How important can a mediocre scientific study be? Extraordinarily important, even decisive — if it appears in the right place.

JAMA, the lead scientific outlet of the American Medical Association, has just published an overview of the "Effectiveness of Long-term Psychodynamic Psychotherapy." It concludes that for a variety of what it calls complex mental disorders — including personality disorders, eating disorders, and chronic anxiety and depression — long-term pscyhodynamic psychotherapy (LTPP), is a much more effective treatment than short-term psychotherapy.

The hitch is that there have been few rigorous studies of LTPP. Partly, the problem is that psychoanalysts and other psychodynamic practitioners have been unenthusiastic about intrusive, coarse-grained research. Partly, the problem concerns logistics. If you want to test a treatment that involves, say, 300 sessions over three years, what do you do with the control group? Teach them tennis or meditation for 300 hours? Who pays for this sort of study, and what kind of academic career can be built on such slow research? And then, medications muddy the mix. Over time, control patients may end up on antidepressants; because antidepressants work, they create a difficult standard against which to measure talk-treatment effects. Finally, there are problems with consistency. What studies we have tend to reflect the field's theoretical rather then society's economic interests - comparing one long-term treatment to another rather than contrasting LTPP with a less expensive intervention. Altogether, those trials that have been carried out bear little relationship to one another.

Ideally, a study justifying long-term treatment would compare LTPP with a short-term intervention that is sustained by a normal sort of follow-up, say, twelve sessions of cognitive therapy to start, with monthly refreshers for a year and then bimonthly refreshers for six months. Few if any such trials exist. Worse, there are almost no studies in which the main treatment is carried out by team that has no allegiance to it; reseach on short-term therapy shows that allegiance effects, presumably resulting in subtle biases, are strong.

For a grand overview, the JAMA integration includes only a small number of patients: 1053 in all. And that level was achieved by counting studies (of psychodynamic therapy) that have no comparison condition, so there are only 257 control cases. And then, the short-term therapy represented is reasonably extensive: 33-plus sessions over 39 weeks, on average — versus, for LTPP, 102-plus sessions over 53 weeks. So the issue is as much intensity as duration. Some of the control conditions are psychodynamic, and some are not. And there was no attempt to shadow the tested treatment with a less specifically active one, sports or relaxation.

In sum, it is not clear that the JAMA researchers, Falk Leichsenring and Sven Rabung, had the raw materials to do the task required for public health purposes, comparing longer and shorter treatments. In many cases that Leichsenring and Rabung (they are researchers at a German university and hospital) analyzed, the comparison intervention was no treatment at all. That said, the results are striking. The analysis found that the effect size for long-term treatment was 1.8, which means that after treatment 96 per cent of patients in the long-term therapy were better off than the patients in the comparison groups. Long-term treatment works very well - probably.

So: not much to go on, but the results are suggestive, and they are the best we have. And then, there's the brand label: JAMA. The study passed a rigorous peer review. There is no comparably prestigious study denying that long-term therapy is the treatment of choice. Moreover, the positive result agrees with expert opinion of the relevant specialty, psychiatry, and with patient experience, as documented, for instance, in a widely circulated Consumer Reports survey.

What makes the JAMA study particularly important is the new mental health parity law. The expert consensus, bolstered by now by uncontroverted medical evidence, should make it harder for insurers to deny extensive psychotherapy. And note what we're talking about: not a limping-along approval of four sessions this month and, depending on how those go, four sessions the next. No, the treatment involves an advance commitment to methods that are understood to take one to three years to work. The treatment is not this or that brief set of sessions, it is long-term exploration of the unconscious, transference, and the rest of the targets of psychoanalysis and its progeny.

I don't say that subscribers will get this result — non-intrusive insurance coverage for serious psychotherapy - but when they don't and they appeal, arbitrators or judges should rule in patients' favor. If insurers don't like this state of affairs . . . then they should underwrite the more targeted research (for purposes of public health policy) that should have been undertaken already. For now, it's good for patients and good for the field that the weightiest evidence available favors what most thoughtful observers believe already.

Of course, there is a certain dumb luck in this result. Thinking about this stroke of fortune raises important questions about what standards should have prevailed, absent JAMA's timely publication of an assiduous but paper-thin study. What should count as evidence, and how should we set policy regarding reimbursement? I hope to take up those issues in a future posting.

Bonus addendum: I have reported here that the financial bailout is really a mental health parity bill. It turns out that it is also a bike-to-work act!

Most Recent Posts from In Practice

What Bill Taught

W. P. Kinsella believed in output, and the man could write.

Why Doctors See Antidepressant Effects That Researchers Miss

Good results that drug trials consistently hide.

Stronger Drugs, Stronger Placebos

New research on placebo responses improves our understanding of drug efficacy