skin test seen as manageable with trained staff.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1156",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1156",
      "pageStart": 1156,
      "pageEnd": 1156,
      "target": "../sections/page-1156.md",
      "narrative": null,
      "content": "21 \n \ntricky at first, but ease of use came with practice and appropriate training: \n“I mean, they’re fairly easy to use. I mean they’re a bit fiddly but they’re not, once you get  \nused to it it’s just getting the technique… But it’s yeah, it’s not difficult, once you get the \nhang of it.” (ID012, HIC, Provider) \n \n“We found that we have to retrain our staff at least every six months of the year… with the \nresult sometimes it’s incorrect so every, for example in the community they accepting every \ntime they go to a village, I have to spend the first two or three times intensively to check \nev",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents provider views that the tuberculin skin test is fiddly at first but easy with practice, though staff may need frequent retraining. It notes judgements on the ease of novel skin tests and introduces contextual feasibility factors: resource needs, the COVID-19 pandemic and the health system setting.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1157",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1157",
      "pageStart": 1157,
      "pageEnd": 1157,
      "target": "../sections/page-1157.md",
      "narrative": null,
      "content": "22 \n \nFigure describing the contextual factors influencing test feasibility \nResource Requirements \nThe TBST, TST, and IGRA all require access to specific test equipment and trained staff – as such, the \navailability of a trained workforce was considered key for test feasibility. However, the IGRA also \nrequires access to a laboratory for test analysis and a hospital or clinic for the blood draw. As a \nresult, there is also a mandatory transport requirement for the IGRA on the part of the consumer to \nthe hospital or clinic. Patient transport may be a requirement of the TBST and TST in some se",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains that all three test types need equipment and trained staff, but IGRA also needs a laboratory, a blood draw and patient travel, making it less feasible in resource-limited settings. It introduces equity factors, noting tests needing less infrastructure were seen as more accessible.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1158",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1158",
      "pageStart": 1158,
      "pageEnd": 1158,
      "target": "../sections/page-1158.md",
      "narrative": null,
      "content": "23 \n \n \nFigure describing the factors influencing test health equity \n \nAccess and Transport Issues \nThe IGRA typically requires a visit to a healthcare facility for test administration, whereas the TBST \nand the TST are more portable with less resource requirements, thus a visit to a healthcare facility \nmight not always be required. This is a distinct advantage of the skin tests in low-resource settings, \nas the tests can be made accessible to consumers living a great distance from a healthcare facility, to \nhelp them avoid any of the challenges associated with attending a hospital appointme",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes how portable skin tests can reach people far from health facilities, avoiding travel, crowded clinics and long waits, an advantage in low-resource settings. It notes IGRA was seen as less suited to low-income countries because of cost and resource needs.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1159",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1159",
      "pageStart": 1159,
      "pageEnd": 1159,
      "target": "../sections/page-1159.md",
      "narrative": null,
      "content": "24 \n \nrequired for implementation compared to the TST and the TBST.  \n“I guess the issue of making it more affordable, its bringing the potential of a different  \ntechnology in a more affordable way to people who have a high burden of the problem.” \n(ID014, HIC, Provider) \n \n“We’ve got access to phlebotomy and can get the IGRA done and it’s quite easy, it sort of \nfeels to me like well what would be the benefit changing. But I see that actually in areas \nwhere actually maybe there isn’t phlebotomy, you haven’t got that sort of access, then \nhaving something quite simple to do and actually you",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents quotes on making testing more affordable where tuberculosis burden is high and phlebotomy is unavailable, and on the scalability of skin tests for large screening in hard-to-reach communities. It concludes that equity impacts may vary by population group.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1160",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1160",
      "pageStart": 1160,
      "pageEnd": 1160,
      "target": "../sections/page-1160.md",
      "narrative": null,
      "content": "25 \n \n3.2.4 Summary of Work Package B findings from User Interviews \nEnd user values and preferences \nHealthcare providers and consumers were found to value TBI test accuracy, convenience, a positive \npatient experience, low cost and resource requirements, and these factors affected their \npreferences for TBI tests. The novel antigen-based skin tests (TBSTs) were discussed as having \ncomparable accuracy to the IGRA, however with better affordability and lower resource \nrequirements. In comparison to the TSTs, the TBSTs were viewed to have better accuracy and \ncomparable affordability, resource",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Summarises interview findings that providers and consumers value accuracy, convenience, good patient experience, low cost and low resource needs. Novel skin tests were seen as similar in accuracy to IGRA but cheaper and less resource-intensive, and more accurate than the tuberculin skin test, though still needing a return visit. It adds that wider access barriers in low-income settings will not be solved by introducing the new tests alone.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1161",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1161",
      "pageStart": 1161,
      "pageEnd": 1161,
      "target": "../sections/page-1161.md",
      "narrative": null,
      "content": "26 \n \nConclusion \nIn conclusion, TBST were deemed to be more acceptable and equally feasible to the TST, with \ngreater benefits for health equity driven by higher sensitivity and specificity. TBST may be more \npreferable for use in more resource-limited settings compared to the TST. IGRA appears the \npreferred test option in high-income countries with the infrastructure to support their use; it is \nunlikely that widespread use of the IGRA would change in these settings. End-users were found to \nvalue attributes such as test accuracy, convenience, a positive patient experience, cost, and resour",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Concludes that novel skin tests were judged more acceptable than and as feasible as the tuberculin skin test, with greater equity benefits, while IGRA remains preferred in well-resourced high-income settings.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1162",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1162",
      "pageStart": 1162,
      "pageEnd": 1162,
      "target": "../sections/page-1162.md",
      "narrative": null,
      "content": "27 \n \nWork Package C - Discrete choice experiment (DCE) \nThis activity intends to answer following objective: \nObjective D: To explore end-user values and preferences for TB testing by investigating the trade-\noffs end users are willing to make between test attributes derived from systematic reviews and in-\ndepth user interviews, and determine the order of preference of different test attributes using a \ndiscrete choice experiment. \n3.3.1 Aims \nThe aim of this work package was to conduct a discrete choice experiment (DCE) in order to explore \nend-user (consumer and provider) values and prefere",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Introduces Work Package C, a discrete choice experiment to explore how test users and providers trade off test attributes. It describes the online cross-sectional survey design and how attributes covering acceptability, accuracy, feasibility and equity were chosen.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1163",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1163",
      "pageStart": 1163,
      "pageEnd": 1163,
      "target": "../sections/page-1163.md",
      "narrative": null,
      "content": "28 \n \nA choice task is where a participant is given a choice between different scenarios based on different \nattribute and level combinations. We constructed a DCE with choice tasks that included two options \nof hypothetical TBI test alternatives and an opt-out alternative of choosing neither, which was \nincluded to estimate preferences more relevant to real-world choices.   \nSeparate DCE choice tasks were created for respondents who were providers of TB services and for \nrespondents who were consumers. The attributes included in each DCE differed slightly, though \nsome were common across both",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains how choice tasks offered two hypothetical tests plus an opt-out, with separate versions for providers and consumers, and how a 10-task efficient design was built from the many possible profiles using NGENE software after piloting.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1164",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1164",
      "pageStart": 1164,
      "pageEnd": 1164,
      "target": "../sections/page-1164.md",
      "narrative": null,
      "content": "1 \n \nTable C1: List of attributes and levels included in both provider and consumer DCEs  \n \nProvider DCE \nConsumer DCE \nAttribute \nLevels \nDescription  \nLevels \nCost to healthcare practice \nUSD 100 \nTime required for test results to \ncome back \n1 day \n  \nUSD 50 \n  \n3 days \n  \nUSD 25 \n  \n7 days \nIf a follow-up appointment is \nneeded to get test results \nIn-person follow-up appointment needed \nIf a follow-up appointment is \nneeded to get test results \nIn-person follow-up appointment needed to get \nresults  \n  \nNo follow-up appointment required \n  \nNo follow-up appointment required \nChance of th",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Table C1 lists the attributes and levels in the provider and consumer choice experiments, such as cost, waiting time, follow-up visits, false positive and false negative rates, test location and how the test is done.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1165",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1165",
      "pageStart": 1165,
      "pageEnd": 1165,
      "target": "../sections/page-1165.md",
      "narrative": null,
      "content": "3.3.2.3 Recruitment  \nParticipants for the DCE were recruited from the same research networks used to recruit the users for \nthe interviews for Work Package B e.g. from a clinical, research and community setting: \n• \nTB Treatment Action Group \n• \nTB Civil Societies \n• \nMembers of the WHO Guideline Development Group for TBST \n• \nGlobal Coalition of TB Activists (GCTA) \n• \nStop TB Partnership \n• \nWHO TBI Task Force \nPotential participants were provided with an online link, which led them to the study information \nsheet. At the bottom of the information sheet webpage, participants were informed t",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Describes recruitment of participants through tuberculosis networks and civil society groups, online consent, and analysis using multinomial logit models. It begins reporting respondent characteristics, with most respondents from India, the United States, South Africa, Pakistan and Zimbabwe.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1166",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1166",
      "pageStart": 1166,
      "pageEnd": 1166,
      "target": "../sections/page-1166.md",
      "narrative": null,
      "content": "Other \n1 (1%) \n \nAge Category \n \n \n    18-35 \n32 (17%) \n13 (27%) \n    36-55 \n106 (57%) \n24 (50%) \n    56-75 \n44 (24%) \n11 (23%) \n    75+ \n4 (2%) \n0 (0%) \n \nFigure C1: Location of DCE respondents  \n \n3.3.3.2 Choice analysis \nTable 4 shows choice model results, analysing DCE data for providers and consumers using multinomial \nlogistic regression models. Where coefficients are significantly greater than zero, an attribute can be \ninterpreted as having a relatively positive impact on participant utility. The magnitude of coefficients \ncan be directly compared within each model. The statistical sig",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Gives the age breakdown and location of choice experiment respondents and explains how to read the multinomial logit results. It begins presenting the choice model results for consumers and providers.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1167",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1167",
      "pageStart": 1167,
      "pageEnd": 1167,
      "target": "../sections/page-1167.md",
      "narrative": null,
      "content": "Chance of the test wrongly saying there is TB infection when \nthere is not (false positive) \n-0.790*** \n(0.181) \n-0.189** \n(0.0819) \nChance of the test wrongly saying there is no TB infection \nwhen there is (false negative) \n-0.967*** \n(0.184) \n-0.169* \n(0.0960) \nWaiting time in days for test result \n-0.0953* \n(0.0523) \n(not presented as an \nattribute) \nPlace of the test (Reference: hospital) \n  \n      Primary care (e.g. family doctor, GP, pharmacy) \n0.427** \n(0.179) \n(not presented as an \nattribute) \n      Community (e.g. workplace or your house) \n0.395* \n(0.221) \nHow the test is performed (R",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Presents the choice model coefficients, showing that both consumers and providers strongly disliked false positive and false negative results and that consumers preferred testing in primary care or the community over hospitals. It notes that consumers weighed false negatives slightly more than false positives, while providers did the reverse.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1168",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1168",
      "pageStart": 1168,
      "pageEnd": 1168,
      "target": "../sections/page-1168.md",
      "narrative": null,
      "content": "Panel A shows results for consumers, and panel B results for providers. A positive value indicates that \nrespondents are willing to accept a lower likelihood of a false-positive result for that attribute.  \nPanel A, example interpretation: Consumers would be willing to accept an 8.4 percentage point greater \nfalse positivity rate to avoid a seven-day wait for test results.  \nPanel B, example interpretation: Providers would be willing to forego 8 percentage points of false \npositivity rate for a test which did not require an in-person follow-up appointment \n3.3.4 Summary of Work Package C findi",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Explains how to read willingness-to-accept estimates, for example that consumers would accept a higher false positive rate to avoid a week's wait for results. It summarises that 236 respondents most valued test accuracy, followed by other attributes such as convenience.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    },
    {
      "id": "library/10665-381036/blocks/prose-page-1169",
      "types": [
        "folio-assistant-core:Prose",
        "doco:Section"
      ],
      "kind": "prose",
      "title": "Page 1169",
      "pageStart": 1169,
      "pageEnd": 1169,
      "target": "../sections/page-1169.md",
      "narrative": null,
      "content": "systematic review evidence to design an initial DCE task for piloting.  \nThis DCE study has several limitations. First, DCEs require choices between hypothetical alternatives \nwhich may not reflect real-world choices accurately or comprehensively. Although there is evidence \nthat DCEs predict real-world choices reasonably well29, there remains the potential for hypothetical \nbias in these data if respondents chose differently in the online survey, than they would given different \nalternatives in real-life. Nevertheless, that preferences align with a priori expectations is an indication \nof int",
      "truncated": true,
      "provenance": "ingested",
      "summary": {
        "status": "draft",
        "text": "Sets out limitations of the choice experiment, including possible hypothetical bias, recruitment of only about a quarter of the target consumer sample, and too few responses to explore differences between geographic settings.",
        "state": "draft",
        "draftedBy": {
          "kind": "agent",
          "id": "claude-code",
          "model": "not-disclosed"
        },
        "draftedAt": "2026-11-11"
      }
    }
  ]
}
