OBJECTIVE: To describe infant mortality in the State of São Paulo, in 1999, based on multiple causes of death and to compare data from underlying and multiple causes of death. METHODS: Data came from 12,793 infant death records in 1999, of Seade Foundation (State Data Analysis System Foundation). Causes of death were coded according to the Tenth Revision of the International Statistical Classification of Diseases and Related Problems and were classified into 28 meaningful groups for the purpose of this article. In order to analyze multiple causes of death, simple frequencies were used in addition to principal components analysis to obtain the main groups of causes that contributed to death. RESULTS: The most frequent multiple causes of death were respiratory and cardiovascular diseases of the perinatal period (24.2% of all multiple causes), diseases related to growth and maturity of the fetus and the newborn (20.2%), congenital malformations (8.6%) and perinatal infections (7.6%). Principal components analysis revealed three major interpretable components: "post-neonatal, infectious and avoidable deaths", "complications of procedures and external causes" and "perinatal disorders, not related to low birth weight and/or immaturity". CONCLUSION: By using principal components analysis it was easier to understand how the multiple causes were associated. This is more interesting for Public Health purposes, because it may help clarify the steps in disease causation.
Infant mortality; Multiple causes of death; Principal components analysis
