When a company hired by a pharmaceutical manufacturer offered to help Lisa M. Shulman write a research article, Dr. Shulman was very interested in aid. She had just reached the rank of assistant professor of neurology at the University of Miami and was overworked and under-resourced, serving a fellowship on Capitol Hill.
Dr. Shulman let the company, DesignWrite, select background information on connections between estrogen and Parkinson’s disease, and then let it draft a proposed summary of the existing data.
She used that summary to write her article, which appeared in the June 2002 edition of the journal Parkinsonism & Related Disorders without any mention of DesignWrite or of Wyeth, the pharmaceutical maker that paid it and marketed estrogen pills. The lack of attribution was “an oversight,” Dr. Shulman told The Chronicle.
Despite such defenses, instances of “ghostwriting,” in which drug companies help researchers write articles that appear in medical journals, are drawing sharp scrutiny. The practice raises the specter of hidden bias in published papers that favor the effects of the company’s drugs. Doctors rely on such papers when making life-or-death choices about treating their patients.
A study presented last week at a major conference on medical publishing held here found that in The New England Journal of Medicine, at least 11 percent of the articles had “ghost” authors. Another study tracked attempts in the late 1990s by Parke-Davis, now a subsidiary of Pfizer Inc., to get articles published concerning its medication Neurontin. The pharmaceutical company succeeded in placing 11 articles in seven journals, none of which disclosed its role in authorship and only two of which acknowledged its financial support, according to researchers at the conference, the Sixth International Congress on Peer Review and Biomedical Publication.
In recent weeks, disclosures of researchers’ use of “ghosts” included Barbara B. Sherwin, a psychology professor at McGill University, who put her name on an April 2000 article that was actually written by a freelance author under contract to Wyeth. “I made an error in agreeing to have my name attached to that article without having it made clear that others contributed to it,” she told a Montreal newspaper.
Some institutions are trying to resolve the problem. They include Columbia University, which now explicitly requires faculty members and students to list all contributing authors when submitting articles to medical journals. Others are trying to remove “publish or perish” pressure from professors by limiting the number of papers that can be cited by a candidate for promotion.
But many of those who proffer solutions acknowledge that effective policies are hard to craft, particularly because the definition of ghostwriting-as-wrongdoing seems as hard to pin down as a phantom.
Dr. Shulman sees little harm in accepting outside professional help, since she takes full responsibility for the published contents. Her article, she says, is a dispassionate examination of whether estrogen has any connection with Parkinson’s disease.
The article opens by stating that “increasing evidence” supports the use of estrogen for guarding against Parkinson’s, although it notes conflicting findings based on variations in age and dosage. “There is nothing in my paper that is favorable to Wyeth,” says Dr. Shulman, who denies that her actions constituted ghostwriting.
Others have a different definition.
“The solution is essentially zero tolerance,” says Steven E. Nissen, chairman of cardiovascular medicine at the Cleveland Clinic. “No amount of editing,” he argues, can justify an author’s using an undeclared contributor.
Researchers are tempted to accept such help because it can be a career booster, says David J. Rothman, a professor of social medicine at Columbia University. A list of published articles is one of the first items checked at a medical school when a researcher seeks a promotion.
That suggests that one solution to the problem of ghostwriting might be for medical schools to place greater emphasis on quality rather than quantity, says Mr. Rothman, president of Columbia’s Institute on Medicine as a Profession. Harvard University has taken a step in that direction by sharply limiting the number of research papers it wants to see from candidates, he says. Harvard recommends only two published items from a candidate for assistant professor and five for associate professor. But more typical, Mr. Rothman says, are universities facing lists of 60 to 80 articles.
Too Much Ado?
There are those who argue that the worry over ghostwriting is largely overblown. Matthew R. Weir, director of the division of nephrology at the University of Maryland Medical Center, in Baltimore, says such articles typically appear only in lower-tier medical journals, which are recognized as less reliable. The study of ghost authors at The New England Journal of Medicine, presented at the conference, casts doubt on such assertions, however.
Dr. Weir has himself been drawn into a related controversy, accused in a lawsuit against the drug maker Merck of signing his name to an article that played down the chances that the company’s Vioxx medication might raise the risk of heart attacks. The co-author on the report was a scientist employed by Merck. Dr. Weir says he stands by the data in that article.
Articles suspected of being influenced by ghosts aren’t necessarily glowing tributes to a particular company’s product. Many of the articles instead discuss the advantages and disadvantages of a treatment, before conveying a general sense of the medical profession’s acceptance of the overall benefits for some patients.
For some scientists involved with corporate partners, that inclusion of competing viewpoints helps deflect accusations that they are participating in industry propaganda. Some also note that their articles don’t mention specific drugs by name, and instead just give a general endorsement of a particular therapeutic approach that happens to align with the medication offered by the company.
Rogerio A. Lobo, a professor of obstetrics and gynecology at Columbia University, says DesignWrite contributed work, unattributed, to one of his published papers. But the company’s role was limited to assembling existing research on a subject, providing statistics and charts, and copy editing. He says he substantially changed the final version by cutting out entire sections and eliminating the endorsement of a particular product.
“I don’t consider that ghostwriting,” he says. “I’m the responsible person, and I stand by it, and I wrote it.”
DesignWrite, a medical-communications company based in New Jersey, regards its service as providing a public benefit. “We stand behind the accuracy of every article we have been involved in,” says the company’s president, Michael Platt.
Both ghostwriting’s detractors and its alleged participants agree that the difficulty of defining the practice makes it tough to eliminate. “There is no perfect solution to this,” Dr. Nissen says.
Universities must allow their researchers some interaction with industry, Dr. Lobo says. Because of that, he continues, it’s never possible to know who is involved in some form of ghostwriting. “There’s really very little ability to police this,” he says.
Sen. Charles Grassley, Republican of Iowa, wrote to the National Institutes of Health last month, asking it to use the power of its budget—the NIH is the primary provider of federal research dollars to universities—to crack down on ghostwriting. But the senator isn’t offering any specific suggestion about how to accomplish that.
The NIH, while designed to promote research, isn’t well suited to police misconduct at universities or medical journals, Mr. Rothman says. Yet the temptation to press the agency is strong, he says, given its financial clout and its recent success in enforcing new standards for the protection of humans in medical research. “If the NIH barks,” Mr. Rothman says, “everybody listens.”
