Halloween's ghouls may have been banished for another year, but a certain kind of ghost is proving hard to exorcise from the medical literature.

Concern appears to be mounting about ghost authors - writers whose contribution to academic literature goes undisclosed - who are employed by pharmaceutical companies to draft medical papers describing new products the firms have developed.

Xavier Bosch, associate professor of medicine at the University of Barcelona, said medical ghostwriting constituted a threat to public health because ghost authors could be induced by companies to underreport side effects or exaggerate the benefits of new drugs. This could influence doctors to prescribe the drugs even if they were more expensive, less effective or more risky than alternative treatments.

Such instances have come to light in several recent court cases involving controversial medical treatments. According to James Parry, acting head of the UK Research Integrity Office, growing concern about the issue is illustrated by the increased mentions of "ghost author" on Google Scholar: from three in 1999 to 29 in 2009 and 21 last year.

ÖйúAƬ

A major survey published last month in the British Medical Journal suggests that although ghost authorship has declined in recent years, it remains prevalent. The survey of the corresponding authors of more than 600 articles published in 2008 in the top six medical journals by impact factor indicates that 7.9 per cent had ghost authors - rising to 11.9 per cent for original research papers.

This compares with figures of 11.5 per cent and 13.4 per cent respectively recorded in 1996 by the same researchers, all of whom are staff on the Journal of the American Medical Association.

ÖйúAƬ

The survey also probes the flip side of ghost authorship: "guest" or "honorary" authorship, in which writing credit in papers is ascribed to individuals - often senior figures - who in reality contributed very little.

The practice has been reported in a range of academic fields, but again, opprobrium tends to centre on the senior medical researchers who lend their names to papers written by ghostwriters funded by industry. Guests are drawn either by the kudos of having extra papers on their CVs or, especially in the US, by pharmaceutical company fees.

According to the BMJ paper, "Honorary and Ghost Authorship in High Impact Biomedical Journals: a Cross Sectional Survey", the proportion of guest-authored articles only declined from 19.3 per cent in 1996 to 17.6 per cent in 2008. For original research articles, the figure rose from 16.3 to 25 per cent.

The authors think the reality is likely to be worse. Although it was made clear that responses to the survey would be treated in confidence, some of the corresponding authors who responded may have been reluctant to report dubious practices. Moreover, all six of the journals surveyed have "rigorous" authorship criteria in comparison with those of some other publications.

Proscriptions of inappropriate authorship - although sometimes only implicit - are common in the guidelines of publication ethics organisations such as the International Committee of Medical Journal Editors.

However, some observers point to the difficulty of defining authorship in a sufficiently rigorous and universally applicable way.

Meanwhile, although Mr Parry agreed that it was primarily up to journal editors to enforce authorship standards, he admitted they have "limited capacity" to do so.

According to University of Toronto law professors Simon Stern and Trudo Lemmens, journals' motivation may also be compromised by the "significant revenue" they gain from pharmaceutical firms advertising their products and ordering reprints of articles (ghostwritten or otherwise) describing them.

ÖйúAƬ

In a recent PLoS Medicine paper, "Legal Remedies for Medical Ghostwriting: Imposing Fraud Liability on Guest Authors of Ghostwritten Articles", professors Stern and Lemmens also question the will of research institutions to get tough with senior, successful faculty members who become guest authors.

The power of councils compels you?

Funders are increasingly adopting their own authorship standards. The Medical Research Council, for instance, will this month consult on tougher research-ethics guidelines that include specific prohibition of ghost and guest authorship.

ÖйúAƬ

Richard Ashcroft, professor of bioethics at Queen Mary, University of London and chair of the committee that drew up the guidelines, said that inappropriate authorship amounted to scientific misconduct, adding that the MRC was taking a "stronger line" on investigating misconduct. However, he was unaware of any grant rescinded on ethics grounds.

"We are focusing on improving good practice by scientists rather than thinking very strongly about how to punish institutions," he said.

In the US, the National Institutes of Health recently extended its definition of conflicts of interest to encompass "paid authorship" in order to send what a spokeswoman for the NIH called "a clear message" in support of "transparency and accountability in research".

In response to calls to adopt an explicit ban on ghostwriting, NIH director Francis Collins suggested that the practice would already fall foul of the agency's prohibition on plagiarism. However, Dr Bosch thinks this is conceptually misconceived. "Guest authorship approximates plagiarism because an individual's naming implies credit for work done by others. However, a plagiarist copies text without consent, whereas a ghostwriter intentionally and willingly creates text for attribution to others," he said.

Mr Parry does not regard inappropriate authorship as the most serious integrity issue facing research, and thinks attempts to stamp it out would only "drive it underground". For him and many other commentators, the best solution would be for journals to insist that each author - including ghosts - sets out exactly what they contributed. According to a recent estimate, only around 20 per cent of journals currently require such statements.

But the authors of the BMJ paper found no significant reduction in inappropriate-authorship levels among papers published in journals with such requirements.

They say journals and institutions should go further, requiring corresponding authors to endorse signed contribution statements from each author.

They also suggest that institutions establish clear procedures for the confidential reporting and objective assessment of concerns about inappropriate authorship.

Meanwhile, the clamour for more severe punishment of perceived miscreants is growing. Some observers call for publication bans, while the most red-blooded are starting to advocate legal action.

Professors Stern and Lemmens argue that guest authors could be liable to civil fraud claims from journal subscribers. In the US, subscribers would be able to team up in class-action lawsuits that could lead to "significant damages" being awarded against guest authors. This threat, in their view, could be enough to deter the practice.

Dr Bosch acknowledged that suing or prosecuting ghostwriters would be more difficult since they were just doing their job. But he said that several other legal mechanisms for punishing guest authors were being developed in the US.

"Since self-regulation has not produced results, journals may be part of the problem and governments have failed to have any significant impact, the only option may be the legal system," he said.

ÖйúAƬ

paul.jump@tsleducation.com

Additional reporting by Andy Wright

?