Effects of distances on newborn outcomes in Brazil

Abstract

Travel distances can affect the delivery outcomes of pregnant women. Using detailed national data from the public health system (SUS), this paper identifies the effect of the distance to the place of delivery on infant mortality between 2006 and 2017 in Brazil. We focus on SUS cases to reduce demand bias because preferences and constraints to join the private system might differ. The paper follows a two-part empirical strategy. First, we focus on a sample of low-risk pregnancies, among which traveling for childbirth would most likely reflect differences in geographic accessibility. Using a linear regression model, we find that moving to another municipality for childbirth increases the infant mortality rate by 17%, and each additional ten kilometers of journey raises the infant mortality rate by 0.1 points per 1000 live births. Results are robust to introducing multiple fixed effects, socioeconomic and risk factors. Second, we use a sample of highrisk pregnancies and a self-made complexity-of-care classification to identify the effect of the distance to specialized medical care on infant mortality. For this sample, the distance to any facility is less relevant than that to facilities with specialized medical technology: living one standard deviation farther away from municipalities with level- II and level-III technologies increases the infant mortality rate by 0.94 and 0.4 points, respectively. Furthermore, the geographic availability of neonatal intermediate and critical care beds is relevant for explaining the effect of the distance to specialized medical care on infant mortality.