Provocative Thoughts Emerging From a Study of Psychotherapies for Panic Disorder

Psychotherapies for Panic Disorder: A Tale of Two Sites was an exciting, landmark study that my colleagues (Dianne L. Chambless, PhD; Jacques P. Barber, PhD; and others) and I conducted, and several of its important points deserve highlighting. It’s an important achievement that this rigorously conducted study demonstrated that a time-limited, panic-focused psychodynamic psychotherapy (PFPP) had efficacy for patients who had panic disorder with and without agoraphobia. While smaller studies have shown the efficacy of PFPP, this is the first larger-scale, multisite study to do so. The evidence-based psychotherapy most often provided for anxiety disorders these days is cognitive-behavioral therapy (CBT). CBT is a powerful, effective treatment for panic disorder, and it performed well in this study, too. Nonetheless, this approach doesn’t benefit all patients, so having a variety of types of psychotherapy with different strategies and focuses is important to help patients in need.

Unfortunately, in our study, outcome analyses revealed site-by-treatment interactions. No investigators would wish to find site-by-treatment differences in outcome in a randomized controlled trial, and we are no exception. Patients in PFPP improved more at Cornell than those at University of Pennsylvania, whereas patients receiving applied relaxation training fared better at Penn than at Cornell. While the populations differed from one another in a number of crucial ways, these differences did not fully explain the site-by-treatment differences we observed. Like all unexpected research findings, this unexplained difference has led the investigators to explore more carefully which specific elements of the psychotherapies yielded greater improvements among which specific patients. Without scooping my co-investigators in the analyses they are conducting, I can say that we are currently using this wonderful dataset to analyze the following questions:

1. Did specific moderators (pre-existing characteristics that patients have, like their age, for example) affect outcome within and across psychotherapies?
2. Can specific (a priori) mediators of change be pinpointed as leading to improvement?
3. How did responders to the different psychotherapies fare at 12-month follow-up?
4. What, if anything, differed about the manualized, mostly adherent psychodynamic therapy that the therapists delivered to the patients at the 2 sites? (This question is in some ways the one most interesting to me, as a dynamic therapist.)

We expect that some of the answers to these questions will help the field of psychotherapy for anxiety to move forward with better understanding as to which treatments work better for which patients, how they work (which may or may not be related to underlying causes of illness), and how to improve the psychotherapy we offer to patients.

Financial disclosure:Dr Milrod has received a Cornell Clinical and Translational Science Center grant; receives support from the New York Community Trust in a fund established by DeWitt Wallace; has received honoraria for various academic talks; and has received book royalties from Taylor & Francis.​