Health is affected by the places in which people live, work and interact, yet many epidemiological studies overlook the characteristics of places and instead focus solely on the people who inhabit them. Place-based investigations of disparities in health outcomes are concerned with the healthiness of places and not merely the healthiness of the populations in these places. A place-based approach has been used within medical geography and medical sociology, typically in the study of health differentials associated with long-term, cumulative exposures to a wide range of environmental variables. This approach has rarely been extended, however, to health research that looks at the effects of extreme events (such as industrial accidents or hurricanes). The purpose of this paper is to incorporate a place-based framework into extreme event health research. The paper first discusses methodological considerations for a place-based approach and then illustrates the use of spatial analysis techniques as the first step in identifying place-based risk factors in mortality associated with heat waves. The study centers on St. Louis, Missouri, a city where hear waves are frequent and heat-related mortality is high. The results show that heat-related mortality rates during the most severe heat waves were generally higher in the warmer, less stable and more disadvantaged areas of St. Louis and lower in the cooler and more affluent parts of the city. During the milder years analyzed, there was little evidence of a relationship between place-based characteristics and the distribution of heat-related mortality. These findings about extreme event mortality risk would not have been evident from a population-based analysis. Ongoing dialog between epidemiologists and social scientists can help to bring place into the arena of extreme event research and to increase understanding of the role of place in risk. (C) 1998 Elsevier Science Ltd. All rights reserved.
