Psychodynamic Therapy or Cognitive Therapy for Social Anxiety Disorder

To the Editor: I read with interest the article by Leichsenring et al. (1) in the July issue of the Journal. Cognitive therapy was significantly more effective than psychodynamic therapy on four of five social anxiety measures, including all those mentioned in the trial registration document (ISRCTN53517394). As one of the developers of cognitive therapy, I was naturally delighted to see the trial extend the range of credible psychological therapies to which cognitive therapy has been shown to be superior (2). However, several design features suggest the trial is likely to have underestimated the benefit of cognitive therapy relative to psychodynamic therapy. Readers may wish to bear these in mind when considering the relatively small differences in outcome that are reported.

First, to match psychodynamic therapy, the time over which cognitive therapy was delivered was extended from the usual 3–4 months to 8–9 months. There is good evidence that slowing down the delivery of cognitive-behavioral treatments for social anxiety disorder reduces their effectiveness (3).

Second, “type of treatment” is confounded with “therapist experience.” On average, psychodynamic therapists were qualified and had 8.0 years of clinical experience, whereas cognitive therapists had only 1.7 years of clinical experience and many would still be trainees. In our clinic, cognitive therapy for social anxiety disorder achieves greater effects when delivered by fully trained therapists.

Third, the average competency with which cognitive therapy was delivered seems to be low. Mean ratings of therapy videotapes on the cognitive therapy competence scale (unpublished 2013 paper of U. Stangier) suggest a substantial number of sessions fell below the “redline” minimum of 3.5 (out of 6.0). This is important because there is good evidence that competence predicts outcome in cognitive therapy for social anxiety disorder (4).

Fourth, behavioral experiments in which therapists leave the office with patients to test their beliefs in real life are a central feature of cognitive therapy. To make such experiments feasible, the standard cognitive therapy protocol recommends 90-minute sessions. Almost all Leichsenring and colleagues’ sessions were 55 minutes, making it unlikely that such experiments were used regularly. Comparisons between previously published trials (2) show that cognitive therapy is associated with approximately 50% more improvement in social anxiety when sessions are 90 minutes compared with 55–60 minutes.

In summary, Leichsenring and colleagues’ conclusion that cognitive therapy and psychodynamic therapy are both efficacious (superior to no treatment) is justified, as is the conclusion that cognitive therapy is superior to psychodynamic therapy. However, one might reasonably question their estimate of the relative effects of cognitive therapy and psychodynamic therapy. A further trial in which cognitive therapy is delivered as recommended and at an adequate level of competence would be required to clarify matters.