An application of Bayesian spatial statistical methods to the study of racial and poverty segregation and infant mortality rates in the US

摘要

The infant mortality rate is a fundamental measure of population health used internationally. In the United States, the infant mortality rate is higher than what would be expected for a country of its affluence. We present an analysis of US county infant mortality rates using modern Bayesian spatial statistical methodologies. Our key predictors in our statistical analysis are residential racial and poverty segregation, measured by the dissimilarity, interaction and spatial proximity indexes. We use both Exploratory Spatial Data Analysis methods and Hierarchical Bayesian spatial regression models to examine the influences of these segregation measures on the infant mortality rate for each county, net of income inequality, degree of rurality and relative socioeconomic deprivation. The spatial measures of racial segregation suggest that when blacks live in close proximity to each other, this tends to increase the infant mortality rate. The results for poverty segregation suggest the same pattern, when poor populations live in close proximity to one another this is generally detrimental to the county infant mortality rate. However, interaction between blacks and whites and poor and non-poor residents of an area is protective for infant mortality.