Examining birth outcomes hotspots in Maryland

Background: Research suggests spatial autocorrelation of birth outcomes exists at the Census tract level in Maryland, yet identification and analysis of birth outcomes hotspots has not occurred.

Methods: Tract-level LBW and PTB rates were drawn from 2010-2012 aggregate vital statistics data in Maryland, and demographic data were drawn from 2008-2012 American Community Survey. Global Moran’s I (GMI) is a global measure of spatial autocorrelation of LBW and PTB rates. Local Moran’s I (LMI) identified hotspots. T-tests examined demographic differences between “High” and “Low” hotspots.

Results: GMI suggests spatial autocorrelation for LBW (z-score: .291, p=.001) and PTB (z-score:.173, p=.001). LMI identified 127 LBW hotspots, with 77 percent (99 of 127) being “High LBW” (≥ 11 percent LBW). There are 88 PTB hotspots, with 68 percent (60 of 88) being “High PTB” (≥ 13 percent PTB). High LBW hotspots had higher mean poverty (23.39 percent) and mean concentration of Black non-Hispanics (81.75 percent) than Low LBW hotspots (p<.001). High PTB hotspots had higher mean poverty (24.74 percent) and mean concentration of Black non-Hispanics (85.69 percent) than Low PTB hotspots (p<.001).

Discussion: This study provides evidence that geographic concentration of birth outcomes does occur, and allows for the identification of areas with poor outcomes. Results suggest hotspots with poor outcomes have high rates of poverty, and are minority-majority areas. This can result in programs and policies to improve birth outcomes that are both geographically targeted, and culturally appropriate. Future research should include individual- and area-level factors to better understand clustering of birth outcomes.