| The main goal of this paper was to demonstrate area effect on infant health outcomes by developing and estimating two spatial models that can be used to determine risk areas in Allegheny County, Pennsylvania, and to show how space is important in determining both the individual and community health outcomes. Spatial data analysis was performed to look for spatial effects and spatial autocorrelation between the explanatory variables and the dependent variable. These models incorporate both the community level (i.e., income, unemployment, poverty level, housing value, and age of housing) and individual level (i.e., education level, alcohol consumption, and smoking) factors that are considered as determinants of adverse infant health outcomes.; Secondary data analysis was performed, which incorporated both spatial and non-spatial data. GIS was used to capture, store and manipulate both spatial (geo-referenced) data obtained from digitized maps, and non-spatial (attribute) data obtained from birth files and socioeconomic data from 1990 census.; The independent variables included, primarily, unemployment, percent below poverty level, housing median value, education level, alcohol consumption while pregnant, and smoking cigarettes while pregnant by census tracts for the first model, and by zipcode for the second model. The dependent variables included low birthweight by census tracts, and blood lead level by zipcodes within Allegheny County.; The findings of this study support the hypothesis that the characteristics of area of residence during pregnancy have an effect on an expectant mother's risk of delivering a low birthweight infant. Similarly, communities with higher rates of poverty level, lower values of education level, and higher values of age of housing are at risk of having children with elevated blood lead level >= 10 mμg/dL.; Being able to spatially identify areas at high risk enable local policy makers to provide the right interventions at the right place, determine where certain interventions are mostly needed, identify where gaps exists, and identify where services overlap. While past research on differences in infant birthweight has largely been limited to individual-level analysis, the results presented here illustrate the need to include contextual characteristics. |
