Wednesday, October 1, 2008

Long-term Therapy Wins for Complex Cases (i.e., borderline etc.)

Wow. JAMA just published a stunning meta-analysis article evaluating the efficacy of long-term psychodynamic psychotherapy. This is gonna get blogged a lot. I was eager to peek at it, and post a note abt it.

Pubmed does not have this sited yet. Here is the ref:

Leichsenring F, Rabung S. Effectiveness of Long-term Psychodynamic Psychotherapy: A Meta-analysis. JAMA 2008; 300 (13): 1551-1565. October 1, 2008.

Long-term psychodynamic psychotherapy seems to have a good effect for patients with this mixed bag of difficult psychological problems. Problems include chronic depression, bipolar disorder, and others.

Especially to the credit of the authors is the section in which they assess the degree that results may have been biased by non-publication of any unknown studies which hypothetically could have, and reasonably would be expected to have, negative results. That is a long sentence, so I will say it another way: the authors reviewed all avalable decently conducted studies; they obviously can only review those published; there could be unpublished studies that failed to get published because they demonstrated no significant beneficial effect for long-term psychodynamic psychotherapy. The authors conducted some processes to add this unknown factor into their conclusions.

This is a conservative, cautious, judicious move. And commendable.

The conclusion, after that adjustment, or reinterpretation for hypothetical non-significant results that might be out there somewhere, in the "file drawer" of some researcher who could not or chose not to get it published, still remained.

I am gonna have to read the article more carefully than the one minute I spent looking this over.

But the bottom-line looks like: meds vs. therapy, and therapy has won this one. Long-term psychodynamic psychotherapy resaerch too often does not have the data to support this type of evaluation. It is commendable for the authors to pull all of this together.

Why all the big deal about this problem of hypothetical missing unpublished studies?

Well, on the other side of the meds vs. therapy issue, many of the unpublished studies CAN be discovered and included. When they ARE included, as was done last year, it greatly influences the result to show that antidepressants have a very modest effect for treating depression (and that is being generous to say 'modest').

Here is one main reference:
Kirsch I, Deacon BJ, Huedo-Medina TB, Scoboria A, Moore TJ, Johnson BT. Initial severity and antidepressant benefits: a meta-analysis of data submitted to the Food and Drug Administration. PLoS Med. 2008 Feb;5(2):e45. PMID: 18303940.

You just won't see this "file drawer" problem addressed in studies from Big Pharma. You won't see CME on this topic. It is likely rare that physician researchers are taught the issue and methods for addressing the file drawer problem. In my training, I was, which helps me appreciate the problem. And to understand how it is one of the ways that Pharma can lie with statistics.

Meds Vs. Therapy. Meds won this one. Good night.

No comments: