Ghost guardians

Are ghost authors vanishing from medical journals? A new study published last week in the BMJ suggests that ghost authorship may be in decline.

Ghost authorship is the lack of acknowledgement for individuals who may have contributed substantially to the production of an article, such as a paid medical writer or statistician. Honorary authorship, in comparison, is granting authorship status to individuals who may have had little (if any) involvement in a study, such as a dean or department head. Both ghost and honorary authors create problems for establishing credit and accountability in research.

The authors of the study, Joseph Wislar, a survey researcher at the American Medical Association, along with three editors from JAMA (Annette Flanagin, Phil Fontanarosa, and Catherine DeAngelis), updated an earlier study of inappropriate authorship from 1996. Surveying corresponding authors of papers published in 2008 in the top six medical journals, as measured by their impact factors (Annals of Internal Medicine, JAMA, Lancet, PLoS Medicine, NEJM, and Nature Medicine), the researchers report on the frequency of ghost and honorary authorship by journal and article type.

In 2008, self-reports of ghost authorship was 7.9%, down from 11.5% in 1996. In comparison, rates of honorary authorship remained statistically similar over time (17.6% in 2008 versus 19.3% in 1996). Prevalence of honorary authorship in research articles was higher in 2008 than in 1996, but lower for review articles and editorials.

Surprisingly, journals that require authors to detail their contributions showed no difference from journals without such author requirements.

While this study was beautifully and rigorously executed — with a response rate of over 70% — the researchers acknowledge that respondents may not be forthright with reporting inappropriate authorship practices, especially considering the social stigma against ghost authorship. Indeed, a study of members of the American Medical Writers Association and European Medical Writers Association put the incidence of ghostwriting at 42% for 2008, down from 62% in 2005. If the incidence of ghost writing is truly declining, it still has a long way to go.

Commenting on the study in an editorial, Patricia Baskin and Robert Gross, editors of the journal Neurology, question the very definition of scientific authorship and wonder whether the definition of authorship established by the International Committee of Medical Journal Editors (ICMJE) contributes to the problem of inappropriate authorship. They write:

The ICMJE criteria leave “substantial contributions” incompletely defined and, moreover, require three conditions to be met. However, in practice as the numbers of contributing specialists increase author groups tend to ignore the ICMJE criteria, adding “honorary authors” when contributors do not fulfil all three criteria. Thus the author byline may not have accurate information about individual contributions to a study or may not reflect important contributions from people who were not named.

Baskin and Gross’ solution for the journal Neurology was to make authorship as inclusive as possible and require that any medical writer who wrote the first draft of the article or responded to reviewer comments to be listed as an author with full disclosure.

It is unknown whether alternate policies will change the prevalence of inappropriate authorship. With hope, they will make it more difficult for ghost authors to hide.

Phil Davis

Phil Davis

Phil Davis is a publishing consultant specializing in the statistical analysis of citation, readership, publication and survey data. He has a Ph.D. in science communication from Cornell University (2010), extensive experience as a science librarian (1995-2006) and was trained as a life scientist. https://phil-davis.com/

Discussion

7 Thoughts on "Vanishing Ghost Authorship"

You have misunderstood what the survey of EMWA and AMWA members shows. It does not put the incidence of ghostwriting at 42%.

What it showed was that 42% of contributions from professional medical writers were unacknowledged in 2008. That is not at all the same thing as saying that 42% of publications are ghostwritten, unless you believe that 100% of all publications are written by professional medical writers. The proportion of publications written with the help of medical writers is, to the best of my knowledge, unknown, but is surely considerably less than 100%.

The conclusions of both this study as well as the findings of the AMWA and EMWA are incomplete because they focus only on North America and Europe. Nearly 1/3 of human clinical trials (in 2011) archived on PubMed (year-to-date) are authored outside of North America and Europe. The percentage of trials conducted in Asia has increased from less than 10% in 1997 to almost 15% in 2011. Both ghostwriting and honorary authorship are common and unmeasured in Asia. The authors of this study cannot claim that ghostwriting is in decline because their sample results (88% of all respondents were from North America and Europe) were overwhelmingly biased toward western authors. Twenty years ago I suppose that one could survey western authors and extrapolate to the total population, but that is certainly no longer appropriate. At current growth rates, Asian authored research will surpass American authored research by the early 20s. The Euro-American centric focus of American researchers and editors is flawed and begins to border on racism. How could the authors conclude that ghostwriting and honorary authorship is in decline when their response results are so obviously skewed?

Upon reflection, I see that was probably a bit harsh in my assessment of this paper. The only issue I have with this study is that the authors were remiss in the limitations section of the paper by not mentioning that the response rate to their survey was overwhelmingly geographically centric in that 88% of all responses were from North America and Western Europe. Clearly the skewed response from North American and European authors would introduce an additional bias not mentioned in the limitations section. I fear I was exhibiting my own biases in my earlier response. I extend my apologies to the authors.

Our study’s results came from a stratified random sample of articles published in the 6 leading general medical journals with the highest impact factors in 2008. We achieved a 70% response rate from the corresponding authors of those journals. We are confident that the authors in our sample are representative of those who published in those 6 journals in 2008. As we reported in table 3 of the online supplement to our article, 2.7% of the corresponding authors who published in these journals came from Asia, 2.5% came from Australia and New Zealand, and another 3% came from elsewhere in the world, including Africa and South America. These data only tell us where the corresponding author is based, but do not necessarily tell us where the research was conducted or where the co-authors of a particular paper are located.

We were clear in our article about what our results represent. As we stated, “Our findings suggest that 21% of articles published in 2008 in the generalmedical journals with the highest impact factors had an inappropriate honorary author, and that nearly 8% of articles published in these journals may have had an unnamed important contributor.” In addition, as we stated in the Limitations section, because only 6 general medical journals with high impact factors were included, these results may not be generalizable to other medical journals or to authors who did not publish in these journals. We hope in future analyses to examine factors, such as geographic location of authors. In addition, perhaps our study could be replicated among regional journals or those that publish papers primarily from regions other than North America and Europe.

An interesting preliminary study presented results on the prevalence of inappropriate authorship in the Chinese Medical Journal in 2008. For those interested, see “Ghost Writers and Honorary Authorship: A Survey From the Chinese Medical Journal” by Xiu-yuan HAO, Shou-chu QIAN, Su-ning YOU, Mou-yue WANG, which was presented at the 6th International Congress on Peer Review and Biomedical Publication in 2009. The abstract is available on the Peer Review Congress Web site at http://www.ama-assn.org/public/peer/abstracts_2009.html#6.

Thank you Annette. I fear that I have done a very poor job in communicating my point, hopefully you will allow me to extend my apologies and start over.

I don’t think it necessary to study regional journals to measure the activities of authors outside North America and Western Europe. A review of the PubMed database shows that nearly 40% of all articles and 20% of human clinical research articles published (2011 YTD) in the core journals (as defined by PubMed) are authored by researchers who reside outside of the United States, Canada and the leading research centers of Western Europe. Note: All of the six journals in this study are included in PubMed’s core journal list.

My question regarding this study does not concern the 70% who responded to the survey but the 30% who did not. The authors state that 88% of the respondents to this study were from North America and Western Europe. Therefore, I would pose this question; do you think your study results are projectable to all authors in the six journals or only to authors from North America and Western Europe?

Comments are closed.