Their analysis is the first effort to account for unpublished tests of such therapies. Treatments like cognitive behavior therapy and interpersonal therapy are indeed effective, the analysis found, but about 25 percent less so than previously thought.
Doctors have long known that journal articles exaggerate the benefits of antidepressant drugs by about the same amount, and partly for the same reason - a publication bias in favor of encouraging findings. The new review, in the journal PLOS One, should give doctors and patients a better sense of what to expect from various forms of talk therapy, experts said, if not settle long-running debates in psychiatry about the relative merits of one treatment over another.
Five million to six million Americans receive psychotherapy for depression each year, and many of them also take antidepressant drugs, surveys find. Most people find some relief by simply consulting a doctor regularly about the problem, experts said. Engaging in a course of well-tested psychotherapy, according to the new analysis, gives them an added 20 percent chance of achieving an even more satisfying improvement, or lasting recovery. Before accounting for the unpublished research, that figure was closer to 30 percent, a difference that suggests that hundreds of thousands of patients are less likely to benefit.
The new paper is the latest chapter in a broad retrenchment across science in which researchers are scrutinizing past results to weed out publication bias and other, more deliberate statistical manipulations.
"We need to seriously consider publishing all completed studies," whether encouraging or not, said Jelte Wicherts, an associate professor in the department of methodology and statistics at Tilburg University in the Netherlands, who was not involved in the study.
Other researchers cautioned that the analysis was hardly the final word on the effectiveness of psychotherapy. "The number of trials they looked at was fairly small, and the different psychotherapy approaches were all pooled together," said Stefan Hofmann, a professor of psychology at Boston University.
In the study, a research team led by Ellen Driessen of VU University in Amsterdam tracked down all the grants funded by the National Institutes of Health to test talk therapy for depression from 1972 to 2008. The team found 55, most of which used so-called manualized approaches, in which therapist and patient use a standardized manual to guide the treatment.
The most commonly studied of these are cognitive behavior therapy - in which people learn to identify and defuse automatic, self-defeating assumptions, like "I'm unlucky in love" or "I always choke" - and interpersonal therapy, which focuses on reshaping how people interact with others. These studies typically have subjects engage in weekly, hourlong sessions with a therapist for three to four months.
The researchers found that 13 of these funded studies were completed but never published, usually because those who did the trials did not think a finding of no benefit stood much chance of being published. The team contacted each of the 13 investigators originally paid to do the work and requested their data. Once that data was included with those from the other, published papers, the effectiveness of the therapies dropped significantly - by about a quarter.
"That seems to be the magic number, a quarter - about the same as you see in the pharma trials" of antidepressants, said a co-author, Dr. Erick Turner, an associate professor of psychiatry at Oregon Health & Science University and the lead author of the 2008 paper detailing bias in those drug trials.
The team did not have enough information about the original studies' designs to determine whether the authors massaged any data to make the treatment look better than it was, as happened in some of the drug trials. Had study designs been available, Dr. Turner said, the benefits of psychotherapy might have been lower. His co-authors were Steven Hollon of Vanderbilt University, Claudi Bockting of the University of Groningen, Pim Cuijpers of VU University, and Dr. Driessen.
The way to think about the results, Dr. Hollon said, is that antidepressant drugs and talk therapies are modestly effective, and the combination is better than either approach alone. But for those who do well or fully recover, "psychotherapy, particularly cognitive behavior therapy, seems to be most effective in cutting the risk for a relapse long-term," Dr. Hollon said.
The new report's findings did not come as a big surprise to many therapists who use talk therapy. "Depression is a tough disorder to treat, and it's very difficult also to judge treatments because the symptoms of the depression naturally wax and wane - it's a moving target," Dr. Hofmann said. "There's a sense of desperation out there because we do need something new, and there's very little on the horizon."
Return to News Home