The experimental paradigm and observational studies of cause-effect relationships in clinical medicine.

The paradigm of the randomized clinical trial (RCT) is proposed as a heuristic that can serve as a unified approach to guide the design not just of cause-effect studies of therapy, but also of studies of the etiology of disease. Three themes are developed in detail: that variability in the scientific paradigm of the randomized trial results in a wide range of techniques and methods being employed for the RCT, and that this extensive variability in clinical trial methods contributes substantially to the occurrence of conflicting trial results; that the scientific validity of observational surrogates for the RCT could be enhanced if investigators designed observational studies by incorporating the principles inherent in the RCT; and that there are two distinctive and competing strategies for designing case-control studies. The traditional strategy views case-control designs as statistical acts of sampling for cases and controls, but ignores the scientific reasoning that should guide the performance of case-control research. The alternative strategy requires that case-control studies adhere to the principles inherent in the RCT.