Scientific Discourse, Corporate Ghostwriting, Journal Policy, and Public Trust

Advancements in science, clinical care, and medical education require a discourse among basic and clinical scientists, clinicians, and medical educators. Peer-reviewed journals such as Journal of General Internal Medicine (JGIM) serve a critical service by providing a medium for such discourse. To be most effective in advancing medical science, patient care, and education, published articles must have relevant content that pushes back the interface between what is known and what is yet to be discovered. Articles' content must be based on high-quality reproducible methods. Ideally, each article should be completely unbiased and widely generalizable. However, perfection is an asymptote, a goal that can be approximated but never reached. As journal editors, we work hard to identify, improve, and publish manuscripts with the highest quality and widest possible relevance to our readers.

We rely heavily on the peer-review process to elevate the level of discourse presented in the Journal's pages by culling the best manuscripts from among the hundreds we receive each year. In most cases, manuscripts that are rejected either are not appropriate for our readers, have important flaws in their methods, or are insufficiently new and interesting to compete for the relatively few pages in which we can publish articles. Sometimes, our reviewers identify a fundamental problem that not only affects the fate of an individual manuscript but highlights a problem we need to address from the perspective of Journal policy.

In this issue, Fugh-Berman describes an egregious case of unethical behavior by an author, a pharmaceutical manufacturer, and a medical education company. In an attempt to promote its new oral anticoagulant, this drug company commissioned a medical education company to perform a review of the literature highlighting potential adverse interactions between warfarin and herbal products and nutritional supplements. We received the manuscript and assigned it to a deputy editor who judged that it was of sufficient interest to JGIM's readers to send it out for peer review. One of the reviewers was Dr. Fugh-Berman, who is an expert in herb-drug interactions. As it turns out, Dr. Fugh-Berman had been approached by the medical education company to “author” a draft manuscript that was very similar to the one we sent to her for review. Indeed, it was the same manuscript, although the version submitted to the Journal had been added to and edited by the author. Knowing the literature, Dr. Fugh-Berman found the review to be highly biased against warfarin, including the title that warns of “hidden dangers.” Dr. Fugh-Berman wrote:

[This manuscript] consists of collated case reports with little to no assessment of case report quality. It is transparently biased. Controlled trials assessing interactions are only mentioned spottily; the writers seem unaware of many more trials that exist. When trials are mentioned, it is dismissively, as if the case reports showcased were just as credible. The fish oil section is particularly egregious. There is only a single case report that I know of that reports a possible interaction. Two clinical studies specifically designed to test fish oil with warfarin show no interaction, and long-term studies of fish oil show no effect on bleeding parameters. PC-SPES didn't interact with warfarin. It was ADULTERATED with warfarin. I could go on with other errors and misleading statements ….

Dr. Fugh-Berman provided us with a copy of the original manuscript and cover letter sent to her by an education company offering her the opportunity to put her name on the manuscript as its author. She was asked to edit it as she saw fit after which the manuscript would be submitted to a journal read by “pharmacists, anticoagulation nurses, and general practitioners.” No compensation was offered to Dr. Fugh-Berman; her reward for putting her name on this manuscript would be the final publication. It is important to note that one or more persons from the medical education company had performed the original literature review and had written a full manuscript, including a title page (with Dr. Fugh-Berman's name on it), tables, and references. A cover letter indicated the target audience and clearly stated that the paper had been commissioned by a specific pharmaceutical manufacturer. Yet the manuscript received by JGIM said that the author had received “editorial assistance” from the medical education company and that the work had been supported by “educational funding” from the pharmaceutical manufacturer.

We had no problem with the manuscript being commissioned by a pharmaceutical manufacturer or that someone from a medical education company had performed a literature review and written the draft manuscript. However, there were two substantial problems. First, the contribution of the manuscript's original author(s) was not recognized by coauthorship or a note in the Acknowledgements, some indication that the person or persons employed by the medical education company were taking direct responsibility for the work. Second, the financial relationship between the original author(s) and the pharmaceutical company was not acknowledged. Would exposing such a relationship automatically result in rejection of such a manuscript? No, but it would have resulted in closer scrutiny for potential bias in favor of the company's products (currently on the market or in development). It is important to mention that the author listed on the manuscript submitted to JGIM was not offered money in return for “authoring” this manuscript. Her reward for putting her name on this manuscript would be the final publication. The medical education company preyed upon academicians' general need to “publish or perish.”

Publishing biased literature is not simply “getting the message out” for the pharmaceutical client of the medical education company. It injects bias and untruth into the scientific dialogue in order to enhance corporate profits. The fundamental conflict of interest is between having a balanced discussion of the evidence and the pharmaceutical company's responsibility to its shareholders to sell as many pills as possible. How much is sullying the medical literature worth in market share? What are the long-term effects on scientific discourse and the bond of trust among the scientific and clinical communities, pharmaceutical manufacturers, and the public? What is the ultimate effect on market share if that trust is breached?

Not all interactions between the private sector and academia are unethical or biased. By encouraging appropriate management of acute and chronic conditions, pharmaceutical manufacturers can benefit to the degree that use of their products is encouraged by evidence-based guidelines established by independent bodies (e.g., government agencies or nongovernmental organizations such as the World Health Organization and the Cochrane Collaboration).

In this case, the peer-review process worked and worked well. This biased manuscript would have been rejected even if there had been no prior contact between Dr. Fugh-Berman and the medical education company. But it brings to the fore holes in JGIM's editorial policies that we have since filled with new policy. As is now stated specifically in our Instructions for Authors,¹ everyone who had input into the writing of a manuscript, including generating data from reviewing published articles, must be specifically named either as a coauthor or in the Acknowledgments. All financial arrangements between anyone listed as an author or named in the Acknowledgments must be stated explicitly in JGIM's Conflict of Interest Statement, which must be completed by every author before we will review a manuscript.² In addition, the World Association of Medical Editors (WAME) has in the past published a policy on authors' responsibilities in taking credit for work done and declaring conflicts of interest.³ In response to this specific incident, WAME has broadened its focus on authors' responsibilities to include those who would commission such articles and those medical education companies who agree to ghostwrite them. This new WAME policy is published in this issue of JGIM.⁴ Diligence of medical editors and reviewers will continue to be the first line of defense against the sullying of the medical literature by biased “reviews” whose motivation is commercial, not scientific. We believe that a positive dialogue between journal editors and the pharmaceutical industry would enhance the appropriate use of their companies' products and reestablish trust and cooperation among industry, academia, journal publishers, and most of all, our patients who are the ultimate consumers of medical advances.