Scanty smears associated with initial loss to follow-up in South African tuberculosis patients.

Background: Smear-positive patients should be started on anti-tuberculosis treatment promptly. However, studies show that up to 38% of diagnosed patients are initial loss to follow-up (LTFU), meaning they do not start treatment after diagnosis. We investigated determinants of initial LTFU at primary health care facilities.
Design: In a facility-matched case-control study, health care facilities were visited from October 2010 to September 2012. After identification from registers, patients were traced and invited to complete a questionnaire.
Results: Of 973 participants, 233 (24%) were cases and 740 (74%) controls. Initial LTFU was associated with smear grade (pooled adjusted odds ratio [aOR] 0.73, 95% confidence interval [CI] 0.64-0.90, scanty at baseline) for participants identified at facilities, but not with age (overall P = 0.80) or sex (aOR 0.83, 95%CI 0.58-1.20). Of the 233 cases, 197 (85%) were traced in the community, of whom 58 (29%) were found. Among the group found, initial LTFU was associated with age (aOR 3.38, 95%CI 1.15-9.95) and smear grade (aOR 0.08, 95%CI 0.02-0.34, scanty at baseline).
Conclusion: Scanty smear positivity was associated with initial LTFU. Tuberculosis programmes should start scanty smear-positive patients on treatment early and develop alternative community tracing strategies. Health care worker training could address the first aspect, and the use of technology to improve treatment initiation, such as mobile phone applications, the second.