Content provided by the Faculty of the Harvard Medical School

In Brief: Testing antidepressants: How methods affect results

In brief

Testing antidepressants: How methods affect results

In many clinical trials of antidepressants eventually approved by the FDA, there was little or no difference in therapeutic effect between the drug and a placebo. An analysis published in the American Journal of Psychiatry aims at teasing out features of experimental design or choice of patients that made for success or failure of the drug.

The authors examined FDA reports for 52 clinical trials of 9 antidepressants evaluated between 1985 and 2000. On average, the difference between the antidepressant effect and the placebo effect was a modest 3.1 points on a standard rating scale for depressive symptoms. The analysts defined as successful the 26 trials in which the drug-placebo difference was greater than that.

Surprisingly, the average number of subjects was about the same in successful and unsuccessful trials, and subjects took the drug (or placebo) for about the same length of time whether the trial succeeded or failed.

Another surprise was that successful trials included a lower proportion of women, although almost all other available evidence indicates that antidepressants are equally effective in both sexes. The authors point out that their unusual finding is difficult to interpret because they did not have access to results recorded individually by sex.

The most common ingredient of a successful trial was flexible dosing, that is, permission to adjust the amount of drug (or placebo) instead of using a fixed dose. Since that is how antidepressants are used in standard practice, they could be more effective in everyday life than the average clinical trial suggests.

Khan A, et al. “Research Design Features and Patient Characteristics Associated with the Outcome of Antidepressant Clinical Trials,” American Journal of Psychiatry (Nov. 2004): Vol. 161, No. 11, pp. 2045–49.

Last updated:	August 21, 2006