A natureza da relação entre escolaridade materna e mortalidade infantil e na infância no Brasil

Abstract

This dissertation investigates the nature of the relationship between maternal education and infant mortality (early neonatal mortality, neonatal mortality, post-neonatal mortality) and child mortality in Brazil using data from the Demographic and Health Survey (DHS). Maternal education is an important socioeconomic determinant of mortality during the childhood, but few studies highlight the nature of this relationship. Based on the socioeconomic gradient and threshold model, this work defined two hypothesis related to the relationship between maternal education and infant and child mortality: graduated returns to maternal education on child survival, and existence of one or more thresholds along the single years of schooling maternals curve. The study was performed using three complementary analyses. First, I used Poisson regression models to study the average effect of additional maternals education year on the infant and child mortality prevalence, controlled by selected associated factors (maternal and child demographic characteristics, maternal reproductive history, prenatal care, maternal exposition to information, household characteristics and marriage market). The results of the Analysis 1 showed null effects of maternal education on child survival for all samples and refused the hypothesis related to socioeconomic gradient. Second, I use Poisson regression models, but maternal education categories were used in order to verify different effects of maternal schooling levels on child survival. In general, the results of Analysis 2 showed protector maternal educations effects up to high levels of schooling (college) and null effects before, mainly in 2006. Finally, I applied spline regression models to verify whether there were threshold and how many there were along the curve that describe the relationship between maternal education and child survival. Results of Analysis 3 showed trend of change to linear from threshold structure of the relationship between maternal education and infant and child survival. It was verified high thresholds (ten and eleven years of maternal schooling) in contrast of few years indicated for previous studies. In brief, the results indicated that differences in the nature of the relationship between maternal education and infant mortality components and child mortality can be the result of change in mortality structure of causes, and the introduction of new medicine technologies in the field of child heath and the pattern of social inequalities in the access of these resources.