Ghostwriters for medical research criticized, reforms urged

Two new reports involving the painkiller Vioxx raise fresh concerns about how drug companies influence the interpretation and publication of medical research.

The reports claim Merck & Co. frequently paid academic scientists to take credit for research articles prepared by company-hired medical writers, a practice called ghostwriting. They also contend Merck tried to minimize deaths in two studies that showed that the now withdrawn Vioxx didn't work at treating or preventing Alzheimer's disease.

Merck called the reports in Wednesday's Journal of the American Medical Association false and misleading. Five writers of the articles were paid consultants for people who sued Merck over Vioxx's heart and stroke risks; the sixth testified about Merck and Vioxx's heart risks before a Senate panel. Merck says those connections makes the reports themselves biased.

While Merck is singled out, the practices are not uncommon, according to JAMA's editors. In an editorial, they urge strict reforms, including a ghostwriting crackdown and requiring all authors to spell out their specific roles.

Dr. Catherine DeAngelis, JAMA's editor-in-chief, said those are already policies at JAMA but not at many other journals.

"The manipulation is disgusting. I just didn't realize the extent," she said.

The practices outlined in JAMA can lead editors to publish biased research that can result in doctors giving patients improper and even harmful treatment, she said.

DeAngelis said doctors, medical researchers and journal editors bear some responsibility for those harms.

"We're the ones who have allowed this to happen. Now we've got to make it stop," she said.

Drug studies involve several steps, including designing and performing the research, analyzing the results and writing them up for submission to a medical journal. Pharmaceutical companies sometimes pay for a study but have independent scientists perform all those steps. Sometimes companies and their own scientists are involved in some or all the steps, and those were the studies scrutinized in the JAMA reports.

The articles are based on reviews of company documents from court cases over Vioxx, which was pulled in 2004 because of its heart and stroke risks. Merck agreed to pay $4.85 billion last November to settle thousands of lawsuits.

One JAMA report says internal company data showed in 2001 that Vioxx patients in two Alzheimer's studies had a higher death rate than patients on dummy pills. Merck didn't publicize that "in a timely fashion" and provided information to federal regulators that downplayed the deaths, the report said.

But Jim Fitzpatrick, a Merck attorney, said "it's completely not true" that Merck tried to minimize those deaths. He said a Merck analysis found the excess deaths were not related to Vioxx.

The other JAMA article says one Alzheimer study was designed and conducted mainly by Merck scientists. But when published, the lead authors listed were academic scientists not named in a study draft.

Peter Kim, head of Merck Research Laboratories, said those authors "were intimately involved in the studies." One was New York University Alzheimer's specialist Steven Ferris. He also disputed the implication that he had little to do with the study, and said Merck paid him for his work.

Fitzpatrick acknowledged that Merck has hired outside firms to write drafts of other studies that later list scientists as first authors. In those cases, the scientists are expected to review the manuscript and can suggest changes, he said.

The Alzheimer study was published in 2005 in the journal Neuropsychopharmacology. Its new editor, Dr. James Meador-Woodruff, said the journal's policies have been strengthened to ban ghostwriting.

---

On the Net:

JAMA: http://jama.ama-assn.org