Some Problems Arising in Prospective and Retrospective Studies of the Etiology of Disease

Abstract

SIX years ago Bradford Hill, in his Cutter Lecture on the role of observation and experiment in preventive medicine, pointed out that although the experimental method clearly is preferable whenever it can be used, many, if not most, of the problems arising in the investigation of the etiology of disease in human populations, for ethical or other reasons, must be studied by the observation of naturally occurring events. Therefore he did not "repudiate or even. . . underrate the claims of accurate and designed observations." But, he emphasized, in using the method of observation, one should not be content with . . .