Jonathan Leo, Ph.D.
Over the past several years, I have written about the potential impact of conflicts-of-Interest in medicine (COI). I have also watched how the mainstream media reports the results of medical research with great interest. As a neuroanatomist, I was particularly interested in a recent study published in the Journal of the American Medical Association (JAMA), which reported that the prescription of escitalopram lowered the rate of depression in patients who had recently suffered a stroke. I co-authored a letter to JAMA with Dr. Jeffrey Lacasse of Arizona State University, in which we pointed out the problem of selective reporting within the study: Problem-solving psychotherapy was statistically equivalent to escitalopram, but this was never mentioned. In the mass media, the principal investigator recommended that all stroke victims should be prescribed antidepressants without mentioning that problem-solving therapy was statistically equivalent.
In the process of reading this body of research, my co-author and I noticed a problem with the COI disclosures in this positive trial of escitalopram published in JAMA. The principal investigator on the project, Dr. Robert Robinson, had not declared that he had previously received funding from Forest Laboratories, the makers of escitalopram. The financial relationship with Forest Laboratories was well-documented and easily discoverable via a Google search, as evidenced by Dr. Robinson’s previous self-disclosures in varied sources such as here, here, here, here, and here.
I urge my students to carefully consider COI when they read an article as such conflicts can be important in how one interprets research. I do not consider myself a whistleblower, but I do think that the full story behind clinical trials should be transparent so that patients and doctors can make informed decisions. This instance of unreported COI in a gold-standard study published in JAMA seemed like a perfect case study of problematic COI issues. First, I informed JAMA of what we had found. We then co-authored a commentary describing this saga and the potential implications, linking together the unreported COI, selective reporting in JAMA, and in the mass media. We submitted it to the British Medical Journal (BMJ), who agreed to publish it as a Rapid Response on their website.
Months later, before publishing the BMJ letter, I called JAMA editorial staff twice to talk about the upcoming piece, and I also sent an email. Given the topic of the letter, I wanted to give JAMA an opportunity to correct any factual errors. No one from JAMA responded to my phone calls or email. The letter underwent legal review at BMJ and was approved for publication. We then published what we thought of as a fairly unremarkable letter primarily of interest to researchers who study COI.
The immediate, aggressive response from JAMA has been well documented. Threatening phone calls, personal attacks, and emails were accompanied by demands that we retract the entire BMJ letter.
Our letter was published without any negative commentary regarding JAMA itself, and included the following statement: ≥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.≤ Our letter did not attack JAMA and, if anything, presents the facts of the matter flatly.
JAMA continued to ask that the entire piece be retracted. We were stunned by the continued, heated reaction from JAMA, and were concerned that we might have inadvertently made an error of fact in our letter. I made the following offer repeatedly: Please tell me what we have written that is factually incorrect, and if we have made a mistake or misinterpretation, we will retract the entire piece and issue a public apology. No factual errors were ever pointed out. I remain confused as to why JAMA felt they could demand that we retract an accurate letter.
Phone calls were followed by a scathing JAMA editorial which pinpoints my actions as the cause of the problem- with no mention of the culpability of the researchers who failed to disclose their COI, or the fact that the undisclosed COI was unearthed by a 5-minute Google search.
JAMA now insists that it was inappropriate to disclose the COI while they were conducting an investigation. This is curious, for several reasons. First, their investigation was complete by the time our letter was published. Second, the undisclosed COI information contained in the article is publicly available on the Internet (again, here, here, here, here, and here). JAMA has never clarified how the re-publication of publicly available information after the fact could interfere with a completed investigation. This investigation, which took five months, resulted in a short correction published in JAMA, along with a letter from the authors apologizing for their lapse of memory resulting in undisclosed COI. JAMA has claimed that the result of their investigation was more comprehensive than our BMJ piece. I only ask that readers actually compare the material published in JAMA with that published in BMJ. The material published in JAMA does not include any analysis of the context or potential implications. I believe our BMJ letter presents a more complete (and troubling) story.
Importantly, I am under the impression that JAMA objected not to the timing of the publication of the letter, but to us publishing the letter at all. In their most recent editorial, JAMA seems to assert that they have some right to control the publication of publicly available information outside their own medical journal. I do not believe they have any such right. It would seem to be an infringement of academic freedom to threaten academics who analyze publicly stored information. This information was available to anyone with access to the Internet. The view that JAMA should control such information is anachronistic at best. At worse, it is a reflection of a scientifically and ethically inappropriate effort to suppress the free exchange of information, which is at the heart of productive scientific discourse.
The implications of the JAMA’s reaction to our letter are significant. For instance, the pharmaceutical industry is often criticized for their impact on evidence-based medicine. In the past, I have criticized direct-to-consumer advertising of psychiatric medications, which is not helpful to Big Pharma. However, I have never been telephoned or threatened by representatives from Big Pharma. In contrast to my experience with JAMA, any exchanges have been civil and appropriate.
The claim that JAMA can control the flow of information in the public record should be considered by bioethicists and other academics who study the process of medical research and publication. In my opinion, this claim has shifted this issue markedly. What began as a short (and potentially obscure) letter about undisclosed COI has now led to questions about the limits of institutional authority in the medical publishing industry, the extent of academic freedom, and even the role of the First Amendment.
Competing Interests: None.
Acknowledgement: Jeffrey R. Lacasse, Ph.D., provided editorial assistance in the preparation of this letter.