Validity of clinical trials of antidepressants

Recent reports have criticized the design of antidepressant studies and have questioned their validity. These critics have concluded that antidepressants are no better than placebo treatment and that their illusory superiority depends on methodologically flawed studies and biased clinical evaluations. It has been suggested that the blind in randomized trials is penetrable-since clinician's guesses exceed chance-and that only active placebo can appropriately camouflage the difference between drug and placebo response. Furthermore, evidence has been cited to suggest that psychotherapy is as effective as antidepressants in both the acute and maintenance treatment of depression. These positions are often accepted as valid and have been broadly discussed in both the lay press and scientific literature. The purpose of this review is to reassess the cited data that support these assertions.

The authors examined the specific studies that were cited in these reports, evaluated their methodology, and conducted aggregate analyses.

Analyses of the original sources failed to substantiate 1) that standard antidepressants are no more effective than placebo, 2) that active placebo offers an advantage over inactive placebo, or 3) that substantial evidence of a medication bias is suggested by raters' treatment guesses exceeding chance. The authors also note that some researchers have suggested that the interpretation of psychotherapy trials can be complicated by "allegiance effects."

The issue of bias or allegiance effects for both antidepressant and psychotherapy research is real. Investigators of all orientations must guard against potential bias. However, studies cited as supporting the questionable validity of antidepressant trials fail upon closer examination to support assertions that these trials are invalid.