Steps toward discovering causes: divergence and convergence of epidemiology and clinical medicine

Ways in which frames of reference about causation differ across disciplines are exemplified. The necessity for a multivariate concept of causation is emphasized. On the epidemiological side, divergences of clinical approaches from requirements for meeting preconditions for cause are each illustrated by several examples. Association is discussed in terms of comparability of groups, independent observation of associated variables, sampling and measurement. Securing time order among variables is a matter of sequential positioning by design, and the difficulties that inhere in the clinical situation are again illustrated by example. On the clinical side, strengths in generating hypotheses, in discovery of associations without conventional comparison and in potential precision in measuring outcomes are outlined. On the epidemiological side, the discipline adds to clinical medicine unique dimensions analogous with the basic biomedical sciences. These are essential to capturing environmental causes and the antecedent experience of sick individuals. Convergence between the two perspectives is bound to yield, and has yielded, much profit.