Ghostwriters and ghostbusters: drugmakers routinely hire writers to prepare 'research papers' touting their products for publication in prestigious medical journals, and some physicians are happy to sign their names to them--for a fee. The practice raises serious questions about the integrity of medical literature.

In the early 1990s, Troyan Brennan, a physician and lawyer at Harvard University, received a generous offer: $2,500 just to sign his name to an article that was already written and slated to be submitted to a prestigious medical journal. Immediately suspecting that a drug company was behind the deal, Brennan--who is endowed with a stubborn sense of ethics--declined the offer. He then wrote his own article, exposing the ghostwriting scam and naming the company behind it.

Brennan sent me his article when I was the editor of the New England Journal of Medicine (NEJM), and we published it right away. (1) It was one of the first to expose the practice of medical ghostwriting by pharmaceutical companies.

What is ghostwriting? In the medical context, a nonphysician writer pens a manuscript that is published under the name of a "guest" physician--a doctor who, in most cases, has been paid to lend his or her name to the article. Both the ghostwriter and the physician are at fault: The doctor gets credit (and money) for work he or she didn't do, and the pharmaceutical company can promote the drug under a mantle of scientific respectability. Both threaten the integrity of science.

A doctor who goes along with this practice violates a fundamental publishing principle--that people should publish only their own work. The practice also represents a financial conflict of interest. (2) And to the extent that the ghostwriters simply parrot the drug company's merchandising messages, they are complicit in producing biased medical information.

For all these reasons, ghostwriting damages the authenticity of medical literature and can lead to the spread of misinformation among doctors, inappropriate uses of drugs, and needlessly expensive medical care. Ghostwritten articles have promoted off-label use of some drugs; downplayed the risks of drugs for obesity, arthritis, and psychiatric disorders; and exaggerated the benefits of other drugs. (3)

Generally, two types of articles are ghostwritten: reports of clinical trials and reviews of previously published medical literature.

Many clinical research trials are funded by industry. When testing a new drug, for example, a pharmaceutical company will hire a researcher as the "principal investigator" and then engage other physicians--anywhere from dozens to hundreds--from all over the country or the world to conduct tests of its new drug. The company collects and analyzes the data, prepares a manuscript, and then presents this work to the study's principal investigator and his or her colleagues.

[ILLUSTRATION OMITTED]

In an ideal world, these researchers would then review the raw data, examine and edit the manuscript, and publish only the results and conclusions that they believed were warranted. But in the real world, it works differently. Often, the researchers lack access to the study s raw data, do little independent analysis, and perform minimal editing. They may not even have a final say in the manuscript's conclusions because of contractual arrangements made when they signed up. The resulting article may or may not be slanted to favor the company's viewpoint--and it may be difficult for a reader to tell if it is.

The second type of commonly ghostwritten article is a review of a medical subject or, sometimes, an editorial. Because these articles do not contain original data, a ghostwriter selects from already-published information to produce a clinical recommendation about the use of a drug, diagnostic test, or medical procedure. Given this preselection process, it is easy for the writer to introduce bias by ignoring any studies that might conflict with the company's viewpoint. The bias may not even be introduced by the writer consciously.

To produce this type of manuscript, a pharmaceutical company typically hires a so-called medical education or medical communication company, which hires the ghostwriters. These writers usually have strong science backgrounds in addition to sharp writing skills. The company then finds a "guest" physician, someone with a prestigious name, who agrees to sign on as the author. (As for why a doctor would do so, the reason is simple: The pay is good.)

Who is an author?

Ghostwriters can play a legitimate role in producing medical literature. There is nothing fundamentally wrong with a doctor hiring a skilled science writer to write up a clinical trial report, review article, or even opinion piece. (4) Doctors and academics are notoriously poor writers, as I have ample reason to know from my experience at the NEJM. Manuscripts we accepted for publication went through four or five levels of editing, becoming a blur of colored pen and pencil marks; the published articles often bore little resemblance to the original submissions.

If a contributor wants to hire an experienced, knowledgeable wordsmith to help craft an article for publication, and does so in an honest and upfront manner, that is perfectly legitimate. The concern is with the writer's level of involvement with the piece.

What constitutes authorship of a scientific paper? How much does a physician have to contribute to be considered a bona fide author? What about the ghostwriter, working at the behest of a pharmaceutical firm?

Guidelines issued by the International Committee of Medical Journal Editors (ICMJE) propose that authorship credit should be based on three criteria:

* "substantial contributions to conception and design, or acquisition of data, or analysis and interpretation of data"

* "drafting" or "revising" the article "for important intellectual content"

* "final approval of the version to be published." (5)

The guidelines of the World Association of Medical Editors (WAME) are similar. (6) They urge editors to have authors precisely describe the role of each coauthor of the study, review, or analysis, and his or her role in preparing the manuscript. Even when these guidelines are followed, determining what constitutes ghostwriting and what constitutes genuine authorship can be difficult.

Some say that as long as the invited guest author has examined the ghostwritten paper extensively and made substantial editing changes, the manuscript should be considered the work of the guest, not the ghostwriter. I disagree. A few years ago, while conducting research for a book I was writing, I came across the Web site of the National Anemia Action Council. It contained material that was extensively slanted to promote the off-label use of a drug sold by Amgen--the company that sponsored the site.

To my surprise, one of my own faculty members in Boston was among the 26 prominent physicians who had contributed clinical material and opinions to the site. In a conversation with him about his involvement, he admitted that he hadn't spent much time reviewing the content because he trusted the people who were developing it.

Undoubtedly, there are many other guest authors who simply sign their names to a ghostwritten paper after giving it a quick once-over. Even extensively edited ghostwritten articles may not be free of company-introduced bias.

Nailing down the numbers

No one can say for sure how much influence drug companies exert over the medical literature. But recently, some prominent journal editors have raised cause for concern.

Richard Smith, a former editor of the British Medical Journal, has written that medical 'Journals [have become] an extension of the marketing arm of pharmaceutical companies," (7) and Richard Horton, editor of the Lancet, concluded that "journals have devolved into information-laundering operations for the pharmaceutical industry." (8) Though neither offers much more than a "gut" feeling about the degree of the industry's influence, their opinions must be taken seriously.

Attempts to quantify the frequency of ghostwritten medical articles have yielded varied results. A survey of more than 800 authors of research articles, review articles, and editorials in three major medical journals found that 11 percent had evidence of ghost authors, (9) and a self-administered survey of participants in the Cochrane summaries of medical information found evidence of ghost authorship in only 9 percent of the summaries. (10) A survey of authors of 1,000 published articles from 10 international journals found evidence of writing assistance in only 6 percent of papers. (11)

Even if only 5 percent to 10 percent of journal articles are ghostwritten, this represents a large number. More troubling, in a recent survey of scientists conducting research funded by the National Institutes of Health, 15.5 percent of respondents said they had adjusted the design or interpretation of research to mollify a sponsor, and 10 percent admitted to inappropriate assignment of authorship credit. (12)

And a study published in Denmark early this year offers an even more worrisome view. It compared the protocols of 44 industry-initiated clinical trials with the published articles and concluded that "ghost authorship in industry-initiated randomised trials is very common, and we believe that this practice serves commercial purposes." (13) The authors found that in 26 of the 44 published studies, there was an explicit statement that "the company conducted the statistical analyses or wrote the clinical study report or the manuscript." (14) In 31 of the 44 studies, the researchers found evidence of ghost authorship.

All these studies have substantial limitations. Most involve self-reporting, and the exact frequency of company-inspired writing is well hidden. Nonetheless, it is evident that companies have important input not only into clinical trial reports but also into review articles and editorials, and not only in minor medical journals.

Ghostbusters

Recently, Adriane Fugh-Berman, a physician at Georgetown University, had an experience similar to Troyan Brennan's. The pharmaceutical firm AstraZeneca offered her an honorarium to append her name to an editorial prepared by a medical communication company. The company sent her a draft of the article, complete with her byline and biographical information. Like Brennan, Fugh-Berman declined the offer. Several months later, she received a modified version of the same article to review--only this time, it had someone else's byline on it. She too went public with her ghostwriting story. (15)

Medical editors and legitimate science writers have been at the forefront of the attempt to expose and limit industry ghostwriting. In the dozen or so years in which concern about the practice has escalated, medical journal editors have issued tougher requirements for authors.

Many journals subscribe to the ICMJE and WAME guidelines on authorship. Some require authors to assert that they had control of the data, carried out the statistical analysis themselves, and take full responsibility for the integrity of the data. In some journals, authors must state their specific contribution to a study and to a manuscript. Some even require that statistical evaluation of a study be carried out by an independent analyst at a medical research center if the study is funded by industry. (16)

One proposal for "cleaning up" authorship and reducing industry's influence calls for the radical restructuring of peer review. (17) This would include open, Web-based disclosure of protocols, data, and statistical assessment, with publication only of systematic reviews based on the study data. While this approach sounds solid in theory, its economics and practicality are questionable, and journal editors have not rallied around it.

Guidelines from the new online-only, open-access journal Open Medicine seem like the soundest approach. The journal says its editors will carefully review components of study protocols that are vulnerable to manipulation by the sponsor, and it will publish clinical trials only if they are registered in a public database endorsed by the ICMJE. This requirement ensures that the sponsor cannot suppress a study with unfavorable results. Open Medicine requires that the roles of sponsors be transparent and will reject papers in which the decision to publish is not controlled by the researchers. (18)

Certainly, it is discouraging that professionalism and respect for medical integrity are not enough to keep physicians from embracing too-close relationships with the pharmaceutical industry. It is disheartening that examples of ghostwriting continue to emerge and that editors must remain constantly alert to it, implementing ever more restrictive policies that affect all authors.

The ties between industry and the authors of medical studies, review articles, and editorials are strong, pervasive, and dangerous. Physicians must work harder to prevent bias from creeping into the medical literature. Academic medical centers must help them negotiate contracts with industry that keep test data pristine and prevent manipulation of results.

Physicians should never claim authorship of review articles or editorials prepared by industry or medical communication companies--no matter how much they participate in revising the drafts. The potential for bias is simply too great.

Physicians who are tempted to sign their name to a ghostwritten article would do well to stop and ask themselves the fundamental ethical question raised by Howard Gardner, a professor of education at Harvard: "If all workers in my profession adopted the mind-set I have, or if everyone did what I do, what would the world be like?" (19)

Notes

(1.) Troyan A. Brennan, Buying Editorials, 331 N. Eng. J. Med. 673 (1994).

(2.) See generally Jerome P. Kassirer, On the Take: How Medicine's Complicity with Big Business Can Endanger Your Health (Oxford U. Press 2004).

(3.) Id. at 31-35; Anna Wilde Mathews, At Medical Journals, Writers Paid by Industry Play Big Role, Wall St. J. Al (Dec. 13, 2005).

(4.) See Jim Cozzarin et al., The Role of Professional Writers, www.clinicaltrialstoday.com/ centerwatch_clinical_tria/2006/07/the_role_of _pro.html (july 31,2006).

(5.) Intl. Comm. of Med. J. Editors, Uniform Requirements for Manuscripts Submitted to Biomedical Journals: Writing and Editing for Biomedical Publication, www.icmje.org. (Feb. 2006).

(6.) WorldAssn. of Med. Editors, Ghost Writing Initiated by Commercial Companies, www.wame. org/resources/policies#ghost (june 20, 2005).

(7.) Richard Smith, Medical Journals Are an Extension of the Marketing Arm of Pharmaceutical Companies, 2 PLoS Med. e138/0364, 0365 (May 2005).

(8.) Richard Horton, The Dawn of McScience, 51 N.Y. Rev. of Books 7 (Mar. 11, 2004).

(9.) Annette Flanagin et al., Prevalence of Articles with Honorary Authors and Ghost Authors in Peer-Reviewed Medical Journals, 280 JAMA 222 (1998).

(10.) Graham Mowatt et al., Prevalence of Honorary and Ghost Authorship in Cochrane Reviews, 287 JAMA 2769 (2002).

(11.) Karen L. Woolley et al., Declaration of Medical Writing Assistance in International Peer-Reviewed Publications, 296 JAMA 932 (2006).

(12.) Brian C. Martinson et al., Scientists Behaving Badly, 435 Nature 737, tbl. 1 (2005).

(13.) Peter C. Gotzsche et al., Ghost Authorship in Industry-Initiated Randomised Trials, 4 PLoS Med. e19/0047, 0050 (Jan. 2007).

(14.) Id. at 0049.

(15.) Adriane Fugh-Berman, The Corporate Coauthor, 20J. Gen. Internal Med. 546 (2005).

(16.) See Philip B. Fontanarosa et al., Reporting Conflicts of Interest, Financial Aspects of Research, and Role of Sponsors in Funded Studies, 294 JAMA 110 (2005).

(17.) See Richard Smith & Ian Roberts, Patient Safety Requires a New Way to Publish Clinical Trials, 1 PLoS Clin. Trs. 1 (May 2006).

(18.) www.openmedicine.ca/about/editorial policies.

(19.) Bronwyn Fryer & Howard Gardner, The Ethical Mind: A Conversation with Psychologist Howard Gardner, Harv. Bus. Rev. 51 (Mar. 2007).

JEROME P. KASSIRER, a former editor of the New England Journal of Medicine, is a professor at Tufts University School of Medicine in Boston, a visiting professor at Stanford University in Palo Alto, California, and the author of On the Take: How Medicine's Complicity with Big Business Can Endanger Your Health (Oxford U. Press 2004).