[Infant mortality from preventable causes in Brazil: an ecological study in 2000-2002].

This study aims to test the associations between infant mortality from preventable causes in Brazil and socioeconomic factors, including those pertaining to health services and investments. This was an ecological study using 296 Brazilian counties (municipalities) with more than 80,000 inhabitants each as the analytical units. Kruskall-Wallis and ANOVA tests were performed to compare independent variables according to infant mortality quartile, and Pearson and Spearman's correlation coefficients were computed to test the associations. As the infant mortality quartile from preventable causes increases, there is a gradual decrease in the municipal human development index, per capita gross domestic product, households with bathrooms and indoor plumbing, total health expenditures per inhabitant, and physicians per 1,000 inhabitants, and an increase in the Gini coefficient. Improved socioeconomic conditions and public health investments are strongly associated with reduction of infant mortality from preventable causes. This knowledge should permeate actions aimed at minimizing the number and unequal distribution of such deaths.