Uso de regressão logística para identificar os fatores de risco associados à ocorrência de anomalias congênitas em recém-nascidos

Abstract

The logistic regression models have been extensively applied in several areas of knowledge, particularly in the area of health. The condition of studying binary response variables in terms of a set of explanatory factors have become increasingly common in epidemiological studies. Thus, this study aims to use a logistic regression model to investigate the risk factors associated with the occurrence of congenital malformation in children from a hospital in João Pessoa - PB. The results showed that age, mother's education, the use of corticosteroids during pregnancy, type of childbirth and APGAR measures 1 and 5 minutes were associated with the probability of birth of children with congenital abnormality. Furthermore, it was found that the adjusted model could correctly classify over 93% of the cases examined.