Abstract |
Given that a major task for environmental epidemiology is to provide evidence of long-term health risks early enough so that appropriate preventive measures can be taken, biochemical and biological markers of potentially harmful environmental exposures are of great interest and possibly of great value. Such markers are in two discrete classes, those quantitatively related to exposure itself, reflecting the magnitude of such exposures or of the body burden of the pollutants, and those markers which reflect the biological response to such exposures. This paper suggests guidelines for the appropriate use of cross-sectional, retrospective, and prospective designs in epidemiologic investigations of markers of exposure and disease. Also presented are suggestions for the application of sensitivity, specificity, reliability, and acceptability in marker studies since the effectiveness of any marker will ultimately be judged on these characteristics. |