Let medicine lead the way beyond the registration of clinical trials: registering hypotheses

There has been much recent interest in improving the quality and reporting of clinical research. Major journals now require clinical trialists to register studies a priori and follow specific reporting guidelines. While best developed for randomized, controlled trials, guidelines for other study types are being developed. A hypothesis is presently available for 4 of 5 publications reporting randomized, controlled trials, much less common for other study designs and scarce in the general scientific literature. The declaration of a hypothesis or its absence should be a standard feature of all scientific reporting. The recording of the hypothesis in a registry before data collection would assure readers that the hypothesis was a priori. Medicine has been in the vanguard in requiring methodological rigor in the reporting of science and should again push the envelope by developing a hypothesis registry. If successful, the gauntlet would be thrown down for the rest of science.