Epidemiological designs in radioepidemiological research.

Most of the information currently available on radiation effects comes from the observation of individuals who received high doses, either from one acute exposure or a series of exposures. Because of the uncertainties in extrapolating risks from these studies to the exposure circumstances of most cancers today (in particular, low-level environmental exposure), direct epidemiologic studies of populations receiving low doses chronically are needed. The effects which we want to quantify are likely to be small, however, and extreme care must therefore be put into the design of such studies if they are to be informative for risk assessment purposes. The areas of particular importance in designing risk assessment studies are reviewed. The advantages and limitations of various epidemiological study designs, and the problems and importance of dosimetry are discussed. Examples from the recent epidemiological literature are presented to illustrate the discussion.