OBJECTIVE: To describe the infant mortality coefficient and its components in the municipality of Embu, São Paulo, Brazil, between 1995 and 1998, according to strata of living conditions. METHODS: A cross-sectional descriptive study was conducted using cluster analysis in the 135 census session of the municipality of Embu, grouped into four strata of living conditions: stratum 1, with the best living, income and schooling conditions; strata 2 and 3 with regular conditions; and stratum 4, in which all houses were subnormal agglomerates or slums. The neonatal, post-neonatal and infant mortality rates, the population's ascribable risk, the ratio between neonatal and post-neonatal deaths and the mortality due to ascribable causes were calculated for 1995 to 1998, for each stratum. RESULTS: In all studied years, the stratum 4 presented the highest infant mortality rates and higher population's ascribable risk compared to intermediate strata. This stratum has also presented the lowest ratio between neonatal and post-neonatal mortality. The population's attributable risk in the stratum 4 was higher for perinatal affections (159.4), respiratory (271.4) and infectious (415.6) diseases. Similar demographic data in areas close to the geographically limits of the study and heterogeneity of events in the same territory were present. CONCLUSIONS: Relationship between social inequality and infant mortality was identified according to the living conditions criteria. However, heterogeneous distribution was shown within each stratum, precluding its use for evaluating social inequalities in big urban centers.
infant mortality; health inequalities; ecological studies; health indicators; social indicators
