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

1Tuberculosis is the leading infectious cause of death worldwide. An estimated 3 million people with active tuberculosis were either not diagnosed or were diagnosed but not notified through national reporting systems in 2019. The so-called missing millions of people with undiagnosed or untreated active tuberculosis are at risk of death and severe illness, and can transmit tuberculosis to others in their households and communities. Declines in global tuberculosis incidence have been slow and, at the rate of current progress, are unlikely to meet the WHO End TB Strategy targets to reduce incidence by 90% and tuberculosis deaths by 95% by 2035. Therefore, implementation of effective, evidence-based strategies that can increase diagnosis and treatment of tuberculosis, and potentially reduce tuberculosis transmission, are urgently required.

2Community-based tuberculosis screening, delivered through active case-finding interventions, has been widely implemented throughout the 20th and 21st centuries, but with varying levels of intensity between regions and over time. Because tuberculosis care and prevention interventions that rely primarily on passive case detection and health facility-based screening strategies have insufficiently reduced tuberculosis incidence, many national tuberculosis programmes have promoted community-based active case-finding interventions.

3Active case-finding encompasses a wide range of activities that range in intensity from health promotion campaigns and community mobilisation, through to systematic identification and offering screening and diagnosis to entire populations. Generally, active case-finding aims to diagnose tuberculosis either in those who do not recognise that they have symptoms, or those who do recognise symptoms but for whatever reason do not, or cannot, access services at health-care facilities., We expect that an effective community-based active case-finding intervention would initially increase the number of people diagnosed with tuberculosis and started on tuberculosis treatment (ie, increase case notifications) in a given setting. When this occurs, tuberculosis transmission might decline because people are diagnosed earlier in their disease course, potentially reducing the length of time in which an individual is infectious to others., If tuberculosis active case-finding is successful, we would expect to see a reduction in tuberculosis disease prevalence and in prevalence and incidence of tuberculosis infection in children.

4Evidence before this studyActive case-finding for tuberculosis is one of the longest running and most widely implemented screening interventions. We did preliminary scoping review searches in PubMed and MEDLINE in February, 2019, using medical subject headings, keyword, and title word search terms including “tuberculosis”, “mass screening”, and “case finding”. We also sought expert opinion (in sessions convened to facilitate the 2020 WHO tuberculosis screening guideline development process) to identify studies related to active case-finding for tuberculosis. We identified a systematic review from 2013 on the individual-level and community-level effects of tuberculosis active case-finding, which covered literature published up until December, 2011. The review concluded that the benefits of active case-finding for tuberculosis disease remained uncertain.Added value of this studySince the previous systematic review published in 2013, several large randomised and non-randomised studies evaluating the effectiveness of community-based active case-finding for tuberculosis have been published. Our systematic review synthesises this new evidence and includes data from 36 studies from 16 countries, comprising at least 110 million person-years of follow-up in studies done between 1980 and 2020. With new evidence from two large cluster-randomised trials done in South Africa and Zambia and in Vietnam that were not included in the previous systematic review, we found moderate quality evidence from some of the reviewed studies that active case-finding, when implemented with sufficient coverage and intensity in high-prevalence settings, can positively affect the community epidemiology of tuberculosis.Implications of all the available evidenceHealth planners and national tuberculosis programmes should consider the implementation of active case-finding for tuberculosis interventions as part of well designed research protocols in urban populations with a high prevalence of undiagnosed tuberculosis and in other populations, to contribute evidence to outstanding knowledge gaps.

5Evidence before this study

6Active case-finding for tuberculosis is one of the longest running and most widely implemented screening interventions. We did preliminary scoping review searches in PubMed and MEDLINE in February, 2019, using medical subject headings, keyword, and title word search terms including “tuberculosis”, “mass screening”, and “case finding”. We also sought expert opinion (in sessions convened to facilitate the 2020 WHO tuberculosis screening guideline development process) to identify studies related to active case-finding for tuberculosis. We identified a systematic review from 2013 on the individual-level and community-level effects of tuberculosis active case-finding, which covered literature published up until December, 2011. The review concluded that the benefits of active case-finding for tuberculosis disease remained uncertain.

7Added value of this study

8Since the previous systematic review published in 2013, several large randomised and non-randomised studies evaluating the effectiveness of community-based active case-finding for tuberculosis have been published. Our systematic review synthesises this new evidence and includes data from 36 studies from 16 countries, comprising at least 110 million person-years of follow-up in studies done between 1980 and 2020. With new evidence from two large cluster-randomised trials done in South Africa and Zambia and in Vietnam that were not included in the previous systematic review, we found moderate quality evidence from some of the reviewed studies that active case-finding, when implemented with sufficient coverage and intensity in high-prevalence settings, can positively affect the community epidemiology of tuberculosis.

9Implications of all the available evidence

10Health planners and national tuberculosis programmes should consider the implementation of active case-finding for tuberculosis interventions as part of well designed research protocols in urban populations with a high prevalence of undiagnosed tuberculosis and in other populations, to contribute evidence to outstanding knowledge gaps.

11Despite widespread implementation of active case-finding interventions globally, the evidence for effectiveness and the optimal approaches to delivering active case-finding interventions remain uncertain. Therefore, we aimed to systematically appraise evidence for the effectiveness of active case-finding interventions on tuberculosis case notifications, tuberculosis disease prevalence, and tuberculosis infection incidence and prevalence.

12We systematically reviewed the literature for studies that reported the effects of active case-finding interventions on tuberculosis epidemiological indicators. Our literature search was an update of a 2013 systematic review by Kranzer and colleagues, which covered the period between Jan 1, 1980, and Oct 13, 2010, with additional searches by that group up to the end of 2011. We did a systematic search of PubMed, Embase, Scopus, and Cochrane Library for papers published between Nov 1, 2010, and Feb 14, 2019 (subsequently updated to April 13, 2020). The search terms used are described in the appendix (pp 15–16).

13We included studies that evaluated at least one active case-finding intervention and contained data to permit a comparison of tuberculosis epidemiology between populations exposed and not exposed to active case-finding (or populations exposed to two different methods of active case-finding). Eligible study designs included randomised controlled trials, non-randomised parallel group studies with outcome measurement before and during the intervention period (referred to as controlled before-after studies), and studies that compared outcomes before and after the intervention period in the same population (referred to as before-after studies). Because the epidemiology of tuberculosis differs substantially between children and adults, we excluded studies that were done only among children (aged <15 years). Studies must have screened at least 1000 people for tuberculosis because the prevalence of tuberculosis disease will rarely exceed 1% in any given community. If tuberculosis screening was targeted at a subset of a population but effects were measured in the wider population, the target population must have comprised at least 10% of the whole population. We excluded studies that were published before Jan 1, 1980, and studies not published in English.

14We reviewed the full text of studies included in the systematic review by Kranzer and colleagues, as well as those meeting eligibility criteria at title and abstract screen of the updated search. Each full text was reviewed by two of RMB, MN, and HRAF, and discrepancies were resolved by consensus discussion with ELC and PM. Reference lists from the included studies were examined and expert opinion on other available studies was sought from members of the WHO TB Screening Guideline Development Group.

15Data were extracted from the studies independently in duplicate (by two of RMB, MN, and HRAF) into a case record form; discrepancies were resolved by discussion and data were entered into a spreadsheet.

16Outcomes were comparisons between intervention and control groups of tuberculosis case notification rates per 100 000 population, prevalence of pulmonary tuberculosis disease (measured during a population prevalence survey following the active case-finding intervention period), and incidence or prevalence of tuberculosis infection in children (measured by tuberculin skin test or interferon γ assay surveys). For tuberculosis case notification rates, we used the number of people who started tuberculosis treatment as the numerator; however, if studies reported only numbers diagnosed with tuberculosis, we included this as a proxy for case notifications.

17To investigate the effects of active case-finding on tuberculosis case notification rates, if possible, we extracted or calculated person-years of follow-up and numbers of tuberculosis cases notified in each group. We used simple arithmetic to estimate person-years of follow-up if this was not directly reported. For randomised studies and before-after studies, case notification rate ratios (in intervention vs control populations or baseline vs endline populations) were calculated. For studies that had a non-randomised comparator and compared tuberculosis case notification rate trends over time in two groups (controlled before-after studies) we calculated the difference between case notification rate ratios in the groups with and without exposure to active case-finding. We additionally reported the authors' effect estimates (or measures of association) and CIs, if provided, and summarised any statistical adjustments for clustering and confounding. We did not calculate CIs from available grouped summary data because this would require adjustment for effects of clustering and confounders, neither of which were typically reported.

18For studies that reported effects of active case-finding on tuberculosis prevalence we extracted the size of intervention population, number of people screened for tuberculosis during active case-finding, method of tuberculosis screening, number of people in the prevalence survey or surveys, definition of a tuberculosis case, and numbers of people with tuberculosis disease. We reported summary measures of the effect of active case-finding on tuberculosis prevalence and uncertainty intervals as reported within the studies.

19Active case-finding was defined as interventions implemented in a community that endeavoured to systematically screen people for tuberculosis. A tuberculosis screen could take any form but required a personal interaction between a screener and the person being screened (eg, leaflet distribution alone would not meet this definition). The following interventions are examples of active case-finding: mobile tuberculosis screening or diagnostic clinics or sputum drop off points; mobilisation and training of community health workers and volunteers as screeners to detect tuberculosis symptoms and potentially do tuberculosis diagnostic tests in community members; door-to-door tuberculosis screening with symptom interview, sputum collection, or both. We included tuberculosis screening in closed community settings (eg, prisons) or occupational groups (eg, among miners). Tuberculosis screening interventions delivered at permanent health facilities and for contacts of people with tuberculosis did not constitute active case-finding interventions for this review.

20We classified studies according to the population groups they targeted, including general populations, remote rural populations, people living in informal urban settlements, people in prison, people experiencing homelessness, refugees or displaced people, and indigenous populations. Active case-finding interventions were often delivered concurrently alongside a wider set of tuberculosis screening and care activities (co-interventions, such as facility-based screening or laboratory strengthening). We recorded the presence of co-interventions.

21To assess risk of bias, we used Cochrane RoB 2 for randomised trials and the ROBINS-i tool for non-randomised studies. Quality assessment was done collaboratively by two authors (RMB and PM). Because we did not do a meta-analysis, we did not stratify assessments on the basis of study quality.

22WHO facilitated discussions among authors at the design stage but had no role in data collection, data analysis, data interpretation, or writing of the report.

23The literature search from Nov 1, 2010, to Feb 14, 2019, returned 23 466 unduplicated titles and abstracts; the updated search on April 13, 2020, identified a further 4417 titles and abstracts. 921 articles from these searches were identified for full text review. An additional 67 articles were identified from the systematic review by Kranzer and colleagues (published from Jan 1, 1980, to Dec 31, 2011) and from searching reference lists, resulting in a total of 988 articles that underwent full text review (figure 1). A total of 36 studies were included in our systematic review.

24We identified 30 articles reporting 28 studies on the effects of active case-finding interventions on tuberculosis case notification rates (Table 1, Table 2, Table 3; appendix pp 2–7). These studies included six cluster-randomised trials (two of which compared two active case-finding interventions to each other), 13 controlled before-after studies, and nine before-after studies. One of the cluster-randomised trials, which compared two strategies to each other, was also included as a before-after study.

25Of the 28 studies, five were done in general populations,,,,, seven were done in high-density, low-income urban areas,,,,,,, two were done in camps for internally displaced people,, four were done in remote rural populations,,,, four were done among indigenous populations (two of which were also in high-density, low-income urban areas),,,, four were done in prisons,,,, one was done in gold mines, and two were done among people experiencing homelessness.,

26Several types of active case-finding intervention were used and some studies used more than one (Table 1, Table 2, Table 3, appendix pp 2–7). The active case-finding interventions included door-to-door screening (14 studies);,,,,,,,,,,,,, sputum collection by community health workers or volunteers (13 studies);,,,,,,,,,,,, and community mobilisation combined with mobile tuberculosis screening clinics (six studies).,,,,, 17 studies included co-interventions that could affect tuberculosis detection in the community, including financial incentives for tuberculosis detection;,,,, facility-based tuberculosis screening;,, laboratory or health facility upgrading;,,, household contact tracing;,,,,, and latent tuberculosis infection treatment.,,

27Most studies (21 of 28) used tuberculosis symptom screening as the first step in the screening algorithm. Five studies used chest x-ray regardless of symptoms.,,,, Three studies used a tuberculin skin test as the first screening test.,, In one study, chest x-ray was used to screen people for tuberculosis, but sputum was additionally collected regardless of symptoms or chest x-ray findings.

28Four randomised trials assessed the effect of active case-finding on tuberculosis case notifications compared with no active case-finding.,,, Two trials showed an increase in tuberculosis case notifications,, whereas the other two trials did not show effectiveness (table 1, figure 2).,

29In non-randomised studies, populations who received active case-finding interventions consistently had higher tuberculosis case notification rates than comparison populations, with the highest case notification rate ratios in prisons, remote rural communities, and indigenous populations (figure 2). There was considerable variation in comparison and measurement periods. For the randomised trials, risk of bias was assessed as low (four studies) or as having some concerns (two studies; appendix p 14). The majority of non-randomised studies had a severe (ten studies) or critical (nine studies) risk of bias.

30Two cluster-randomised trials compared the effects of active case-finding versus no active case-finding on tuberculosis prevalence in general populations (table 4)., One further cluster-randomised trial allocated urban clusters in Zimbabwe to one of two types of active case-finding, and also evaluated change in tuberculosis prevalence before and after implementation of active case-finding, a non-randomised comparison. Six other non-randomised studies investigated the effect of active case-finding on tuberculosis prevalence in a variety of populations (table 5).,,,,,

31The ZAMSTAR study was a cluster-randomised trial in 24 communities in Zambia and South Africa. The active case-finding intervention (referred to as enhanced case-finding) included community mobilisation, education about tuberculosis in schools, fast-track sputum collection points in health-care facilities, and mobile community sputum collection points. Tuberculosis diagnosis in the active case-finding intervention was based on smear microscopy. In a post-intervention survey, the overall prevalence of culture-positive tuberculosis among those with valid sputum samples (with 90% survey participation, 73% sputum collection, and approximately two-thirds with an evaluable sputum sample) was 1277 per 100 000 people in areas without active case-finding (505 people with tuberculosis disease) and 1485 in areas with active case-finding (389 people with tuberculosis disease, adjusted mean tuberculosis prevalence ratio of 1·09, 95% CI 0·86–1·40). Among schoolchildren serially tested with tuberculin skin test before and after the intervention period, positivity among children who had been tuberculin skin test negative at baseline was 1·41 per 100 person-years in active case-finding clusters (391 children with incident tuberculosis infection) and 1·05 in non-active case-finding clusters (342 children with incident tuberculosis infection, adjusted rate ratio 1·36, 95% CI 0·59–3·14).

32In the ACT3 study, Marks and colleagues evaluated an active case-finding intervention in Vietnam that involved 3 years of annual household tuberculosis screening using sputum Xpert MTB/Rif assays for all people aged 15 years or older, regardless of symptoms, in 120 communities. A tuberculosis prevalence survey was done in the fourth year, with the denominator for the primary outcome being the total number of people who consented to be in the survey, regardless of sputum production (sputum obtained in 33·2% in the intervention group and 40·7% in the control group). In the active case-finding intervention group, the prevalence of tuberculosis (one sputum sample positive by Xpert) was 126 per 100 000 people (53 people with tuberculosis disease) and 226 per 100 000 (94 people with tuberculosis disease) in the control group (adjusted prevalence ratio of 0·56, 95% CI 0·40–0·78). A prespecified secondary outcome was prevalence of positive QuantiFERON tests among children born in 2012 (who would have been aged 1–2 years when the intervention started in 2014), as a proxy of incidence of tuberculosis infection. Among children born in 2012, 1409 children had QuantiFERON tests; 23 (3·3%) of 701 were positive among children in the intervention group and 18 (2·6%) of 705 were positive among children in the control group (prevalence ratio 1·29, 95% CI, 0·70–2·36; table 6).

33In the DETECTB study in Harare, Zimbabwe, the prevalence of culture-positive tuberculosis among a random sample of 12% of households in each of 46 clusters (23 allocated to mobile van active case-finding and 23 to door-to-door screening with symptoms and smear) before the active case-finding intervention was compared with prevalence after five rounds of active case-finding. The adjusted risk ratio for tuberculosis disease after active case-finding versus before active case-finding was 0·59 (95% CI 0·40–0·89). A further six non-randomised studies were identified from India,,, China, Brazil, and Ethiopia; three were done in the general population,, and three were done in populations with risk factors for tuberculosis (two in prisons, and one in an indigenous community). The reported estimates of effects on tuberculosis prevalence were mixed (table 5).

34The two cluster-randomised trials comparing effects of active case-finding on tuberculosis prevalence and tuberculosis infection incidence (ZAMSTAR and ACT3), both had some concerns of bias relating to participation in endline tuberculosis prevalence surveys and completeness of outcome sputum evaluation (appendix p 14). The risk of bias for DETECTB (before-after comparison) was assessed to be serious; the six other non-randomised studies had a critical risk of bias.

35Community-based active case-finding programmes for tuberculosis are some of the most widely implemented and longest-running screening interventions ever delivered. However, their effect on tuberculosis epidemiology remains uncertain. In this systematic review, we aimed to synthesise evidence from evaluations of community-based tuberculosis active case-finding interventions to determine whether active case-finding affects tuberculosis epidemiology in communities. The review included 36 studies from 16 countries, comprising at least 110 million person years of follow-up in studies done between 1980 and 2020. Our main findings were that there is mixed evidence that active case-finding is effective at initially increasing tuberculosis detection when measured by case notification rates, and that active case-finding could reduce community prevalence of tuberculosis if delivered with sufficient intensity and coverage.

36Active case-finding interventions aim to screen, diagnose, and link to treatment people who have asymptomatic or symptomatic tuberculosis disease and who have, for whatever reason, not been diagnosed through facility-based services. Of note, a single round of active case-finding, no matter how well implemented, will not have a lasting epidemiological effect. If active case-finding is implemented with sufficient intensity and over a sufficiently long period or in repeated rounds, we anticipate that the community tuberculosis transmission would be reduced. The intensity of interventions will depend on how many people in the target population are reached, how often people are reached and what diagnostic algorithm is used (eg, who is eligible for sputum-based tests). Although a rapid effect on undiagnosed tuberculosis disease prevalence is possible, subsequent epidemiological effects might accumulate over several years. In the absence of a test of recent infection that could be used to directly measure the effect of active case-finding on tuberculosis transmission, the effectiveness of active case-finding interventions must be measured through indicators such as case notification rates, tuberculosis disease prevalence, and through measures of community transmission, including tuberculin skin test and interferon γ release assay surveys among children of preschool age and schoolchildren. Analysis of the percentage of cases that are clustered through genomic data holds promise as a measure of changing community tuberculosis epidemiology, but it relies on high coverage of tuberculosis culture positivity and has not been widely used to date.

37Summarising data for the effectiveness of active case-finding on tuberculosis case notification rates, we found that there is inconsistent evidence from a small number of high-quality studies to suggest that community-based tuberculosis screening delivered from active case-finding interventions might initially increase tuberculosis case notification rates. In four randomised controlled trials that compared an active case-finding intervention to a non-active case-finding comparison, two showed non-statistically significant initial increases in tuberculosis case notifications (in urban Brazil and rural Ethiopia), and two showed an increase that reached statistical significance (in rural Ethiopia and prisons in Ethiopia). In a further 22 non-randomised studies with a wide range of designs and interventions assessed, data with low quality of evidence suggested that community-based active case-finding might increase case notification rates. The wide range of study designs and interventions evaluated, limited reporting of data within many studies, and the high percentage of studies classified as being at serious or critical risk of bias meant that only cautious conclusions should be drawn from these studies. Furthermore, we do not have information on the costs or opportunity costs of active case-finding compared to other approaches that could be undertaken to detect tuberculosis.

38We identified two cluster-randomised trials that had varying results on the effects of active case-finding on prevalence of tuberculosis disease and incidence of infection in children. The more intensively delivered door-to-door active case-finding intervention of ACT3 in Vietnam, which used a screening strategy comprising Xpert for all, regardless of symptoms, reported a statistically significant relative reduction in the prevalence of microbiologically confirmed tuberculosis of 45%. By contrast, the less intensive enhanced case-finding intervention in the ZAMSTAR trial in Zambia and South Africa, which used a symptom-based and sputum smear-based screening approach, did not show an effect. The before-after evaluation that pooled data from both intervention groups of the DETECTB trial in Zimbabwe, in which active case-finding was delivered through moderate intensity interventions (mobile vans and door-to-door symptom-based and smear-based screening), showed a relative reduction in culture-confirmed tuberculosis of 41%. Other non-randomised studies had inconsistent and imprecise results, and they were at critical risk of bias due to confounding by secular trends and selection of participants for inclusion and measurement of effectiveness. Evidence for reduced tuberculosis transmission was lacking, with two studies (ZAMSTAR and ACT3) reporting no significant difference in childhood tuberculosis infection (according to prespecified analyses in each study).

39The effects of active case-finding for tuberculosis are likely to be highly context-dependent, varying with tuberculosis prevalence, built environment, access to health care, and social norms, among other factors. There are many possible reasons why ZAMSTAR and ACT3, which were done nearly 10 years apart and in different continents, showed differing results. ZAMSTAR used a less intensive case-finding approach with the aim of enabling community members to identify tuberculosis symptoms themselves and improving access to sputum diagnostics for tuberculosis. By contrast, ACT3 used more intensive screening, involving enumeration of community members and door-to-door tracing of all community members to request sputum, regardless of symptoms. Whether the reduction in tuberculosis prevalence in ACT3 (which was not seen in ZAMSTAR) was due to the more intensive nature of screening in ACT3 or due to other context-specific factors is not known.

40None of ZAMSTAR, ACT3, or DETECTB report directly on harms related to tuberculosis screening. In ACT3, the estimated positive predictive value for a positive Xpert result to detect a true case of tuberculosis disease in the context of community-wide screening was between 61% and 84%, depending on the reference standard that was applied. It is not known whether any individuals experienced harm (such as anxiety, unnecessary further investigations, or unnecessary tuberculosis treatment) as a result of false positive Xpert tests. We would expect that an intervention in which people identify their own symptoms and sputum diagnostics are readily and easily available to these people, such as that used in ZAMSTAR or DETECTB, would be less likely to cause individual harm from false positive results than an approach in which all individuals have sputum tests, such as in ACT3, because presumably the pre-test probability of tuberculosis is higher in those who choose to submit sputum than the rest of the community; however, no data are available that directly address this hypothesis. The resource implications in terms of cost and laboratory capacity are likely to be higher for the approach used in ACT3 compared with that used in ZAMSTAR, although in practice sputum submission during ACT3 was substantially below the universal target. Lastly, it is important to explore population values and preferences around acceptability of various community-based tuberculosis screening approaches, acknowledging that this is likely to vary substantially between communities and countries.

41This systematic review had several limitations. We included only manuscripts published in English. We reviewed the full text of 988 published manuscripts drawn from more than 25 000 titles and abstracts, but we did not include unpublished data or grey literature. Publication bias is possible; we are aware of several active case-findings evaluations which are not published (eg, from TB REACH-funded projects). Studies generally did not distinguish between the number of people diagnosed with tuberculosis and the number started on tuberculosis treatment (ie, they did not account for pretreatment loss to follow-up). We did not assess individual-level effects of active case-finding, such as whether people with tuberculosis detected through active case-finding had less extensive or severe disease or better outcomes than those with tuberculosis detected through usual care-seeking.

42We recognise that community-based studies that set out to evaluate active case-finding interventions are expensive, logistically challenging, and require very large sample sizes and long follow-up periods, as well as careful analysis to minimise bias and allow valid inference to be drawn. Given these challenges, we strongly recommend that future evaluations of the impact of active case-finding on tuberculosis case notification rates (which provide an important source of evidence under programmatic conditions) are carefully designed to minimise selection and ascertainment bias, have prespecified protocols and analysis plans, and undertake appropriate statistical analysis to adjust for confounding and the effects of temporal trends with effect estimates and measures of uncertainty appropriately adjusted for clustering.

43Tuberculosis active case-finding interventions are necessarily highly context-dependent. Different methods of delivering tuberculosis active case-finding and different diagnostic algorithms (eg, initial screening using symptom interview vs using chest x-ray) might be used in different settings, depending on factors such as resources available, physical geography, health systems capacity, expected prevalence of tuberculosis (ie, pre-test probability of tuberculosis), prevalence of drug resistant tuberculosis, prevalence of HIV, and laboratory infrastructure and capacity. In areas with high HIV prevalence, Xpert MTB/Rif might be a more appropriate diagnostic test than sputum smear, and false negatives from symptom screening might be expected to be more common. Future studies should describe their context and intervention in as much detail as possible and fully report all numerators and denominators for total population targeted, number of individuals screened, number requiring a diagnostic test, number receiving a diagnostic test, number testing positive, and number starting treatment. When appropriate, false positive results should also be reported.

44In conclusion, we found evidence to suggest that community-based active case-finding for tuberculosis might be effective in changing tuberculosis epidemiology if delivered with high coverage and intensity. The evidence for effectiveness in other settings and using alternative tuberculosis screening approaches was mixed. Policy makers should consider implementing intensive active case-finding interventions in urban populations with a high prevalence of undiagnosed tuberculosis, and in other populations as part of well designed research protocols to contribute evidence to important knowledge gaps.

45All data are included within the Article and supplementary appendix.

46JEG, HA, and ELC are authors of trials included in this systematic review. HA and ELC are members of the WHO TB Screening Guideline Development Group. JEG, HA, ELC, and PM have received research grants to their institutions for projects evaluating community-based active case-finding. All other authors declare no competing interests.