The medical community was disappointed, but perhaps not shocked, several years ago when it was revealed that the efficacy of antidepressant medications in the treatment of adolescent depression might not be as strong as we thought. This bias was due to the fact that the clinical trials, most of which were funded by pharmaceutic companies, that showed a positive effect for the
medication were published prominently in journals while negative trials got much less publicity and often were not published at all. A deliberate effort to downplay unfavorable studies of antidepressants was generally suspected, although some people have argued that other factors may also be at work, such as journals being less enthusiastic to publish negative findings and researchers not wanting to devote lots of time and energy describing hypotheses that did not pan out.
With regard to depression, the publication bias for antidepressants made some clinicians think more strongly about recommending psychotherapy. Now, however, a new study recently published in PLOS One, examines whether or not a similar bias may be present in psychotherapy studies of depression.
The authors looked back at grants awarded by the NIH for controlled psychotherapy studies for depression conducted between the years 1972 to 2008. They found the publications for these studies and examined the effect size by which the therapy improved symptoms relative to placebo. If they found a funded grant but no publication, the authors contacted the investigators and calculated the effect size for those studies. Studies of children and adolescents, unfortunately, were excluded for these analyses.
Surprisingly, nearly a quarter (23.6%) of the 55 funded studies did not result in publication and two studies were never started at all. Most of the unpublished studies were never submitted for publication in the first place (rather than it being written up and rejected over and over by different journals). Most notably, the effect size of unpublished trials was less than half that of published trials. When these unpublished studies were added to the published ones, the effect size for psychotherapy dropped by 25%.
The authors concluded that there is strong evidence to conclude that a publication bias exists for psychotherapy trials for depression and that our understanding of how effective psychotherapy is, at least for depression, may need to be adjusted somewhat. That said, the evidence continues to show that psychotherapy generally works for depression and there is no reason to alter the general recommendation of psychotherapy as a first line treatment.
This study is useful in reminding us that bias can exist in many domains and for many reasons. Most of us are now quite primed to recognize potential bias and conflict when it comes to the pharmaceutical industry, but we need to be vigilant about some of the less obvious sources as well.
Reference
Driessen E, et al. Does Publication Bias Inflate the Apparent Efficacy of Psychological Treatment for Major Depressive Disorder? A Systematic Review and Meta-Analysis of US National Institutes of Health-Funded Trials. PLOS One, epub ahead of print, 2015.
Tags: antidepressants, bias, depression, psychotherapy, therapy