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

BACKGROUND: In Tiruvallur District, South India, tuberculosis cases are detected at health facilities (HF) as part of a DOTS programme, and by screening adults through community survey (CS) as part of ongoing epidemiological research. OBJECTIVE: To compare socio-demographic, clinical and bacteriological characteristics and treatment outcomes of all patients detected at HF with those of all patients detected by CS during a 12-month period. RESULTS: Of 32,663 adults surveyed, 100 had smear-positive and 116 had smear-negative tuberculosis; of 65 smear-positive patients who began treatment, 44 were cured. Compared to HF patients, CS patients were significantly more likely to be older (AOR = 1.9), male (AOR = 2.7), non-literate (AOR = 1.7), and living in poor quality housing (AOR = 2.0), and were less likely to have cough >3 weeks (AOR = 3.4) or smear-positive tuberculosis (AOR = 4.2). Of 61 new smear-positive CS patients, 40 reported chest symptoms; of these, 32 (80%) had already consulted a health-care provider, but remained undiagnosed. CONCLUSIONS: The community survey was of little value in tuberculosis case detection even in this high-prevalence setting. Patients identified by the survey were less symptomatic and less infectious, and less than half were cured. Diagnostic services should be made more accessible to the elderly, the non-literate and men.