Response to Clark

To the Editor: Dr. Clark raises several concerns concerning our trial (1). First, cognitive therapy was implemented by German experts in cognitive therapy (Drs. Stangier, Hiller, Hoyer, Willutzki, and Leibing), and it strictly followed the manual by Stangier, Clark, and Ehlers. No changes were made in the delivery of cognitive therapy to match psychodynamic therapy. Furthermore, the study by Herbert et al, cited by Dr. Clark, examined the Heimberg approach, not that by Clark, and the extended treatment for the completers did not show a smaller, but a descriptively larger, pre-post effect size (1.95 compared with 1.83).

A second concern raised by Dr. Clark is that type of treatment was confounded with therapist experience. However, we have shown that differences in therapeutic experience had no impact on the results (1). Third, Dr. Clark suggests that the competency for the delivery of cognitive therapy was low, and more competent cognitive therapists could have achieved better outcomes. Unfortunately, in their 2006 study (2), Clark et al. did not report therapist competence to allow for comparison. Competency in our study was comparable to another successful German study of cognitive therapy (unpublished 2013 paper of U. Stangier). In our study, 55 therapists treated 209 patients with cognitive therapy under the conditions of clinical practice. In the study by Clark et al., six specialized and prominent therapists treated 21 patients. The high specialization of therapists, their eminent reputation, the small number of therapists and patients, and the lack of any dropouts question the generalizability of the results by Clark et al. Lastly, as a result of carrying out cognitive therapy under the conditions of routine practice, we may have carried out fewer behavioral experiments than did Clark et al.

Some issues raised by Dr. Clark apply to psychodynamic therapy as well. Whereas cognitive therapy was tested in several earlier studies, the method of psychodynamic therapy was newly developed for this trial (1). In addition, the competency ratings for psychodynamic therapy were below the optimum. For these reasons, the true efficacy of the new method of psychodynamic therapy may have not yet been demonstrated.

Furthermore, we have not concluded that cognitive therapy is superior to psychodynamic therapy in social anxiety disorder. We reported superiority of cognitive therapy only for specific outcome measures referring to statistical significance. However, clinically significant differences in the primary outcome measures were not established, as the differences in success rates (8% and 10%) were below the 15% we had a priori defined as meaningful (h=0.30) (1). This conclusion is supported by the following: the probability that a patient treated with cognitive therapy achieves a higher remission rate than a patient treated with psychodynamic therapy is 0.55 (3)—nearly the same chance as when tossing a coin. In addition, for the primary outcome measure of response, no statistically significant differences were found (h=0.16). As the differences in success rates were below the threshold of clinical significance in the primary outcome measures, recommending cognitive therapy over psychodynamic therapy in social anxiety disorders is not warranted (4). In order to compare the true efficacy of treatments when carried out as competently as possible, a research program that clearly goes beyond a single randomized controlled trial is required.