wastes, but their successful application will vary. Thus, studies of trends over time in air pollution and disease patterns have produced a growing body of literature that has associated day-to-day fluctuations in air pollution with daily fluctuations in mortality across a wide range of exposures with no evidence of thresholds. It is not likely to be worthwhile to conduct such studies at hazardous-waste sites, especially because the pollutants are complex and because there are no long-term records, such as exist for a number of monitored air pollutants.
Cross-sectional studies provide epidemiologic snapshots of a given area at one point in time. Recent computer technology has permitted maps to be drawn that show comparative mortality data from different regions of the U.S., Canada, and other industrial countries. Such maps can show county-wide cancer and other mortality data by decade, for example, as a hypothesis-generating tool to detect geographic variations in these diseases and to infer possible causes. Of more relevance to hazardous-waste studies are small-scale comparisons of adjacent counties or ZIP codes, where differences could be better highlighted. The study of health effects that have shorter latency than most cancers—such as birth defects, neurologic effects, and other acute and chronic effects—increases the likelihood that a connection can be drawn between environmental exposures and disease.
Cross-sectional community studies typically compare communities with different levels of air pollution or populations that live different distances from a hazardous-waste site. All such studies have several problems: Measurement error occurs because of the assumption of the same exposure for every subject within a group. There can be undetected differences between communities for risk factors, such as illness, tobacco use, or occupational exposures. There can be “recall bias ” if one group knows it is in the high-exposure category. There is little standardization of the equipment used to measure exposure in different locations.
In spite of these difficulties, successful community studies have been done on air pollution patterns. In contrast to purely descriptive studies, which lack information on potential confounders, community studies generally contain data on nonpollution risk factors. A few studies have involved materials like those which occur at hazardous-waste sites. Excesses of the rare cancer angiosarcoma occurred in residents near a vinyl chloride manufacturing plant. Another study found increased rates of birth defects in children whose parents lived near such plants.
Longitudinal analyses also have been developed for the epidemiologic study of air pollution. These have some direct bearing on the study