Estudo da mortalidade infantil: um modelo de regressão logística múltipla

Abstract

The present study had the purpose of analyzing the data of the System of Information on Born Alive (SINASC) and of the System of Information on Mortality (SIM) in order to identify the risk factors for the infant mortality of born alive from mothers living in the municipal districts belonging to the 4th Local Coordination Department in 2006. A longitudinal study has been developed for the analysis of the respective data by means of a multiple logistic regression model. The Infant Mortality Coefficient (CMI) which expresses the risk of a born alive dying before to complete one year old is considered as one of the most efficient indicatives of the quality of health assistance, as well as of the socio-economical level of a population. In the present study, the dependent variable (outcome) corresponds to the occurrence or not of death in babies with less than one year old. The results of the model adjustment show that the significant risk predicting factors (p ≤ 0,05) for infant mortality include the variables: marital status (mother without stable engagement OR= 2,0) gestational age (premature: OR= 3,3) birth weight (low weight: OR= 4,7), Apgar index in the first minute (≤ 8 OR= 3,9), Apgar index in the fifth minute (≤ 8: OR = 5,6), and congenital anomalies (with anomaly: OR= 20,4). These results are in agreement with most of the results found in the literature about infant mortality in Brazil. This study allowed using birth and death database, suggesting a better surveillance and planning of the public health political, mainly regarding the attention and cares with pregnant women, aiming at the early diagnosis of the possible fetal and maternal diseases in order to provide a better quality of life for the baby.