Methodological issues in epidemiologic studies using biologic markers.

Biologic markers are becoming prominent features of many classical epidemiologic studies. Their existence has also modified the character of some epidemiologic research such that the term "transitional epidemiologic studies" may be warranted. In the latter type of study, collaboration between laboratory and epidemiologic investigators is integral to the research. In this paper, goals, characteristics, and examples of transitional epidemiologic research are presented. Pertinent features are highlighted, including sources of misclassification and confounding, and elements of study design useful in reducing these potential biases. Sample size issues, transformation of variables, and sources of variability acquire enhanced importance. The study examples presented are intended to illustrate the altered substrate for these methodological features in studies using biologic markers.