Generalizing from clinical trials.

Although randomized controlled clinical trials have become the "gold standard" for evaluating new treatments, only a small subset of the population considered for treatment participate in randomized clinical trials. To what extent is it reasonable to generalize beyond the boundaries of a specific clinical trial? This paper argues that several pieces of information are necessary to determine the extent of extrapolation or generalization warranted in a specific clinical trial. The necessary items of information are derived from basic science laboratory studies; animal studies; genetic studies (where applicable); observational, clinical, and epidemiological studies; and other randomized clinical trials in similar settings or with similar treatments. An example from the field of cholesterol reduction is presented.