The study of tuberculosis-attributed deaths as a tool for disease control planning in Rio de Janeiro, Brazil.

Setting: Two tuberculosis (TB) reference hospitals and three general hospitals in Rio de Janeiro (RJ).
Objective: To analyze TB-attributed deaths as a tool for evaluating the TB control program in RJ.
Design: Retrospective study based on 302 medical records selected from the 1998 death database.
Results: Of 1146 registered adult (>14 years) TB-attributed deaths in RJ, 328 occurred in five hospitals, and 302 records were analyzed. Median age was 47.5 (17-89) years; 237 (78.5%) were male. Median time elapsed from onset of symptoms until diagnosis was 60 (7-730) days; median hospitalization was 60 (0-517) days. Acid-fast bacilli sputum smears were performed in 200 (69%) of 290 cases of pulmonary disease. Among 32 (36%) smear-negative patients, culture was done in only one. The recommended regimen (RHZ) was used in 175 (58%). Among 125 re-treatment patients, 55 (44%) were on RHZ instead of RHZE. Notification to health authorities was recorded in 131 (43.4%) cases.
Conclusion: In RJ, young people die from TB. Major issues identified in the public health system were poor detection and notification and a high default rate, perpetuating the spread of TB. Treating professionals do not follow guidelines, and political commitment is needed to ensure TB control in the state and in the country.