Are Antidepressants More Effective than Placebos?

By: Garry Cooper, LCSW

Reprinted with permission from Psychotherapy Networker magazine. Copyright, Psychotherapy Networker, Inc.

 

According to a new met-analysis, pharmaceutical companies’ own data on 47 placebo-controlled antidepressant trials reveals that on average, there’s virtually no clinical difference between antidepressants and placebos. Information for the article, which appeared in the July 15 edition of the American Psychological Association’s electronic journal Prevention and Treatment, was obtained by researchers Irving Kirsch and Thomas J. Moore using the Freedom of Information Act, and is the second such study based directly on drug companies’ own data.

The analysis finds that the average difference between people who improved with placebos and with the six SSRI antidepressants most widely prescribed from 1987 to 1999 is a clinically insignificant two points on the 61-point Hamilton Depression Scale. Kirsch and Moore emphasize that because the results of these trials were submitted directly to the Food and Drug Administration, any bias in the methodology probably favored the antidepressants.

The findings are accompanied by nine commentaries. While none of the commentaries disputes the statistical conclusions contained in the analysis, several maintain that antidepressants are clinically useful anyway. Acknowledging that the meta-analysis has done a service by airing a “dirty little secret” long known by researchers, Vanderbilt University researcher Steven Hollon nevertheless argues that averages don’t tell the whole story. Some people respond by much more than two points to antidepressants and not at all to placebos, he says, and so a conclusion that antidepressants are no more effective than placebos overstates the significance of the statistics.

Other commentaries point out that the narrow differences say as much about study design as about the relative effectiveness of placebos and antidepressants. The purpose of such studies, they say, isn’t to prove how much more effective antidepressants are, but merely to satisfy the FDA that they work. Furthermore, in studies of all kinds of medications and treatments, the rates of positive responses to placebos have been increasing over the years. This calls into question the purpose and effect of giving placebos, which is to provide null treatment. One commentary notes that the increasing effectiveness of placebos raises the bar toward an impracticably high level for any treatment to demonstrate its effectiveness.

In a response to the commentators, Kirsch posits four current treatments for depression. Treatment A, with a large therapeutic response but a long list of adverse side effects; B and C, with nearly comparable results and no adverse side effects (in fact, B produces positive side effects), but validated with far fewer studies; and D, widely proven as effective as A in the short term and more effective in the long term, and with no adverse effects. Kirsch asks his critics which treatment they would choose–SSRIs (A), exercise (B), bibliotherapy (C), or psychotherapy (D).