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 in light of the complex and changing nature of the pollutants, the absence of long-term records, such as exist for criteria air pollutants, and problems of determining the baseline, or expected rates, of a variety of subtle health end points of interest, such as neurological, behavioral, and reproductive problems.
Lessons that have been learned in air pollution studies are relevant to epidemiologic studies of hazardous-waste sites. Of particular importance are the need to measure exposure as precisely as possible, and the value of obtaining longitudinal data on exposures and disease outcomes in order to strengthen time-series analyses. The recent symposium organized by the National Academy of Sciences (NRC, 1991) stressed the paramount importance (and difficulties) of exposure measurement.
Symptom reports appear to be sensitive indicators of adverse health effects. Simultaneous use of air monitoring and diary records could reduce the problem of recall bias, and are particularly valuable when small changes in pollutant levels cannot be detected by the subjects in a study. It is likely that air emissions from hazardous-waste sites have caused a variety of symptoms indicating low-level interference with normal function. These are often comparable to the symptoms reported in the sick building syndrome. There are insufficient data to determine whether or not airborne exposure to toxics from hazardous-waste sites has resulted in nationwide increases in cancer mortality, or adverse pregnancy outcomes. However, the limited number and low power of studies, and the long latency of some cancers and other chronic diseases, mean that these effects cannot be completely ruled out in most areas.
It is not easy to decide whether to launch an epidemiologic study. With more detailed assessment of exposure, the extent of possible adverse effects will be better understood. We think it is important to draw attention to resource needs for adequate study designs. While a decision to conduct an epidemiologic study of hazardous-waste sites must take account of expressed public concern, this concern should, especially in an era of scarce resources, be balanced by the best available scientific evidence, including the quality of exposure and outcome data and the probability that an answer could be obtained that would be interpretable per se or after combination with the data from multiple studies.