JAMA editor: Corporate research funding
poses potential problems for academics
The editor-in-chief of the Journal for the American Medical Association (JAMA) lived up to her reputation for tough talk on conflict of interest in medical research during a lecture at U-M Oct. 27.
The occasion for Dr. Catherine DeAngelis' visit to the MCHC Auditorium at C.S. Mott Children's Hospital was the sixth annual Center for the History of Medicine's Horace W. Davenport Lecture in the Medical Humanities.
While acknowledging the necessary relationship between drug companies and academic research institutions, DeAngelis' purpose for the talk was to outline some of the dangers of that relationshipthe potential for drug company control over conditions and statistical analyses, delays in publication when study results are unfavorable to product claims, and doctors' failure to report financial interests in companies, among othersand to criticize those who she believes have abused the relationship.
It only makes sense that makers of medical devices, biomedical and biotechnology firms pay for almost 100 percent of clinical trials, she said, because they receive the most benefit from them.
"The problem is the 'golden rule'you know, whoever has the gold, rules," she said, explaining that firms often tell researchers how to set up trials, measure statistics and even determine whenor ifresults get published. Within the last 10 years drug companies have come to stand atop the list of firms who lobby U.S. lawmakers, she said, even ahead of insurance and energy companies, banks and the tobacco industry. In the meantime, she said, JAMA's advertising has remained level over that time, while revenue from that advertising has declined, possibly signaling that companies believe they need journal advertising less as their influence over the research process and lawmakers expands.
Sometimes the U.S. Food and Drug Administration will approve a drug before a journal can publish trial results, which can cause a problem for editors. DeAngelis has responded by instituting a policy that no ads are placed near stories related to the advertiser or the product.
"If the FDA approves it, I can't stop it," she said. "But I can make sure the reader knows it's an ad."
In a recent example, one company asked for its ad to be placed in a four-page spread for a particular issue in which the results of a related study were to be published. Her staff told her it was a big ad, and the journal could really use the money.
Out of concern for the implications of placing the ad near the results, DeAngelis said she told the company no at first, then relentedto an extent.
"Then I moved the article two weeks ahead," she said.
Publication delays also have proved problematic for editors. DeAngelis recalled two cases, one in which a drug actually worsened a condition in patients and another when the drug being studied eventually proved to be equivalent to a generic. Publication of both studies' results were delayed three years, she said, meaning that doctors and patients were denied the opportunity to learn about harmful effects or unnecessarily high costs of those products.
The upside, if there was one, DeAngelis said, was that lawyers used those cases to learn how to write contracts. She cited one case where a University of California, San Francisco researcher found an HIV immunogen to be ineffective. This time, however, JAMA published the results and was sued for $8 million; an arbitrator later ruled in favor of the university.
Conflicts of interest also can result in outright fabrication, DeAngelis said. A doctor paid by the Korean government to find and publish study results that said patient-specific stem cells could be derived from blastocysts. Though the doctor later "expressed concern" over the study, "he still has the $75,000," DeAngelis said.
Results publication also has been manipulated to hide bad news, she said, recalling a since-publicized study that called Vioxx dangerous.
DeAngelis recently has faced accusations of sloppiness at JAMA, where several authors' articles have failed to make important disclosures. However, she reiterated her earlier public statements that with 6,000 submissions annually that average six authors each, she simply cannot check every researcher. She also told the U-M audience that with the support of editors from other major journals, she has begun to require academic statistical analysis from researchers for companies that want to publish study results.
In addition, DeAngelis said, JAMA and most other medical journals require participants in peer review and publication processes, whether as an author, reviewer or editor, to reveal ties to activities that could inappropriately affect the outcomewhether or not it is in fact affected.
"We have to be able to talk about funding sources and be transparent to the reader and enable them to interpret things in light of financial interest," DeAngelis said. "Disclosure policies are essential and management of interest is essential for public trust."