Population-based studies assay all persons within a given region. They offer many of the advantages of diary studies because they deal with an entire community, and they allow the examination of rarer outcomes, such as physician visits or hospital admissions. Pope (1989) reports an association between hospitalization for respiratory illness and PM10 concentrations in Utah Valley, located in Utah County of central Utah. The relationship was found at levels well below the current ambient air quality standard (24-hour PM10 standard of 150 micrograms per cubic meter and an annual PM10 standard of an expected arithmetic mean of 50 µg/m3). A particularly striking effect was seen for the opening, closing, and reopening of a local steel mill. Further analysis shows a significant association in Salt Lake County as well (Pope et al., in press). Bates and Sizto (1987) found a significant relationship between summer air pollutants and hospital admissions for acute respiratory disease in southern Ontario. Data from acute-care facilities in the region 's 79 hospitals were taken over a nine-year observation period.
An example of a population-based case-control study relevant to airborne toxic substances, although not specific to hazardous-waste sites, is the one conducted by Linos et al. (1991). Mortality rates for leukemia and non-Hodgkin's lymphoma (NHL) have been rising in the central region of the U.S. Linos and colleagues were interested in the hypothesis that general environmental factors might account for this increase. Subjects consisted of white males in Iowa and Minnesota diagnosed between 1980 and 1983; there were 520 cases of leukemias and 572 of NHL, who were matched to 1130 controls. The relative risks (RR) were adjusted for factors associated with NHL and leukemia, e.g., pesticide exposure, occupational exposure, social class, use of hair dyes, smoking, and family history of hematopoietic malignancy. The authors found a statistically significant increase in the risk of developing NHL (RR=1.4) and a slight, nonsignificant excess for leukemia (RR=1.2) among men who lived 0.8-3.2 km (0.5-2 miles) from a factory. Petroleum or chemical factories were associated with the leukemia risk, while stone, clay, and glass factories correlated with NHL. The authors state: “These environmental associations may provide clues to the unexplained rising morality rates of leukemia and NHL in the central United States” (Linos et al., 1991, p. 73).
Other studies have reported a correlation between visits to hospital emergency departments and pollutant levels. Some of these have been positive (Levy et al., 1977; Bates et al., 1990); in another the association was present but weak (Samet et al., 1981), and in others