Clinical Trials of Therapy versus Medication: Even in a Tie, Medication wins
Central to the idea of evidence-based medicine is that the choices made by patients and doctors to use a certain treatment should at least in part be based on scientific studies published in peer reviewed academic journals. For a patient diagnosed with a mental disorder, the choice often comes down to whether to use behavioral therapy, psychotropic medications, or a combination of the two. We think the following story will shed some light on how conflicts of interest can complicate the decision making process.
Earlier this year, The Journal of the American Medical Association (JAMA) published a study examining the efficacy of both Lexapro, an SSRI, and problem-solving therapy in undepressed, recovering stroke patients.1 The study found that recovering stroke patients treated with either therapy or medication were less likely to be subsequently diagnosed with depression. After one year, 22% of the placebo group developed depression, while only 9% of the Lexapro group and 12% of the therapy group did. Following the study's publication there were numerous articles in the mainstream media extolling the benefits of the SSRIs. For instance, in USA Today, the lead author of the study stated: "I hope I don't have a stroke, but if I do, I would certainly want to be on an antidepressant."2 And as often happens after these types of studies, the media ask another expert to interpret the study. Fox News interviewed an expert psychiatrist from the University of Pittsburgh and reported that, "he hopes doctors will start prescribing preventive antidepressants to stroke patients."3 The expert said nothing about therapy.
When we looked at the study, given that there was only a three percent difference between the results of therapy and medication, we were confused by a subtle but seemingly critical omission from the published paper. While the authors compared both Lexapro and therapy to placebo, they did not report on the direct comparison of therapy to Lexapro - surely of interest to those making treatment decisions, especially since every medication has the potential for adverse effects. In a letter to JAMA, we pointed this out. Five months later, our letter was published along with an acknowledgement from the original authors that indeed the difference between therapy and medication4 was not statistically significant. Since newspapers rarely reflect on their original coverage, the benefits of therapy for stroke patients will continue to remain a mystery to most of the news reading public.
Journals such as JAMA require authors to publish their conflicts of interest at the end of their studies. In JAMA's conflict of interest policy they state that authors should include potential conflicts from the past five years. The study in question did list several conflicts, but the list did not include Forest pharmaceuticals, the manufacturer of Lexapro. During a subsequent internet search we were surprised to learn that four years previously the lead author had been listed on the speaker's bureau for Forest.5 The omission, however innocent or mistaken, is disturbing; neither the JAMA article nor subsequent media accounts noted that the lead author had served on the speaker's bureau for the manufacturer of Lexapro. However, disclosure of the relationship would not have changed the troubling end result: A researcher with a history of being funded by SSRI makers completes a 'gold-standard' federally-funded study of post stroke SSRI use, which is published in one of the most prestigious medical journals in the world, and is given a forum in the national media to tell the general public that anyone who has had a stroke, whether or not they have been diagnosed with depression, should start a prophylactic regimen of Lexapro ...even though non-medical approaches perform just as well.
In addition to the lead author having an undisclosed conflict with the makers of Lexapro, a simple internet search also revealed that the expert who was subsequently asked to interpret the study for two different news outlets had been receiving money (https://mail.lmunet.edu/exchweb/bin/redir.asp?URL=http://www.psych.org/M...) in the form of research grants from Forest since 2004, yet neither outlet reported his conflict. It is certainly true that having a conflict of interest does not automatically negate someone's view - even someone with a large conflict-of-interest can be correct. Yet, when facts emerge demonstrating that crucially important information was selectively omitted by researchers with conflicts-of-interests, it is hard to not question whether the conflict was the root of the problem.
The medical community strives to make decisions based on evidence, but as this case illustrates we have unfortunately arrived at a point where taking the conclusions of clinical trials at face value is apparently a sign of naivette. Conflicts of interest may play an important role in the reporting of scientific findings. The problem is not limited to just a couple of isolated cases but involves the entire culture of medicine that has developed over the past ten years. Put bluntly, the scientific machinery is broken. There is no easy fix, but surely patients deserve better.
1. Robinson RG, Jorge RE, Moser DJ, Acion L, Solodkin A, Small SL, et al. Escitalopram and problem-solving therapy for prevention of poststroke depression: a randomized controlled trial. Jama 2008;299(20):2391-400.
2. Elias M. Study: Antidepressants help stroke victims. USA Today 2008.
3. Reporter. Study suggests antidepressants for stroke victims. FOX News 2008.
4. Lacasse J, Leo J. Escitalopram, problem-solving therapy, and poststroke depression. Jama 2008;300(15):1757-8; author reply 58-9.
5. Robinson RG, Zorowitz RD. Pseudobulbar affect and stroke. Stroke: Clinical Update 2005;15(January/February 2005):1-4.
Note added at 0035 GMT on 6 March 2009 by BMJ after posting on behalf of the authors:
We are fully aware that JAMA is concerned about conflicts of interest and has taken a leading role in promoting policies to benefit the medical community. We are pleased to report that we learned at the end of business on Thursday (3/5/09) that the JAMA Editorial Staff has looked into this matter and will be discussing it in the forthcoming March 11 issue.
Competing interests: No competing interests