CBT vs psychodynamic therapy for Social Anxiety Disorder

In the October 2014 issue of The American Journal of Psychiatry we see an article by Leichsenring et al (18 authors!) comparing the outcome of social anxiety patients who had received either CBT or psychodynamic therapy.  The patients had about 25 sessions of either therapy, over about 9 months time.  They were followed up over the following 2 years after treatment ended.

The study shows that both groups improved similarly over 2 years:  about a 70% response rate, and a 40% remission rate.

But, huge weaknesses in the study here!

1) No placebo group!  
2) No documentation of the homework done in CBT.
3) No detailed description of how the psychodynamic therapy differed from the CBT, other than a passive reference to the technique or manuals used.


I feel that psychodynamic theory is similar to religious belief or theology:  it is finally a set of cultural practices, couched in a therapeutic milieu.  The actual beliefs are substantially fictional, but are grounded in basic ethical principles expressed in scholarly or literary language.    Similar to a great cathedral, a poetic section of a religious text, or a beautiful hymn, the therapeutic impact comes from the esthetics and earnestness of the fellow practitioners, mixed together with the style being a largely accepted cultural norm.  Fragments of accurate science are blended with fictional but culturally vivid therapeutic dogma (e.g. references to Greek mythology), a product of the testimonial accounts and opinions of strong-minded and literary thinkers, who yet are often poor scientists.  In some ways, it is akin to a medieval alchemist or astrologer, whose theories are mostly fictional, but who may still have a loving and intimate appreciation of their subject matter.  In psychodynamic therapy, there would clearly be a sense of attachment, security, a type of friendship or mentorship (even though these qualities would be normally never be admitted, except as "transference"), and an earnest focus on improvement.

In CBT, many of these same factors would be present, though in a more "coachlike" form.  One of the problems with CBT is that the cultural esthetics of the therapy is largely absent, compared to psychodynamic therapy.   If we compare CBT and psychodynamic therapy to religious denominations, it would be as if CBT would have its meetings in an accountant's office, while the psychodynamic sessions would take place in an environment laden with cultural symbolism, such as a church or cathedral, with musical or poetic accompaniment.  

So one of the strong therapeutic elements of psychodynamic therapy (the "cathedral-like" intellectual esthetics) is compellingly absent in most CBT.  I suspect some of the newer forms of CBT, such as mindfulness-based CBT, are introducing some more of this esthetic element, leading to improved effectiveness.

In treating anxiety of any sort, it appears obviously true to me that the therapy must involve the patient having many hours of practice facing anxious situations.  It is limited how much of this practice can actually take place during a CBT session.  Most of the practice would have to take place as homework.  As I have said elsewhere, psychotherapeutic change in many ways is akin to language learning, or to learning a physical skill or sport.  You can have your weekly lessons with the coach, but most of your improvement will take place if you diligently practice every day.


In this study, there was no mention of this most essential therapeutic agent of all:  the practice done, to face social anxiety situations!  Even in psychodynamic therapy,  I would expect that the therapist would facilitate exposure practice between sessions, even if this was not deliberately prescribed.  In some ways, with a resistant patient, a sensitive psychodynamic therapist could be more effective than a CBT therapist to do such encouragement effectively, just as a good priest may simply have a more effective interpersonal manner to encourage someone in a time of distress, compared to a good accountant.  


But no mention was made of how much the patients actually practiced their skills to manage social anxiety.

I find it quite incredible that 18 scholars, all touting their doctoral degrees in the author list, were required to produce such a trivial paper. 

Varenicline plus Bupropion for smoking cessation

Rose and Behm have shown in their November 2014 article in The American Journal of Psychiatry that 12 weeks of a combination of varenicline 1 mg twice daily combined with bupropion 150 mg twice daily, led to substantially improved abstinence rates for highly nicotine-dependent smokers. 

Most smoking cessation strategies have led to quite low abstinence rates.  A typical outcome would be a 25% probability of quitting after a determined attempt.  This is the first study I've seen that shows a strategy that leads to a 50% abstinence rate.  In fact, they found that the combination works best for the heaviest smokers who were most addicted. 

With smoking cessation, as with many other problems, I think that if a pharmacological strategy is considered, why not try the most effective strategy first? Why not try this combination first, rather than trying one much less effective treatment at a time?

Some remaining questions I have about ongoing management would be to question whether long-term varenicline could be necessary (e.g. for a year or more). 

And, with smoking, a big question now concerns the potential benefits and risks of e-cigarettes.  These are probably good harm reduction aids for many smokers, but on the other hand are addicting on their own, and could initiate dependency problems in young people who try them before smoking at all.   Overall, I think e-cigarettes are an important positive development to help people quit smoking, and also to help deplete the tobacco industry further. 

Quetiapine for borderline personality -- journal article review

This is the first in a planned series of posts to summarize a few interesting articles from psychiatry journals published in 2014.

We begin with an article by Donald Black et al.from The American Journal of Psychiatry 171:1174-82.

It's a very simple 8-week randomized controlled study of treating borderline personality patients with either quetiapine XR 150 mg daily, quetiapine XR 300 mg daily, or placebo.  There were about 100 participants in all.   DSM-IV criteria were used for the diagnosis, and the participants could not have active substance abuse, or be in the midst of a major mood or anxiety episode, etc. 

The "Zanarini scale" was used to track symptom changes.  As I look up this scale, I find it appears to be a simple distillation of DSM-IV criteria, with raters giving each item a numerical score.   Unbelievably, I find that I cannot actually look at the questions directly (a fee of over $40 is requested!), which is quite surprising for what amounts to a small collection of very simple questions.

Nevertheless, the quetiapine groups did better than the placebo group on the borderline symptom scales.  But they did not do compellingly better on broader scales including the Sheehan Disability Scale or the GAF.    There was no advantage of the 300 mg dose over the 150 mg dose.

A few criticisms:

 1) I see the placebo group actually had lower baseline symptom scores, which could have biased the placebo group to show less improvement (e.g. through regression to the mean contributing to the larger symptom changes in the other two groups).     The fact that the graph given in the article showed only symptom change, rather than total symptom score, would have further hidden this bias from the reader.  The error bars were not shown in the graph of symptom change.   I see that the total symptom scores are not shown anywhere in the paper! I'm surprised this got past peer review in a major journal!


2) While 150 mg is considered "low dose" here, it would be useful to see what the effect of 25 mg or 50 mg would be. 

3)  As usual with studies of this sort, it is only 8 weeks in duration.  I would be interested in seeing a duration of at least a year.  This would be relevant not only for evaluating effectiveness (including symptom improvements and dropout rates), but also for evaluating side-effect risks (such as weight gain and metabolic changes).

4) The question is not addressed as to whether the more expensive quetiapine XR preparation is actually needed, compared to the less expensive regular quetiapine.  


In summary, a simple, mediocre study, which lends modest support for a practice that most practitioners probably already have done for years anyway -- which is to offer borderline patients treatment with low-dose atypical antipsychotic medications.