| Geographic information systems (GIS) are evolving into a viable public health research tool. Increases in computer speed, display, and memory, combined with developments in epidemiologic methods, are furthering this cause. Epidemiology has a long history of studying the spatial dimensions of disease transmission, rates of occurrence, and resulting morbidity or mortality. Ecologic analyses have an obvious relation to geography because it is the link between exposure and health. Many of the first GIS-based epidemiologic studies published in the early 1990s were ecologic investigations. However, this study design is clearly weaker compared to individual-level analyses. The focus of this dissertation is the application of GIS in case-control epidemiologic investigations. The studies of asthma prevalence and cancer incidence demonstrate approaches for linking individual-level measures of exposure and health outcome by using spatial relationships. Along with these 2 case-control investigations, there is an examination of the uncertainty associated with spatially locating subjects in epidemiologic studies. Together, the analyses in this dissertation illustrate unique approaches to geographically referencing health outcome data, controlling study subjects' residential mobility, and identifying population-based control groups. The case-control studies also demonstrate significant associations between environmental indices and both asthma prevalence and cancer incidence. |
