VANCOUVER, B.C. Journal articles often list a long string of putative authors. I once counted 47 on a physics paper. But where journal articles in the natural sciences often appear overly conscientious about acknowledging all contributors, the opposite has become a nagging problem in biomedicine. Here, not all authors on a research project – or even, necessarily, the most important ones – may be identified as a contributor.
The existence of these ghost authors, as they’re called, evoked frustration and anger yesterday in a large share of the 400 journal editors and clinical research scientists taking part in a quadrennial international workshop on peer review and biomedical publication. This year’s venue: Vancouver, British Columbia.
The concern, speakers complained, is that ghost writers can be little more than hired guns. The expectation is that they will describe research findings in ways that aid their benefactors – drug or biomedical device companies that wish to avoid the appearance of directly influencing a clinical trial’s interpretation.
Often aiding and abetting the situation is another group of poseurs known as honorary authors. Their names show up on papers – despite their having done nothing – to lend credibility to research, to reward a friend of someone connected with the study, or to help hide the contributions of a ghost.
Both types of fraudulent authors have been around for a long time. But what disturbed the legions of editors meeting in Vancouver is that despite increasing efforts to rout these cheaters, the share of articles they influence seems to have changed little over the past decade.
For instance, Joseph Wislar and his colleagues at the Chicago-based JAMA — the Journal of the American Medical Association — surveyed corresponding authors of 630 papers published last year in major medical journals: Annals of Internal Medicine, JAMA, Lancet, Nature Medicine, the New England Journal of Medicine and PLoS Medicine. In hopes of getting honest answers, the JAMA team offered authors anonymity by not linking their responses to particular papers.
A dozen years ago, Wislar noted, a similar study turned up evidence that honorary authors inflated the ranks of roughly one-in-five biomedical research papers, ghost authors contributed to nine percent of surveyed papers and two percent of the published articles analyzed had both ghosts and honorary authors.
Many top journals have attempted to crack down on such author fraud in recent years by asking whoever submits a manuscript to confirm that no one has contributed to the paper beyond those named and that everyone listed as authors contributed substantially. Editors at the New England Journal and PLoS Medicine, for instance, yesterday noted that they have such policies.
Despite this, Wislar reported, rates of author fraud in the 2008 papers that his team has probed show little improvement from the earlier analysis: 21 percent of papers still had honorary authors, eight percent had ghosts and two percent still had both.
Xiu-yuan Hao of the Chinese Medical Association Publishing House in Beijing reported finding the same types of author fraud in the Chinese Medical Journal. The first authors of 220 papers that had been published during the last four months of 2008 responded to his team’s questionnaire. These data indicate that 10.4 percent of the papers had ghosts and 28.6 percent had honorary authors (mostly heads of departments or other institutions).
Jenny White and Lisa Bero of the University of California, San Francisco, analyzed internal company records from Parke-Davis that had been unearthed as part of a lawsuit involving its drug gabapentin (Neurontin). Developed as an anticonvulsant, this medicine is also used for treating some types of intractable pain. The lawsuit argued that Parke-Davis wanted to encourage doctors to prescribe the drug for uses – like migraines and bipolar disease – and doses that had not been formally approved by the Food and Drug Administration.
Some 900 company documents have been archived at UCSF. White and Bero correlated some of these documents from 1996 and 1997 with journal articles published shortly thereafter on new uses for the drug. White reported yesterday that they detailed evidence that Parke-Davis had plotted to get medical writers to place research papers on benefits of the drug in certain target journals. The company documents included records of payments to a medical-writing company it had hired to do this, she says.
Other documents identified huge sums that had been paid to one scientist for speaking engagements and other consulting tasks. This scientist ultimately ended up being the sole author of one notable paper on the drug. And probably an honorary author at that, White says, because the industry documents suggest a ghost author had drafted the paper.
Contends White: “Parke-Davis was really very aggressively marketing gabapentin back in the ‘90s.” Documents in the UCSF archive indicate that the company was attempting to get research data published that it could then use to help sell doctors on prescribing its drug even more.
“That was the primary objective of getting these prestigious ghost authors to put their names to the articles,” White says. In essence, she contends, Parke-Davis’ goal “was really publication as marketing. Which distorts the scientific record and was very harmful, ultimately, to science and how people make treatment decisions.”
The data presented yesterday riled up a number of journal editors. “This is completely outrageous,” said Ginny Barbour, chief editor of PLoS Medicine in Cambridge, England. It suggests that the authors of a nontrivial number of papers – even in the top tier journals – “have been lying to us.”
For instance, the JAMA team reported yesterday that authors of papers in PLoS Medicine – which has always prohibited ghost and honorary authorship – continue to flaunt the rules. Any author who goes unacknowledged “is presumably someone whose work was paid for by a company,” Barbour says. And trying to hide that fact, she maintains, immediately “throws into substantial doubt the whole study.”
In an editorial published in PLoS Medicine this week, Barbour noted, “We’ve said that if we find a published paper has been ghost written, it should be retracted.” Employers of the authors on that paper should be informed of the fraud, she added, and the entire string of authors on the offending paper would likely be barred from publishing in her journal for some as-yet-unspecified period.
What’s especially troubling about the JAMA team’s new report, Barbour adds, is the nature of papers now being ghosted. In the past, she says, there’s been almost a tacit acceptance of it by journals because it was supposedly restricted almost exclusively to review papers. “But we now find it’s occurring in original research as well,” she observes, “where it poses a very serious threat to the integrity of the medical literature.”
Indeed, the new JAMA analysis found: “Ghosts were more prevalent in research articles (12 percent) vs reviews (6 percent) and editorials (5 percent).”