Impact and implementation of active case finding for tuberculosis in homeless populations: a systematic review

BackgroundIdentifying tuberculosis in homeless populations through active case finding (ACF) is recommended to address health inequalities and contribute to wider control strategies for tuberculosis. We aimed to assess the effectiveness of ACF.MethodsThis systematic review assessed studies on ACF done in countries with low or medium burden of tuberculosis across Europe, the USA, and Australia. We systematically searched EMBASE, CINAHL Plus, ASSIA, Pro-Quest, Scopus, and the Cochrane Library and grey literature for English language publications up to Jan 5, 2019 (no earlier date limit). We used concepts of “ACF”, “tuberculosis”, and “homeless person”. We identified studies that analysed ACF and reported on our outcome measures, in homeless populations, in low-burden and medium-burden countries. ACF screening included testing for latent tuberculosis infection (LTBI) or active tuberculosis affecting any site. Studies into outbreak control or other populations were excluded. Primary study outcomes were the effectiveness of ACF (using population measures of tuberculosis prevalence or incidence) and interventions to improve ACF uptake and completion of the diagnostic pathway. Secondary outcomes were yield of ACF, cost-effectiveness, and characteristics of participants.Findings21 studies met the inclusion criteria. Study heterogeneity precluded meta-analysis. Three time-trend analyses produced some evidence that ACF was effective, because it was associated with reductions in tuberculosis incidence, prevalence, or clustering. A modelling study also showed that ACF was more effective than passive case finding in reducing population tuberculosis burden. Material incentives have the strongest evidence for improving uptake of ACF, with mixed evidence for peer educators. Observational evidence shows professional support and mandatory screening might also enhance uptake, and additional community-based support improves completion of the diagnostic pathway. Across all studies, the yield of screening (defined as the proportion of screened individuals who test positive) ranged from 1·5% to 57% for LTBI (total 41 684 individuals screened), and 0–3·1% for active tuberculosis (total 91 771 individuals screened). ACF can be cost-effective; population prevalence and screening modalities are determinants of cost-effectiveness. Considering ACF participants, subgroups most likely to be diagnosed with tuberculosis appeared less likely to accept screening.InterpretationACF should be considered in both tuberculosis and homelessness strategies, with evidence-based interventions to improve implementation. Outcomes varied widely, meaning programmes must be tailored to local populations. Strengths of our study include generalisable results to homeless populations from diverse settings. Limitations include restriction to the English language, the fairly low grade of the evidence identified, and the low number of studies screening for LTBI or using newer screening tests.FundingThe South West Public Health Training Programme.