OBJECTIVE: To propose a tuberculosis-related death surveillance strategy based on the Brazilian Mortality Information System. METHODS: Data on 55 tuberculosis-related deaths, which occurred in two large hospitals in Rio de Janeiro, Southeastern Brazil, between September 2005 and August 2006, were obtained from the SIM. These cases were searched and compared with cases in the National Notification System (Sinan). The increment in the number of notifications and completeness of data were evaluated, as well as entry type and outcome in Sinan. RESULTS: Of the 55 deaths, 28 were registered in Sinan. Comparison between systems allowed for the following corrections: 27 new cases were notified, 14 new notifications performed by the hospitals where death occurred and ten outcomes corrected. This represented an increment of 41/144 (28%) notifications by these two hospitals in 2006. Nine cases, previously classified as unconfirmed tuberculosis were reclassified as bacteriologically confirmed, and another five cases were reclassified from tuberculosis to AIDS as the primary cause of death. CONCLUSIONS: The proposed surveillance system for tuberculosis-related death was useful to increase data completeness, decrease under-notification and cases with unknown outcome, to evaluate epidemiological surveillance and death certificate quality and to trace previously unidentified contacts.
Tuberculosis; Disease Notification; Hospital Mortality; Mortality Registries; Information Systems; Epidemiologic Surveillance
