The Shrink Tank

I read your post and Coyne's. Sorry, he wins hands down. What's with starting by saying that psychology research isn't worth much, then arguing that the research supports psychodynamic therapy?

Your claim that when compared to other treatments, LTPP does not show superiority or inferiority, is incomplete. In answering the question, "is LTPP worthwhile?", I see several further areas to explore regarding the lack of distinction between treatments.

1.The first are, and directly speaking to cost-effectiveness, if truly LTPP and other treatments show similar outcomes, wouldn't the preferred first treatment approach be the shorter, cheaper one? It seems that starting with such a treatment allows more people to recover quicker, and even if they don't respond to the treatment, LTPP is still an option. I would much prefer to start with a shorter option (let's say 16 weeks of CBT) and then switch to LTPP if it doesn't work then start with LTPP only to find no response after a quite a bit of time and then switch to a shorter treatment. When treatment outcomes are similar, then factors like cost, time, risk I think should inform clinical decision-making in treatment selection.

2. All of these meta-analyses and critiques of these meta-analyses mentioned mention that yes, there just isn't enough quality evidence on LTPP. Critics are not complaining that there isn't enough poor quality studies to conduct strong meta-analyses (as your Woody Allen quote suggests), but saying that because there aren't quality studies to begin with, meta-analyses (supporting or not supporting the effectiveness of LTPP) don't have a lot of resonance. The new analysis that Coyne mentions is rigorous in its methodology, but for all of the long term psychodynamic therapy that goes on in the world, it is shocking that so few quality studies have been done to be included in a rigorous analysis. To use your metaphor, researchers want better food and bigger portions in order to do more useful analyses.

3. Along similar lines, it is difficult to show differentiation of effects between LTPP and other treatments when there is little data and poor data on LTPP. The studies that directly compare LTPP to other treatments often have small sample sizes where only a huge difference of effect size between treatments are detectable in the first place. Not detecting a difference in such a study does now show equivalence, just that the effects between them isn't huge. There could be large differences between treatments, but studies being conducted cannot detect them. Analyzing the between group effect sizes of these studies in a meta-analysis and still finding little difference does not add credence the equivalence argument. (A similar argument but to the other extreme can be made about the effectiveness of antidepressants for treating depression. Studies have very large sample sizes that can detect small differences between drug and placebo groups, but the differences are too small to be clinically meaningful.) This does not mean that LTPP isn't effective for many or that it isn't equivalent or even better than other treatments, only that much of the research doesn't answer this question (and it really should be able to.)

4. Lastly, different differences (not very eloquent) might occur for different populations of people who seek treatment. There might be stronger evidence for using LTPP for treating certain populations with certain issues and another modality for treating other people with different problems. Ultimately, it is great for people seeking help when there may be many ways for them to feel better, as not every treatment will work for everyone. It is important to test treatments to see effects for various population/clinical issues (ex. diagnoses) to make the best decisions about choosing treatments. When many LTPP studies are inclusive of many diagnoses, the mean effects often get washed out, leading to more false equivalence.

I want for all sorts of treatment options to show that they are helpful for people in need and for those treatments to be accessible. However, I don't believe we do anyone any favors by not conducting quality research that evaluates meaningful outcomes of various treatments. Psychotherapy research is messier than perhaps some other experimental research, but that does not mean it can't or shouldn't be done, or that the results cannot be informative in clinical decision-making. A lot of people find themselves distressed/impaired, and only a minority actually seek care. With costs of care high and treatment outcomes uncertain, this does not surprise me. If our goal is to provide psychological treatments for people in distress that make their experience better (and I say this acknowledging that psychotherapeutic interventions may not be all that effective of a way to deal with the widespread poor mental health in the US), then we need to do far better research to find ways we can treat folks with the best interventions possible for them, in a way that is affordable, and so that people who need support can access it.

@Thinker. I did not read that from his post AT ALL! Doing psychotherapy research on MBT and seeing the many, MANY, pitfalls of comparing psychotherapy to an experiment myself, I find his opening remarks completely necessary to be familiar with for anyone who wishes to be taken seriously at the top level of psychotherapy research. Expecting clear findings from RCTs which can settle any debate is futile. It also needs careful process research to determine what actually great THERAPISTS do in their work. For instance: In one study of CBT vs. short-term dynamic therapy done in Norway, the researchers found similar effects of CBT and STDP (STDP actually doing a little better) on depression. However, in the CBT group, one (1!) therpapist was responsible for almost all the positive outcome. And they had vast process data suggesting that the active ingredient was number of emotional breakthroughs (which is the postulated mechanism of change in STDP and NOT CBT). This goes to show exactly why health care consumers should not blindly follow the "RCT for diagnosis dogma" in selecting therapy/therapist.

Psychotherapy is a young science and so far is extremely far away from determining with any strong certainty what and why something works. Voices like that of mr. DeFife is absolutely necessary to bring it to a higher level of methodology and critical thinking. And as a therapist I would not take comfort in doing anything "evidence based" alone. The evidence should be in your very own practice from your own outcomes. I've met horrible "evidence based" therapists using their "evidence based" method as sort of a defense agains their horrible results. It matters not what you call what you do, but what you actually accomplish with your patients.

for more on a similar perspective: http://www.youtube.com/watch?v=UJlVCzKM_gA

Zimmerman,

I don't think anyone compared psychotherapy to an experiment. No practicing therapist treats clients of their work together as an experiment (at least I hope.) Practicing therapists should have their practice informed by quality research.

The study from Norway you mentioned seems great, and I'd love a citation. Understanding process is important, and identifying the interesting finding form the study can become a basis for a new study. Have two psychotherapy groups, one that focuses on leading to emotional breakthroughs and one that doesn't, and see if leads to better outcomes. In itself, though, the finding in the study from Norway needs some fleshing out before that is considered applied to practice.