The Spatial Patterns of Low Birth Weight and Infant Mortality at the County-Level: Using Spatial Analytical Techniques

Seung-Eun Song, University of Texas at Austin

In this paper, we use spatial analytical techniques to visualize spatial variation in low birth weight rates (LBWRs) and infant mortality rates (IMRs) at the county level and provide improved understanding of the underlying mechanisms of differentiating outcomes of interest across areas. We utilize the 2002 Area Resource File. The result shows a clear spatial pattern of LBWRs and IMRs. The spatial autocorrelation is much higher for LBWRs than IMRs. The concentration of high LBWRs is observed in the southeastern and middle northeastern regions of the U.S., while southeastern regions have a high level of IMRs. The spatial regression model shows that socioeconomic aspects of areas play an important role in determining the county level of both the LBWRs and IMRs. Physical features (e.g., humidity, temperature and elevation) greatly influence the county LBWRs, while health care relevant resources (e.g., states' Medicaid eligibility, per capita physician supply) have substantial influences on the county IMRs.