Letícia Xander Russo

This thesis consists of three essays about primary care physicians and child health in Brazil. It focuses on the impact of an increase in the number of primary care physicians (PCPs) on improvements in children’s health, as well as the geographical distribution of PCPs in Brazil. The first essay performs an exploratory analysis of the temporal and spatial distribution of primary care physicians in 2005 and 2012. The essay provides a descriptive analysis of the data and uses the Exploratory Spatial Data Analysis (ESDA). The shortage of PCPs is noticed mainly in North and Northeast; these are also the regions with lower GDP per capita, higher percentage of illiteracy, higher infant mortality rate and lower percentage of household facilities. Even after the recent increased in the number of PCPs, the univariate and bivariate analyses show that the geographical maldistribution of PCPs is still persistent. The second essay investigates the effect of primary care physicians on infant mortality rates (IMR) in Brazilian municipalities between 2005 and 2012. A balanced panel of 5,563 municipalities in Brazil over a period of 8 years is estimated. In addition, the study performs sensitivity analyses for alternative samples to address a possible bias due the quality of information on births and deaths as denoted by Frias et al. (2014). The empirical analysis uses the System Generalized Method of Moments (GMM) estimator developed by Arellano and Bover (1995) and Blundell and Bond (1998) for dynamic models. This method accounts for endogeneity, unobserved municipality-specific effects and it improves the accuracy of estimates when the dependent variable is persistent over time. The findings demonstrate the effect of primary care physician supply on reducing infant mortality, even after including time trends and a range of important determinants of child health. The third essay aims to analyze the effects of PCPs on hospitalizations for ambulatory care sensitive conditions (ACSHs) among children under five years old. ACSHs account for more than half of all hospital admissions among children under five years old in Brazil. Spatial dynamic panel data model is estimated using municipality data over a period of eight years from 2005 to 2012. The regression includes the spatial lag of hospital beds and PCPs, which allow spatial interactions between neighboring units. The regression also includes a set of variables known to be associated with child health, as well as year dummies. The results indicate that primary care physician supply contributed to the decline of ACSHs among children. Spillover effects associated with hospital beds and number of PCPs are also relevant in explaining ACSHs among children; these show that there are regional benefits beyond the municipality of observation.