Objective: To analyze the treatment strategies that influence the effectiveness of tuberculosis treatment at primary care clinics (PCCs) in Brazil. Methods: This was a descriptive, retrospective epidemiological survey based on the medical records of 588 tuberculosis patients enrolled in the tuberculosis control programs at two PCCs located in the city of Belém, Brazil: Centro de Saúde Escola do Marco (CSEM) and Unidade Básica de Saúde da Pedreira (UBSP). The survey was limited to patients enrolled between January of 2004 and December of 2008. We included only patients between 18 and 59 years of age, and we excluded those who were transferred or were found to have been misdiagnosed. We collected data regarding age, gender, type of treatment (self-administered or supervised), co-infection with HIV, and treatment outcome. The health professionals involved in the tuberculosis control program at the two PCCs were interviewed regarding the strategies used for tuberculosis control and regarding routine clinical care for tuberculosis patients. Results: There were no significant differences between the CSEM and UBSP patients regarding age, gender, or co-infection with HIV. Supervised treatment was used significantly more frequently and the rate of cure was higher at the CSEM than at the UBSP, whereas the rate of treatment noncompliance was higher at the UBSP than at the CSEM. Conclusions: For patients enrolled in tuberculosis control programs at PCCs in Brazil, supervised treatment appears to be an extremely important strategy for reducing the rate of treatment noncompliance.

Keywords: Tuberculosis; Patient dropouts; Patient care planning; Treatment outcome.