OBJECTIVES: to assess the quality of information on deaths of infants aged less than one year registered in the Basic Care Information System (SIAB). METHODS: a quality assessment of SIAB data in the municipality of Recife. Changes in the SIAB data on mortality were compared over the period between 2000 and 2006, using municipal and federal databases available on the DATASUS site. For 2006, deaths of infants aged under one year registered in the municipal SIAB were given individual details and semi-structured interviews with community health officers (ACS) were used to confirm or reject registered deaths, by identifying errors a weaknesses in the production of SIAB data. RESULTS: there are differences between the federal and municipal SIAB databases. Only 34.0% of deaths registered by ACS were confirmed. The errors that led to over-reporting of deaths in the SIAB included errors relating to the definition of the event (4.6%), the filling in of forms (33.8%), consolidation (7.3%) and typing of data (21.4%). CONCLUSIONS: although there is evidence of the potential of the SIAB to monitor and evaluate primary care, there are still weaknesses in the information production process and efforts need to be made to ensure more reliable data are collected.
Evaluation; Death; Child; Information systems; Primary health care
