News

More than 20% of articles have a “guest” author, study shows

BMJ 2009; 339 doi: http://dx.doi.org/10.1136/bmj.b3783 (Published 15 September 2009) Cite this as: BMJ 2009;339:b3783
  1. Fiona Godlee
  1. 1BMJ

    At least a fifth of articles published in medical journals are likely to have a guest (or honorary) author, and journals are not doing enough to tackle the problem, say two studies presented at the Sixth International Congress of Peer Review and Biomedical Publication in Vancouver last week.

    A guest author is someone who has not contributed sufficiently to the work but whose name is included in the list of authors. In a survey of corresponding authors of nearly 900 articles published in high impact general medical journals in 2008, 20% of respondents admitted that their paper had at least one guest author. In addition, nearly 8% admitted that their article had at least one ghost author—someone who had written the article or otherwise contributed substantially to the work but was not listed as an author.

    The percentages of ghost or guest authors differed between the six journals studied and different types of articles. Both types of authorship misconduct were higher among research articles than in review articles and editorials.

    Compared with a previous similar survey in 1996, rates of ghost authorship had fallen slightly, from 11.5% to 7.8%, but rates of guest authorship were unchanged.

    Responding to the study, Doug Altman, professor at the Oxford Centre of Statistics in Medicine and chief statistics editor at the BMJ, said that as these rates were based on self reporting they almost certainly underestimated the true picture.

    Ginny Barbour, chief editor at PLoS Medicine, which was included in the study, said, “This is dishonesty, and we shouldn’t tolerate it. Editors and readers are clearly being lied to on a daily basis.”

    Annette Flanagin, managing editor at JAMA and one of the study’s authors, said that she had hoped to find an improvement over time. “The fact that we didn’t is a serious cause for concern,” she said.

    Delegates at the meeting had little to offer by way of solutions to the problem other than greater vigilance among journals and naming and shaming of authors, institutions, and drug companies found to have been involved.

    In a separate study also presented at the congress Lisa Bero and Jenny White from the Institute of Health Policy Studies in San Francisco, California, found that journals with less stringent policies in tackling guest and ghost authorship were more likely to fall victim to drug companies using these practices to market their drugs.

    From documents released by Pfizer and Parke-Davis, as part of a 2004 legal settlement for the off-label promotion of gabapentin (Neurontin), the authors were able to identify 24 articles and letters for which the company had selected guest authors and target journals.

    The study’s authors found that 13 of the articles were eventually published, six in the target journals and seven in alternative journals. Four of the articles were published in supplements to the main journal.

    None of the published articles disclosed funding from Pfizer or Parke-Davis or gave the names of the medical communication company they were using. Although all the target and alternative journals had policies on their websites requiring disclosure of conflicts of interest, few of the policies specifically addressed the possibility of guest or ghost authors.

    Of the 24 target journals the six that published one of the articles were found to have less stringent policies than the 18 that did not.

    Professor Bero said that although lots of the journals had policies, many weren’t explicit enough to capture guest or ghost writing.

    She said, “We need to hit authors hard with explicit questions. Was this ghost written? Were you or anyone paid to write this article? And this has to be done universally by all journals so this sort of stuff doesn’t get through.”

    Notes

    Cite this as: BMJ 2009;339:b3783