DOTS in primary care units in the city of Rio de Janeiro, Southeastern Brazil.

OBJECTIVE To describe the implantation and the effects of directly-observed treatment short course (DOTS) in primary health care units. METHODS Interviews were held with the staff of nine municipal health care units (MHU) that provided DOTS in Rio de Janeiro City, Southeastern Brazil, in 2004-2005. A dataset with records of all tuberculosis treatments beginning in 2004 in all municipal health care units was collected. Bivariate analyses and a multinomial model were applied to identify associations between treatment outcomes and demographic and treatment process variables, including being in DOTS or self-administered therapy (SAT). RESULTS From 4,598 tuberculosis cases treated in public health units administrated by the municipality, 1,118 (24.3%) were with DOTS and 3,480 (75.7%) with SAT. The odds of DOTS were higher among patients with age under 50 years, tuberculosis relapse and prior history of default or treatment failure. The odds of death were 52.0% higher among patients on DOTS as compared to SAT. DOTS modality including community health workers (CHWs) showed the highest treatment success rate. A reduction of 21.0% was observed in the odds of default (vs. cure) among patients on DOTS as compared to patients on SAT, and a reduction of 64.0% among patients on DOTS with CHWs as compared to those without CHWs. CONCLUSIONS Patients with a "low compliance profile" were more likely to be included in DOTS. This strategy improves the quality of care provided to tuberculosis patients, although the proposed goals were not achieved.