Observational studies of cause-effect relationships: an analysis of methodologic problems as illustrated by the conflicting data for the role of oral contraceptives in the etiology of rheumatoid arthritis.

We review the structure of the two most common observational study designs, cohort and case-control studies, and outline guidelines to help clinicians evaluate the validity of studies that employ these designs. To demonstrate the use of these guidelines we apply them to the important issue of whether oral contraceptive usage may reduce the risk of developing rheumatoid arthritis. A recent case-control study has contradicted provocative evidence from two previous studies suggesting the existence of such a protective effect. Our analysis leads to the conclusion that the case-control study suggesting no association of oral contraceptives with the subsequent development of rheumatoid arthritis provides the best estimate of the effect of oral contraceptives currently available, and furthermore, the analysis permits recommendations to be made for the design of future investigations on this topic.