mental evidence can enhance the credibility of epidemiologic findings by indicating mechanisms of disease or by corroborating the basic association between exposure and disease (Davis, 1988). However, the absence of such information does not necessarily invalidate a causal association. The underlying mechanisms for many modern diseases are not readily apparent. Thus, the precise mechanisms by which cigarette smoking induces lung cancer are unknown, although the causal relationship is clear (Doll and Peto, 1978).


Among the major challenges for this field are the small populations and small numbers of events usually involved in such studies and the consequent lack of significance of findings, even when the confidence interval is used. The advent of meta-analysis offers an important opportunity to strengthen the inferences that can be drawn from epidemiologic research (Chalmers et al., 1987). Potentially misleading conclusions can be extracted from single studies because of insufficient sample size, inadequacies of exposure determination, or publication and other biases. Meta-analysis can combine data from a variety of studies and reduce the danger of misinterpretation because it allows for pooling of all available information (Greenland, 1990). Meta-analyses must be carried out properly if they are to supply useful information. Retrospective combinations of research must be approached with caution. Searches for primary studies must be as exhaustive as possible. Biases must be minimized by blinding the evaluators of the methods of the studies with regard to the authors, institutional sources, and findings of the original studies. Opportunities for bias in the original research must be tabulated and used to temper conclusions. The statistical methods must be logical and reliable. The interpretation of meta-analyses must also be tempered by the awareness that reporting and publication biases can distort the sample of studies available for pooling. The future widespread application of meta-analytic techniques to studies of hazardous-waste sites will require recognition of deficiencies in primary research and consequent improvements in the gathering and reporting of data in a way that will facilitate meta-analysis later.

Another challenge to environmental epidemiology is the major effect that emerging case and tort law wield on the subject matter. In most tort cases, a plaintiff must demonstrate by a preponderance of the evidence that his or her version of the facts is correct. The number of lawsuits that request monetary payment in compensation for injury induced by exposure to toxic substances has skyrocketed in

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement