Drop-out from the tuberculosis contact investigation cascade in a routine public health setting in urban Uganda: A prospective, multi-center study

SETTING Seven public tuberculosis (TB) units in Kampala, Uganda, where Uganda's national TB program recently introduced household contact investigation, as recommended by 2012 guidelines from WHO. OBJECTIVE To apply a cascade analysis to implementation of household contact investigation in a programmatic setting. DESIGN Prospective, multi-center observational study. METHODS We constructed a cascade for household contact investigation to describe the proportions of: 1) index patient households recruited; 2) index patient households visited; 3) contacts screened for TB; and 4) contacts completing evaluation for, and diagnosed with, active TB. RESULTS 338 (33%) of 1022 consecutive index TB patients were eligible for contact investigation. Lay health workers scheduled home visits for 207 (61%) index patients and completed 104 (50%). Among 287 eligible contacts, they screened 256 (89%) for symptoms or risk factors for TB. 131 (51%) had an indication for further TB evaluation. These included 59 (45%) with symptoms alone, 58 (44%) children <5, and 14 (11%) with HIV. Among 131 contacts found to be symptomatic or at risk, 26 (20%) contacts completed evaluation, including five (19%) diagnosed with and treated for active TB, for an overall yield of 1.7%. The cumulative conditional probability of completing the entire cascade was 5%. CONCLUSION Major opportunities exist for improving the effectiveness and yield of TB contact investigation by increasing the proportion of index households completing screening visits by lay health workers and the proportion of at-risk contacts completing TB evaluation.