key: cord-0813432-2pogtk78
authors: Telisinghe, L; Ruperez, M; Amofa-Sekyi, M; Mwenge, L; Mainga, T; Kumar, R; Hassan, M; Chaisson, L.H; Naufal, F; Shapiro, A.E; Golub, J.E; Miller, C; Corbett, E.L; Burke, R.M; MacPherson, P; Hayes, R.J; Bond, V; Daneshvar, C; Klinkenberg, E; Ayles, H.M
title: Does tuberculosis screening improve individual outcomes? A systematic review
date: 2021-09-22
journal: EClinicalMedicine
DOI: 10.1016/j.eclinm.2021.101127
sha: bb15177d39dfeff910f8b2a94d29736cff2f6c89
doc_id: 813432
cord_uid: 2pogtk78

BACKGROUND: To determine if tuberculosis (TB) screening improves patient outcomes, we conducted two systematic reviews to investigate the effect of TB screening on diagnosis, treatment outcomes, deaths (clinical review assessing 23 outcome indicators); and patient costs (economic review). METHODS: Pubmed, EMBASE, Scopus and the Cochrane Library were searched between 1/1/1980-13/4/2020 (clinical review) and 1/1/2010-14/8/2020 (economic review). As studies were heterogeneous, data synthesis was narrative. FINDINGS: Clinical review: of 27,270 articles, 18 (n=3 trials) were eligible. Nine involved general populations. Compared to passive case finding (PCF), studies showed lower smear grade (n=2/3) and time to diagnosis (n=2/3); higher pre-treatment losses to follow-up (screened 23% and 29% vs PCF 15% and 14%; n=2/2); and similar treatment success (range 68-81%; n=4) and case fatality (range 3-11%; n=5) in the screened group. Nine reported on risk groups. Compared to PCF, studies showed lower smear positivity among those culture-confirmed (n=3/4) and time to diagnosis (n=2/2); and similar (range 80-90%; n=2/2) treatment success in the screened group. Case fatality was lower in n=2/3 observational studies; both reported on established screening programmes. A neonatal trial and post-hoc analysis of a household contacts trial found screening was associated with lower all-cause mortality. Economic review: From 2841 articles, six observational studies were eligible. Total costs (n=6) and catastrophic cost prevalence (n=4; range screened 9-45% vs PCF 12-61%) was lower among those screened. INTERPRETATION: We found very limited patient outcome data. Collecting and reporting this data must be prioritised to inform policy and practice. FUNDING: WHO and EDCTP.

Despite effective, curative treatment, tuberculosis (TB) is a leading infectious cause of death worldwide [1] . In most TBendemic settings, standard case-detection through routine services (passive case-finding [PCF] ), is the mainstay of access to TB diagnosis and treatment [2, 3] . This may be augmented by facility-based TB screening in specific high-risk populations, such as people living with HIV/AIDS. But these measures alone do not identify the substantial burden of undiagnosed TB in these settings, or effectively reach the poor and vulnerable who face barriers to seeking health care [3] [4] [5] . In 2019, »3 million TB patients were either not diagnosed or not notified [1] . If untreated, TB is associated with high mortality and morbidity [6] . Therefore, closing the case-detection gap by improving access to TB diagnosis and treatment is a priority.

One strategy to address this is TB screening, which encompasses a wide range of activities aimed at detecting and treating TB patients earlier in their clinical course [4 ,5] . This should improve the individual's clinical outcomes, [4, 5] a requirement for traditional screening programmes [7] . While infectious diseases screening can have both individual and population effects, [4] understanding whether screening benefits the individual is critical when considering if to screen. The costs borne by people seeking TB services and their households (patient costs) can be high, hindering diagnosis and treatment [8] . Such costs can exacerbate poverty, increasing the vulnerability of individuals, with further social and health consequences [9, 10] . TB screening, by helping individuals navigate the TB care pathway, may also potentially decrease patient costs.

But evidence that TB screening improves clinical outcomes and reduces patient costs is lacking [4, 11] . Therefore, we undertook two systematic reviews to determine if TB screening 1) identifies TB patients earlier in their clinical course; improves linkage-to-care; improves treatment outcomes; and decreases deaths (clinical review) and 2) decreases patient costs (economic review).

We undertook two systematic reviews to identify studies reporting the effect of TB screening on clinical outcomes and patient costs. These were conducted to inform World Health Organization (WHO) TB screening guideline development. The Population, Intervention, Comparison(s) and Outcomes were determined in collaboration with the guideline development group (GDG), consisting of a panel of experts in the field of TB. The methods followed standard procedures for undertaking systematic reviews [12] and grading evidence quality [13] .

Studies conducted in any population group were considered. Screening was defined as any provider-initiated intervention including 1) using health information/education to encourage appropriate health-seeking behaviours, with or without increasing access to diagnostic services (enhanced case-finding [ECF]); and 2) systematic screening using any test/procedure (active case-finding in communities [ACF] and case-finding in health facilities). PCF, the comparator, was defined as the routine diagnosis of symptomatic TB patients selfpresenting to health services.

We included 23 clinical outcome indicators (Table 1) for earlier diagnosis (e.g. smear grade, body mass index), linkage-to-care (e.g. pre-treatment loss to follow-up [LTFU]), treatment outcome (e.g. success) and death (e.g. case fatality, mortality). These outcomes were all rated as critical or very important by the GDG. Clinical outcomes were assessed among bacteriologically-confirmed TB patients (culture, Xpert MTB/RIF or smear positive). Treatment success was defined as cured and treatment completed (without microbiological evidence of cure) [14] . Pre-treatment LTFU was defined as LTFU between diagnosis and treatment start. Patient cost input data (Table 1) were broadly categorised as direct medical (e.g. hospitalisation costs), direct non-medical (e.g. transportation) and indirect (e.g. lost productivity). Patient costs were assessed among all TB patients (bacteriologically-confirmed and clinically diagnosed). Catastrophic cost was defined as total costs for seeking TB care >20% of the annual household income [1] .

Clinical review: we updated the systematic review conducted by Kranzer 2013, [11] which covered the period 1/1/1980-13/10/2010 ( Figure 1 ). Articles addressing the research questions from the Kranzer 2013 review were also included in our review. Our update used the same methods as Kranzer 2013; the search was nested within a systematic review to determine the number needed to screen to detect a TB patient in any population [15] . For the number needed to screen review, Pubmed, EMBASE, Scopus and the Cochrane Library were searched from 1/11/2010-13/4/2020. Subject headings and key words covered the concepts of TB and screening (Appendix 1). The title and abstract screens were broad; articles needed to be original research on TB screening. Full text screens determined eligibility. Articles from the number needed to screen review reporting on screening for all forms of TB were assessed for eligibility for our review.

Economic review: Medline, EMBASE, Scopus and the Cochrane Library were searched from 1/1/2010-14/8/2020. Subject headings and key words covered the concepts of 1) TB; 2) screening; and 3) economic evaluations or economic/financial analysis (Appendix 1). The Global Health Cost Consortium Unit Cost Study Repository was also searched for additional articles [16] .

For both reviews, bibliographies of identified studies were searched, and authors contacted for additional data if needed.

Evidence before this study Tuberculosis (TB) remains a leading infectious cause of death worldwide, and therefore improving access to diagnosis and treatment, closing the case-detection gap and improving patient outcomes is a priority. In 2019, a MEDLINE and EMBASE search for English language articles on TB screening identified a systematic review. Synthesising data published between 1/1/ 1980-13/10/2010, it found little evidence that TB screening benefited individuals screened; patient costs were not assessed.

Synthesising evidence between 1980-2020, our systematic review investigating the effects of TB screening on patient outcomes, found 24 articles (including three trials) from 12 countries. The limited available data suggests that compared to passive case finding, TB screening may be associated with less severe disease; decreased time to diagnosis/first contact with health services; decreased deaths (among risk groups alone); decreased patient costs; and higher pre-treatment losses to follow-up. There was no difference in treatment success between screened and passive case finding groups.

With World Health Organization targets to END-TB calling for decreases in TB deaths, incidence and catastrophic costs, countries have renewed their interest in TB screening, to find, test and treat "the missing millions". We found very limited data on the individual effects of TB screening. Routine/research programme implementation must be combined with rigorous data collection and analysis of critical patient outcomes that allows the benefits and harms of TB screening to be characterised.

Only articles in English, French and Spanish were included. Both (quasi-)randomised controlled trials (RCTs) and observational studies with screened and PCF groups were eligible. Studies comparing two different screening strategies or where screening and PCF occurred in different populations (e.g. screened miners and PCF in the general population) were excluded. Observational studies not disaggregating data by screened and PCF groups were excluded. RCTs (individual and cluster [CRTs]) comparing treatment, death and cost outcomes by randomised arm were eligible, as this design can mitigate biases inherent in observational screening studies. For the clinical review, household contact screening studies where index cases formed the PCF group and household contacts the screened group were excluded as individuals from the same households are clustered.

Study selection, data extraction and risk of bias assessments were undertaken by two independent reviewers (LT, MR, MAS, MH and CD conducted the clinical review and LM, and EK conducted the economic review). Disagreements were resolved through discussion or, if required, consultation with a third reviewer.

For the clinical review, abstracts of articles were searched to shortlist studies with a control population (parallel or before-after design). For the economic review, articles were initially shortlisted based on the title and abstract. For both reviews, inclusion was based on full text review of shortlisted articles.

Data were extracted into case report forms. Variables extracted included study design, population, calendar period, screening strategy, PCF algorithm, TB case definition, participant numbers and outcome data. Methodological quality of cross-sectional studies was assessed across four domains; valid participant selection, valid exposure ascertainment, valid outcome ascertainment, and adequate control for confounders [13] . Quality assessment of CRTs was undertaken using the Cochrane Risk of Bias tool [17, 18] . For economic studies the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) statement was used [19] .

Due to the heterogeneity of included studies (populations, screening tools, effect estimates, etc), data synthesis for both reviews was narrative. For treatment success and on-treatment case fatality calculations, we only included cured, treatment completed, death, treatment failure, LTFU, and not evaluated (including transferred out) in the denominator; other outcomes reported, such as still on The clinical review was nested within a systematic review to determine the number needed to screen to detect a TB patient in any population. *represents the study selection process for the number needed to screen review.

y The starting point of the clinical review, which is reported in this manuscript. treatment, were excluded. Smear grade was recategorized, with grades scanty/1+/2+ combined to reflect lower grades (and less extensive disease) and 3+ reflecting higher grades (and more extensive disease). A sensitivity analysis was conducted recategorizing smear grades scanty/1+ as lower grade and 2+/3+ as higher grade.

Where proportions were reported, 95% confidence intervals (95%CI) were calculated using Stata version 15 (StataCorp).

The WHO commissioned this work to inform TB screening guideline development. The WHO had no role in the conduct of the study or writing the report. The corresponding and last author had access to all data and final responsibility for the decision to submit for publication. Table 2 ); seven were not reported in the previous review [20, [29] [30] [31] [32] 36, 37] . We only identified n=12/23 (52%) of the outcome indicators sought (Table 1) ; no studies reported on the remainder. All studies reported on smear and/or culture positive TB (Table 2) ; no studies reported on Xpert MTB/RIF positive TB.

Fifteen were observational studies. The characteristics of TB patients identified through screening and PCF varied across these studies (Tables 3-5 ). All had a high risk of bias for the outcomes identified (Appendix 2); most (n=11/15) did not adjust for potential confounders.

Eight observational studies were conducted in rural and/or urban populations; all were from South Asia and sub-Saharan Africa [20] [21] [22] [23] [24] [25] [26] [27] . Most (n=7/8) involved one-off house-to-house ACF strategies (n=5/7 were prevalence surveys) [20] [21] [22] [23] [24] [25] 27] . Four (50%) used symptom screening, [20, 22, 26, 27] three (38%) chest radiographs (CXRs) and symptoms, [23] [24] [25] and one (12%) prevalence survey conducted sputum smear and culture on all individuals [21] .

Three studies [20, 21, 25] reported on smear grade (Table 3 showing proportions and prevalence ratios and Appendix 3). All showed screened TB patients were less likely to have higher smear grades, but the small sample size of the screened group gave wide CIs in one [21] . Two studies conducted in the same south Indian population over consecutive calendar periods reported on pre-treatment LTFU (Table 4 ) [23, 24] . In both, the proportion LTFU among those screened was higher (screened 23% and 29% versus PCF 15% and 14%). Among individuals LTFU, none died in the screened group, while nearly 20% had died in the PCF group for whom outcomes were available [23] . Symptom duration was longer in the PCF group in one study (cough <3 weeks 13% in PCF versus 28% in screened group) [25] but shorter in another (mean cough duration 6.8 weeks in PCF versus 10.3 weeks in screened group) [20] . One study found no difference in time to treatment start between screened and PCF groups [22] .

Four studies involving different screening strategies (symptom; CXR; and smear/culture screening) reported on treatment outcomes (Table 5 showing proportions and prevalence ratios). In three the proportions with treatment success among screened and PCF groups was similar, ranging from 68-80% [21, 25, 26] . Two studies also reported on pre-treatment LTFU; both only provided data for the screened group (26-32%) [21, 25] . There was no difference in the proportion who died between screened (range 6-8%) and PCF (range 4-11%) groups in four studies [21, [25] [26] [27] . There was no difference in the proportion LTFU during TB treatment between screened (range 6-20%) and PCF (range 8-19%) groups [25, 26] .

One CRT, conducted in 32 contiguous rural Ethiopian communities with difficult access to health care, used monthly ECF with outreach clinics to initiate diagnosis (continued at health facilities through routine services) over 1 year in 12 intervention communities ( Table 2, Table 5 and Appendix 2) [28] . There was no difference in TB patient characteristics, treatment success, on-treatment case fatality or on-treatment LTFU by study arm. Data on pre-treatment LTFU was not provided. But pre-treatment symptom duration was significantly lower in the intervention group (median difference between intervention and control group -47 days; 95%CI -76 to -19; 55-60% reduction in duration in the last three quarters compared to the first quarter in intervention communities, with corresponding 3-20% fall in control communities). Because of insufficient information to assess one bias domain, the risk of bias assessment raised some concerns.

Seven observational studies reported on risk groups, including prisoners, [29] [30] [31] [32] migrants, [33] miners, [34] and homeless people. [32, 35] . Four involved established European and South African CXR screening programmes [32] [33] [34] [35] . Three studies from India and Brazil reported on one-off/limited ACF using symptoms [29] [30] [31] .

One Indian study found no difference in smear grade among screened and PCF groups (Table 3 showing proportions and prevalence ratios) [29] . Three European and one Brazilian study reported on smear positivity among culture-confirmed TB patients [31] [32] [33] 35] . The proportion with positive smears was lower in those screened in three [31] [32] [33] . One study showed no association but small sample sizes gave wide CIs in both study groups [35] . No studies reported on pre-treatment LTFU (Table 4 ). Symptom duration was shorter in the screened group in two studies (prevalence of diagnosis delay 50 days was 23% lower in the screened group in an Indian study, [30] and the median symptom duration was 7.5 weeks in the PCF versus 0.0 weeks in the screened group in a study from the Netherlands [33] . Time to treatment start in one Indian study [30] found no difference between the screened and PCF groups.

Three studies (including two established CXR screening programmes) reported on treatment outcomes (Table 5 showing proportions and prevalence ratios). The proportions with treatment success among screened and PCF groups was similar, ranging from 80-90% in two [29, 33] . In one Indian study reporting on one-off symptom screening, there was no difference in case fatality among screened and PCF groups [29] . Two studies reporting on »4-5 years of data from established CXR screening programmes among migrants to the Netherlands and South African miners showed higher case fatality among the PCF group (PCF versus screened odds ratio [OR] 15.3; 95%CI 2.0-118.0; adjusted OR 5.6; 95%CI 2.6-12.2 respectively) [33, 34] . There was no difference in the proportion LTFU during TB treatment between screened (range 6-10%) and PCF (range 7-10%) groups [29, 33] .

Two CRTs were identified ( Table 2, Table 5 and Appendix 2) [36, 37] . One among Indian neonates compared fortnightly ACF over 2 years, in 297 intervention communities to PCF in 295 control communities [36] . Screening was associated with lower all-cause mortality compared to PCF (adjusted OR 0.68 [95%CI 0.47-0.98]), which was attributed to decreases in pneumonia/respiratory infections. The risk of bias was high which could work to underestimate the effect of screening on mortality. A CRT among Vietnamese household contacts of TB patients, compared CXR and symptom screening at 0, 6, 12 and 24 months in 36 intervention communities to PCF in 34 control communities [37] . Screening was associated with lower all-cause mortality compared to PCF (risk ratio 0.60 [95%CI 0.50-0.80]). The risk of bias assessment raised some concerns as the data represented a posthoc analysis.

From 2841 articles, six observational studies were eligible [38] [39] [40] [41] [42] [43] ( Figure 2 and Table 2 ); none were included in the previous review. Most were from South Asia (n=4; 67%), [38] [39] [40] [41] with one from South East Asia, [42] and one from sub-Saharan Africa [43] . Most studies included general populations (n=4; 67%); [38] [39] [40] 43] three involved house-to-house screening [38, 39, 43] . Risk groups were those with structural risk factors (n=1), [41] household and neighbourhood contacts (n=1), [42] and social contacts (n=1) [39] of TB patients, and health facility attendees (n=2) [39, 40] . Four studies (67%) used symptom screening alone, [39] [40] [41] 43] whereas two (33%) used CXR and symptoms. [38, 42] . The analyses undertaken varied; four performed cost analysis [38, 39, 41, 42] and two conducted cost-effectiveness analysis [40, 43] . All studies reported findings transparently; three [38] [39] [40] met all CHEERS checklist criteria (Appendix 4).

Data were summarised using different measures (means, medians). The illness periods for which costs were reported varied; two studies reported diagnosis costs alone, [41, 43] two pre-treatment and treatment costs, [39, 42] one diagnosis and treatment costs, [38] and one pre-diagnosis, diagnosis and treatment costs [40] (Table 2 and 6; Appendix 5). While cost inputs and granularity of reporting varied across studies, all calculated aggregated costs for the reported illness period (Table 6 and Appendix 5). In all studies, higher total costs were incurred in the PCF compared to screened group. Four studies assessed catastrophic cost prevalence, which was higher in the PCF (range 12-61%) compared to screened (range 9-45%) group [38, 39, 41, 42] . In two Indian studies, using house-to-house screening among general populations [38] and those with structural risk factors, [41] total costs and catastrophic costs (on multivariable analysis) were significantly lower in the screened compared to PCF groups. In two studies with small sample sizes, among Cambodian household and neighbourhood contacts of TB patients [42] and among mainly outpatient attendees and social contacts of TB patients in Nepal, [39] there was no statistically significant difference in total costs and catastrophic costs on univariable analysis between screened and PCF groups. Two studies did not assess differences in mean total costs or report catastrophic costs [40, 43] .

We synthesised literature published between 1980-2020, to generate up-to-date evidence for the individual effects of TB screening. We found very few studies addressing the review questions. The WHO END-TB strategy sets out ambitious targets to reduce TB death, incidence and catastrophic costs by 2035 [44] . At the 2018 United Nations General Assembly high-level meeting, world leaders reaffirmed their commitment to ending TB [45, 46] . At a time of unprecedented political commitment to find, test and treat TB patients, evidence for strategies such as TB screening to inform in-country decision making globally, is vital. Further, the reversal in TB control efforts and case-detection due to the COVID-19 pandemic [47, 48] may going forward, make TB screening even more important.

A general challenge with interpreting the findings is the observational design of most studies. This is compounded by differences in reported outcome measures, insufficient data on the care cascade, unadjusted analyses, small sample sizes, and length-time bias (where screening may detect individuals with less severe indolent disease who may have different characteristics, longer disease course and better outcomes including survival, than those who are identified through PCF). These limitations must be kept in mind when interpreting results. Definitive evidence for the effects of TB screening requires well-conducted RCTs. However, these require large sample sizes, long term follow-up and are resource intensive. We only identified three RCTs, conducted over relatively short time-periods (1-2 years) [28, 36, 37] . Therefore, insights from routine programme implementation are essential. While overall screening approaches will depend on the context and available resources, general principles dictate that screening is not one-off, is integrated into health systems, with quality-assured diagnosis and treatment services [4, 7] . We only identified four studies (all in risk groups) reporting on established screening programmes [32] [33] [34] [35] . But there was general consistency in most findings, irrespective of the screening strategy used.

TB screening, by engaging individuals earlier into care, should result in earlier diagnosis when disease is less severe [4] . Smear grade and proportion smear positive among culture-confirmed TB patients was lower in the screened group in most studies with larger sample sizes, suggesting screening does identify individuals with less severe disease. Length-time bias may explain this. But the reported reduction in pre-diagnosis symptom duration among those screened, while subject to recall bias, suggests earlier diagnosis plays a role. If individuals are identified earlier, when disease is less severe, and linked to care, this should translate to better outcomes for the individual [4] .

Studies consistently showed no difference in treatment success between screened and PCF groups. This could be a true finding (screening does not improve treatment success). Or it may be due to potential confounders or the inherent limitations of routine data, where identifying TB patients screened from those self-presenting can be challenging and successful outcomes may be over-ascertained, potentially biasing the effect towards the null. Data on pre-treatment LTFU, while limited and not generalisable, suggests pre-treatment LTFU is high among screened TB patients; in one study, no deaths were reported in the screened group [23] . In the PCF group, there was high pre-treatment case fatality, [23] similar to other reports [49] . Therefore, on-treatment outcomes, which ignore deaths pretreatment, may underestimate the effects of screening. Two studies (Churchyard 2000 and Verver 2001) found screening was associated with lower case fatality, [33, 34] but due to their observational nature we cannot exclude length-time bias and uncontrolled confounders. Both report on established CXR screening programmes, with large sample sizes, access to good health systems and better reporting of deaths. While neither study report on pre-treatment LTFU, individuals treated could be more representative of those diagnosed. Churchyard 2000, among miners did not report treatment success by screened and PCF groups [34] . Verver 2001, showed no difference in treatment success, [33] but this study among migrants, had few deaths overall which may reflect a healthy migrant effect, giving better overall outcomes across study groups. Two CRTs (Jenum 2018 in neonates and Fox 2018 in household contacts of TB patients) found screening was associated with lower all-cause mortality, [36, 37] with Fox 2018, showing no difference in on-treatment outcomes (among all TB patients) between study groups [37] . The limitations of these CRTs (generalisability, post-hoc analysis) need to be borne in mind when interpreting findings. But, in line with these [50, 51] . As all data represent risk groups, findings cannot be extrapolated to general populations. Pre-treatment LTFU, while likely to be setting-specific, can be frequent with interventions targeting "well" individuals. Programmes should ensure that all individuals diagnosed are linked to treatment, with context-specific barriers to engaging with care identified and mitigated. A CRT in rural Ethiopia where health care access is difficult, compared ECF to ECF plus community-based care (sputum collection, providing treatment and supporting adherence) by community health workers over one year [52] . Treatment success was significantly higher in the latter group, highlighting how combining screening with strategies that minimise pre-treatment LTFU can increase treatment success. Further, if all individuals diagnosed at an earlier stage are not started on treatment, reducing transmission, population-level benefits [4] shown in trials [53, 54] may not be realised.

Due to the limitations of the identified economic studies (e.g. differences in the cost inputs and illness periods; small sample sizes; recall bias; and unadjusted analyses) we cannot directly compare findings between studies. Further, the data are mostly from South Asia, limiting generalisability. Nevertheless, all studies consistently showed lower total costs and catastrophic cost prevalence among those screened. While we did not assess screening costs/cost-effectiveness from a health system perspective, this can be high. When viewed from a societal perspective, there may be potential offsets to these costs. But, given the limitations of the included studies, only cautious conclusions can be drawn. Patient costs are often reported as barriers to accessing TB care.8,55-57 Therefore, standardising the collection and reporting of patient cost inputs as part of routine programme monitoring could help identify how interventions affect this patient important outcome, guiding policy making.

These reviews have several limitations. We only searched four databases; the grey literature was not searched. Only English, French and Spanish articles were included. The economic review only included articles from 2010. Therefore, some relevant articles may have been missed. As studies were heterogeneous, we could not meta-analyse the data. We did not assess publication bias.

An important finding was the limited data on individual outcomes, despite many publications on TB screening studies/programmes [58] . Going forward, studies/programmes must prioritise reporting this data, along with the screening cascade. Evaluations should be carefully designed, to identify appropriate control groups and adjust for potential confounders, allowing valid comparisons across diagnosed TB patients in screened and unscreened populations.

In conclusion, we found very limited data on the effect of TB screening on individual outcomes. Routine/research programmes must prioritise collecting and reporting this data.

All data are included within the article and supplementary material. LHC reports a contract from WHO TB Programme to Jonathan Golub for systematic review of ACF for TB and sub-contract/consulting for JHU for systematic review of ACF for TB.

JEG received a contract provided to Johns Hopkins University to conduct systematic reviews for the WHO's TB screening guidelines; received an NIH grant to conduct TB case finding in India, a second to test for and treat latent TB infection in Brazil; received UNITAID grants to conduct implementation research around latent TB infection in several African countries; and sat on the Scientific Advisory Board for the Aurum Institute in November 2019.

CM is a salaried staff of the WHO and is involved in policy development on TB. CM alone is responsible for the views expressed in this publication and they do not necessarily represent the decisions or policies of WHO.

ELC has received a Wellcome Trust Senior Research Fellowship in Clinical Science: 200901/Z/16/Z to their institution.

RMB reports salary support from my Wellcome Trust Clinical PhD fellowship, awarded through her institution, grant number 203905/ Z/16/Z; received payment from WHO to her institution for work on systematic review linked to this present review (but different to this review).

PM reports that he is funded by Wellcome (206575/Z/17/Z). EK has a consultancy contract with LSHTM for other work, this work was done under that umbrella.

HMA reports WHO consultancy for the work for the guideline development process; reports that EDCTP fund the larger TREATS consortium as a grant paid to her institution that covers some of her time; reports that she is a member of the technical review panel of the Global Fund and receive honoraria for her work.

All other authors have nothing to declare. The designations used and the presentation of the material in this publication do not imply the expression of any opinion whatsoever on the part of WHO concerning the legal status of any country, territory, city or area, or of its authorities, nor concerning the delimitation of its frontiers or boundaries.

This work was commissioned by the WHO to update its TB screening guidelines and made possible through a grant from the WHO Global TB Programme. LT, MR, MAS, LM, TM, RK, RJH, VB, EK, HMA are funded by part of the EDCTP2 programme supported by the European Union (grant number RIA2016S-1632-TREATS). RMB, ELC and PM are funded by the Wellcome Trust (203905/Z/16/Z, 200901/Z/16/ Z and 206575/Z/17/Z respectively). AES is supported by an NIH grant K23AI140918. The WHO, EDCTP, Wellcome Trust and NIH had no role in the conduct of the study or writing the review.

Supplementary material associated with this article can be found in the online version at doi:10.1016/j.eclinm.2021.101127.

What can National TB Control Programmes in low-and middle-income countries do to end tuberculosis by 2030?

Passive case finding for tuberculosis is not enough

Systematic screening for active tuberculosis: principles and recommendations

Systematic screening for active tuberculosis: rationale, definitions and key considerations

Natural history of tuberculosis: duration and fatality of untreated pulmonary tuberculosis in HIV negative patients: a systematic review

World Health Organization. Principles and practice of screening for disease

Financial burden for tuberculosis patients in low-and middle-income countries: a systematic review

The Social Consequences of Poverty: An Empirical Test on Longitudinal Data

The influence of income on health: views of an epidemiologist

The benefits to communities and individuals of screening for active tuberculosis disease: a systematic review

Cochrane Handbook for Systematic Reviews of Interventions version 6.2

GRADE handbook for grading quality of evidence and strength of recommendations Updated

World Health Organization. Definitions and reporting framework for tuberculosis

Overview and systematic review of the number needed to screen for active TB

Global Health Cost Consortium. The Unit Cost Study Repository

Cochrane Handbook for Systematic Reviews of Interventions version 6

Revised Cochrane risk of bias tool for randomized trials (RoB 2.0): Additional considerations for cluster-randomized trials

Consolidated Health Economic Evaluation Reporting Standards (CHEERS) statement

Are patients with pulmonary tuberculosis who are identified through active case finding in the community different than those identified in healthcare facilities?

Comparison of symptoms and treatment outcomes between actively and passively detected tuberculosis cases: the additional value of active case finding

Prevalence of smear-positive pulmonary tuberculosis in a rural district of Ethiopia

Failure to initiate treatment for tuberculosis patients diagnosed in a community survey and at health facilities under a DOTS programme in a district of South India

Gender disparities in tuberculosis: report from a rural DOTS programme in south India

Are community surveys to detect tuberculosis in high prevalence areas useful? Results of a comparative study from Tiruvallur District, South India

Tuberculosis case finding in remote mountainous areasÀare microscopy camps of any value? Experience from Nepal

Tuberculosis case-finding in Eastern Nepal

Tuberculosis case-finding through a village outreach programme in a rural setting in southern Ethiopia: community randomized trial

Active versus passive case finding for tuberculosis in marginalised and vulnerable populations in India: comparison of treatment outcomes

Patient characteristics, health seeking and delays among new sputum smear positive TB patients identified through active case finding when compared to passive case finding in India

Impact of mass-screening on tuberculosis incidence in a prospective cohort of Brazilian prisoners

Active case finding for pulmonary tuberculosis using mobile digital chest radiography: an observational study

Screening for pulmonary tuberculosis among immigrants: estimated effect on severity of disease and duration of infectiousness

Factors associated with an increased case-fatality rate in HIV-infected and non-infected South African gold miners with pulmonary tuberculosis

The diagnosis and management of tuberculosis in common hostel dwellers

Incidence of tuberculosis and the influence of surveillance strategy on tuberculosis case-finding and all-cause mortality: a cluster randomised trial in Indian neonates vaccinated with BCG

Household-Contact Investigation for Detection of Tuberculosis in Vietnam

Catastrophic costs due to tuberculosis in South India: comparison between active and passive case finding

The role of active case finding in reducing patient incurred catastrophic costs for tuberculosis in Nepal

The costeffectiveness of incentive-based active case finding for tuberculosis (TB) control in the private sector Karachi

Active case finding among marginalised and vulnerable populations reduces catastrophic costs due to tuberculosis diagnosis

Mitigating Financial Burden of Tuberculosis through Active Case Finding Targeting Household and Neighbourhood Contacts in Cambodia

Cost-effectiveness analysis of community active case finding and household contact investigation for tuberculosis case detection in urban Africa

World Health Organization. The End TB Strategy: global strategy and targets for tuberculosis prevention

United Nations. Political declaration of the UN general assembly high-level meeting on the fight against tuberculosis

World Leaders Reaffirm Commitment to End Tuberculosis by 2030, as General Assembly Adopts Declaration Outlining Actions for Increased Financing. Treatment Access

The impact of COVID-19 on TB: a review of the data

Commentary: Lessons from the COVID-19 global health response to inform TB case finding

Pre-treatment loss to follow-up in tuberculosis patients in low-and lower-middle-income countries and high-burden countries: a systematic review and meta-analysis

Twelve-monthly versus six-monthly radiological screening for active case-finding of tuberculosis: a randomised controlled trial2011

Rapid urine-based screening for tuberculosis in HIV-positive patients admitted to hospital in Africa (STAMP): a pragmatic, multicentre, parallel-group, double-blind

Health extension workers improve tuberculosis case detection and treatment success in southern Ethiopia: a community randomized trial

Community-wide Screening for Tuberculosis in a High-Prevalence Setting

Comparison of two active case-finding strategies for community-based diagnosis of symptomatic smear-positive tuberculosis and control of infectious tuberculosis in Harare, Zimbabwe (DETECTB): a cluster-randomised trial

Barriers in the access, diagnosis and treatment completion for tuberculosis patients in central and western Nepal: A qualitative study among patients, community members and health care workers

Patients' perspectives of tuberculosis treatment challenges and barriers to treatment adherence in Ukraine: a qualitative study

Barriers to initiating tuberculosis treatment in sub-Saharan Africa: a systematic review focused on children and youth

Community-based active case-finding interventions for tuberculosis: a systematic review