Aims: In the 1950s, Eysenck suggested that psychotherapies may not be effective at all. Twenty-five years later, the first meta-analysis of randomised controlled trials showed that the effects of psychotherapies were considerable and that Eysenck was wrong. However, since that time methods have become available to assess biases in meta-analyses.
Methods: We examined the influence of these biases on the effects of psychotherapies for adult depression, including risk of bias, publication bias and the exclusion of waiting list control groups.
Results: The unadjusted effect size of psychotherapies compared with control groups was g = 0.70 (limited to Western countries: g = 0.63), which corresponds to a number-needed-to-treat of 4.18. Only 23% of the studies could be considered as a low risk of bias. When adjusting for several sources of bias, the effect size across all types of therapies dropped to g = 0.31.
Conclusions: These results suggest that the effects of psychotherapy for depression are small, above the threshold that has been suggested as the minimal important difference in the treatment of depression, and Eysenck was probably wrong. However, this is still not certain because we could not adjust for all types of bias. Unadjusted meta-analyses of psychotherapies overestimate the effects considerably, and for several types of psychotherapy for adult depression, insufficient evidence is available that they are effective because too few low-risk studies were available, including problem-solving therapy, interpersonal psychotherapy and behavioural activation.
Keywords: Depression; outcome studies; psychotherapy; randomised controlled trials.