Elizabeth Loder reports on a panel discussion held at the International Publication Planning Association meeting in St Louis, Missouri.
In mid-February I travelled to St. Louis to represent the BMJ at the Midwest meeting of the International Publication Planning Association. Along with Brian Mandell , editor-in-chief of the Cleveland Clinic Journal of Medicine, and Patricia Baskin, the executive editor of Neurology, I participated in a panel discussion moderated by Susan Siska of Abbott Pharmaceuticals. The title was “How journal editors view industry-sponsored publications.” Ms Siska invited each editor to briefly describe their journal and then to review the top three matters they encountered in dealing with industry publications. This was followed by a question and answer session which evolved into a lively discussion of the thorny question of authorship. The debate centered on the question of “ghost” authors; in particular, medical writers paid by the pharmaceutical industry who may prepare a manuscript’s first draft, among other things, but who traditionally have not always been listed as authors when the paper is published.
The Cleveland Clinic Journal of Medicine publishes only review articles, and has just instituted a new policy regarding industry involvement in a review’s preparation. Until February of this year, the journal published such articles, with the rationale “that when a new drug or device was introduced, it was those physicians most familiar with the drug or device and the raw data who could best discuss its use.” Over time, though, the journal’s editors became worried about their ability to detect subtle bias or spin, as well as the “burden this places on our peer-review process.” A new policy was introduced: “We will no longer consider for publication in regular monthly issues of the CCJM any review article in which a manufacturer played a part in the preparation of the manuscript, either directly or through payment to the author or a surrogate. We will continue to consider articles written by experts who declare relationships with companies, such as accepting research grants or serving as consultants. Such experts often know the most about new devices and drugs. Still, we feel we must draw the line at a direct pharmaceutical company role in the development and preparation of articles. We will continue to alert our readers to all potential conflicts of interest or bias.”
Neurology, on the other hand, has taken a different approach. The journal publishes original research as well as review articles. Its leadership did not find satisfactory the International Committee of Medical Journals (ICMJE) approach to authorship, which Ms. Baskin described as specifying who deserves to be an author. Rather, Neurology seeks to identify as an author anyone who has influenced the content of the paper, including any paid medical writer. The rationale: “It seems self-evident that a first draft of a manuscript generally serves as the intellectual framework of all subsequent revisions and modifications. We at Neurology think that the author of the first draft of a manuscript should be recognized. ”
What are the merits and drawbacks of these two policies? In the case of the Cleveland Clinic Journal of Medicine, it is likely too soon to draw any conclusions. In my view, excluding direct pharmaceutical company involvement in the preparation of review articles is probably a step in the right direction. One wonders how much of the actual grunt work involved in reviews – the formulation of the search strategy, selection of articles, and data abstraction – academic authors really do in such cases, when it is so easy to rely on the hired medical writers.
On the other hand, such exclusion does not eliminate the possibility of bias or spin, and probably will not greatly lighten the load on peer reviewers. Some of the most egregious examples of bias in the preparation of reviews or research reports, in my experience, stem from academic conflicts of interest. These are more difficult to identify and to root out because doing so is not just a matter, as it so often is with pharmaceutical industry connections, of “following the money.” Consequently, this more pernicious form of bias receives little attention.
The approach taken by Neurology is equally interesting, particularly notable for its novel egalitarianism. As with many well-intentioned schemes, however, execution may prove challenging. I can envision a number of possible problems, including what should be done when someone instrumental in designing and carrying out a study does not agree with the way results are reported or interpreted, and declines to be an author (read more). And then there’s the fact that editors and peer reviewers often have substantial influence on the content of a paper. Editors in particular sometimes require the inclusion or omission of particular information or insist that authors change conclusions that are not firmly supported by the data contained in a paper. Often this means that the main message of the paper changes substantially. Does that make us authors, too?
In the end, these two new approaches to the persistent problems of authorship (What does it mean to be an author? Who “deserves” to be one?) illustrate a troubling lack of consensus. One thing is clear, though: the definition of “authorship” is a work in progress, and the world view of the International Committee of Medical Journal Editors is not universally shared.
Elizabeth Loder is the BMJ’s US based clinical epidemiology editor.