Toward an "unconfounded" classification of epidemiologic research design.

Existing classifications of research designs for epidemiologic studies are inconsistent and confusing because of the conceptual "confounding" of three distinct concepts: directionality in which exposure and outcome are investigated, sample selection criteria, and timing of the study proper with respect to the calendar times of exposure and outcome. The authors have attempted to construct a classification in which these three concepts comprise distinct (if not totally independent) axes. Closer attention to the distinctions described may help the planning and interpretation of epidemiologic studies, as well as facilitate future teaching and communication.