Who wrote that article?

Abstract

The latest revelations about ghost authorship of journal articles are truly frightening

Authorship issues are a common obsession of medical journal editors. We fuss about them a great deal, fretting about who contributed what to a paper, who was responsible for the work and its conclusions, and what should qualify a contributor to assume the august title of “author.” The quantity and, to a lesser extent, the quality of authored publications have a lot to do with who gets promoted in academia, who gets tenured, and who gets jobs at prestigious universities. So naturally there is a great desire among academics to get their names on as many papers as possible, preferably at the head of the (often lengthy) list of authors.

I don’t think anyone in the outside world cares much about all of this. It’s easy to make fun of the competition, quibbling, and controversy surrounding authorship by reminding everyone of the old saw that the reason the politics are so intense in academic medicine is that the stakes are so low. After all, who is really harmed if a few old men’s names are gratuitously added to a list of authors or if a research assistant’s contributions go uncredited?

The stakes are raised substantially, though, when the drug industry becomes involved. In support of their products, drug companies sponsor carefully orchestrated campaigns to pass off sympathetic, if not biased, research and review articles as the work of academic scientists rather than of their own or contracted employees. Ghost authorship takes on a new meaning when health communication companies write papers on contract, recruit prestigious authors for them, and then disappear from view. Former editor of the BMJ Richard Smith, in his 2006 book The Trouble with Medical Journals, recounts a story of an employee of one such firm bragging that she was a leading author of articles in prestigious medical journals despite never having her name appear on the papers.

I knew that ghostwriting had been documented at rates of between 6% and 15% of various types of medical journal articles and that “gift authorship” is even more common. But I never really appreciated the depth of the problem nor its pervasiveness until two weeks ago. The litigation surrounding the drug company Merck and its pain reliever rofecoxib (Vioxx) led to the public release of millions of company documents. In a shocking case study of about 250 of these documents published in JAMA (2008;299:1800-12) on 16 April Joseph Ross and colleagues matched up drafts of company authored research articles and contract authored review articles with the subsequently published papers (BMJ 2008;336:849 doi: 10.1136/bmj.39553.344965.DB). Comparing the lists of authors on the draft and final versions of these almost 100 papers is chilling.

Although we’ve known about these practices for years it’s creepy to actually see the title pages of succeeding drafts of articles, firstly with only drug company authors and then with one to three recruited academic authors’ names magically appearing at the top. In the series looked at in the JAMA article, the first author of 16 of 20 reviewed research articles changed from a Merck author to an external, academically affiliated investigator between the draft and the published version.

The findings for review articles are even worse. At least in the research studies the Merck authors’ names remained on the paper. Most of the review articles in the study were ghostwritten by contract communication companies and “edited” (for an honorarium) by external academics who were then given sole authorship. The JAMA article reprints an email from a contractor in which the progress of eight review articles is communicated to Merck staff, complete with article title, intended journal, and intended (external) author. Seven of the eight review articles were subsequently published. Also reprinted is a contract showing that the going rate for one of these review articles in 2001 was just under $24 000 (£12 000; €15 400).

It is not being overly dramatic to say that public trust in clinical research, in the medical journals that report research, and in medicine in general is at stake here. Unless we can be reasonably certain that research results and review articles are unbiased we can’t know whether doctors are giving—and the public is getting—correct advice and care. And unless we know who wrote the articles and paid for the research we can’t completely assess the work for bias.

What can be done? Scandals like this will presumably help discourage such practices. Proper disclosure of research funding and authorship helps a lot, allowing readers to know who is behind the work. External statistical input and impartial peer review help prevent and identify problems too.

But do we need to go further? Since so much drug research is funded by the drug industry, it is unlikely that a prohibition of such funding is feasible. But maybe journals should refuse review articles from authors with any support from the industry. This was done by some journals a few years ago, but the policy was reversed.

Unfortunately, no matter how many rules and regulations are in place, this is ultimately a matter of trust. Liars and unscrupulous companies will continue to get away with these practices some of the time despite everyone’s best efforts. A depressing thought.

It is not being overly dramatic to say that public trust in clinical research, in the medical journals that report research, and in medicine in general is at stake