Need for Randomized Clinical Trials

To the Editor: It would be unfortunate indeed if the apologia of Gehan and Freireich (N Engl J Med 290:198–203, 1974) should have the effect of undermining our confidence in randomized clinical trials, cultivated so painstakingly over the past half-century and finally beginning to mature. The authors' basic assumption seems to be that an investigator can establish with sufficient precision the probability of a success in a new drug trial by comparing the prognostic features of his patients with a similar group treated at another time or place, making concurrent controls unnecessary. Unfortunately, this assumption is valid only for situations.