[Mortality due to ill-defined causes in Brazil (1979-2002) and a predictive model for age].

This study focused on the percentage of deaths due to ill-defined causes in Brazil, from 1979 to 2002. The objectives were to: (a) describe changes over the 24-year period; (b) identify the age group in which the percentage of ill-defined causes correlated most closely with the total percentage of ill-defined causes; (c) select a predictive model for the percentage of ill-defined causes in such age group, given the total percentage of ill-defined causes; (d) describe changes in the distributions of each age group in the total ill-defined causes; and (e) verify whether the percentage of deaths in-hospital shows any relationship to the percentage of ill-defined causes. Results showed that the total percentage of deaths from ill-defined causes decreased in Brazil. The percentage of ill-defined causes in the > 50-year age group correlated most closely with the total percentage of ill-defined causes, and cubic regression was the most appropriate predictive model. Age > 50 showed the highest increase in its share of total ill-defined causes from 1979 to 2002. The percentage of in-hospital deaths showed an inverse relationship with the percentage of deaths from ill-defined causes.