A pharmaceutical company seeking Food and Drug Administration (FDA) approval for a new drug usually pays for the clinical trials needed to demonstrate that the drug is safe and effective. This arrangement is widely accepted as fair and efficient. The company that stands to profit from the drug pays for the research needed for its approval, and the scientists hired to conduct these trials are paid to follow strict research procedures that yield clear, unbiased results.
Pharmaceutical companies, however, have found many ways to influence the outcome of clinical trials. Some have hired ghostwriters to shape research findings into a flattering portrayal of the drug, and then submit the results under the names of respected scientists to a peer-reviewed journal.
And after approval, manufacturers may conduct another study to demonstrate the drug's superiority to similar drugs produced by competitors. The goal is to produce scientific evidence that can be used to market the drug, which is why some scientists refer to these studies as “experimercials” – a form of advertisement thinly disguised as objective information.
Scientists who conduct these studies are not expected to falsify results to produce favorable evidence, but they may design the study to avoid findings that might reveal shortcomings of the drug. “The companies seem to get the results they want not by fiddling with the results, which would be far too crude and possibly detectable by peer review, but rather by asking the ‘right’ questions – and there are many ways to do this,” wrote Richard Smith, MD, former Editor of the British Medical Journal, in an article titled “Medical Journals Are an Extension of the Marketing Arm of Pharmaceutical Companies” (PLoS Med 2005;2(5):e138).
STUDY DESIGN: EPILEPSY DRUGS
An August 8th editorial in Neurology (2006;67:378–379) reveals how a study comparing the anti-epilepsy drugs lamotrigine and topiramate (Topamax), published in the same issue of the journal, avoided potentially negative results for lamotrigine, marketed as Lamictal by GlaxoSmithKline, which sponsored the study.
The study, by David Blum, MD, and colleagues, focused on the effect of the two drugs on cognitive function, and found that lamotrigine “had significantly less impact than topiramate on measures of cognition when used as adjunctive therapy for partial seizures” (Neurology 2006;67:400–406). Three of the authors were employees of GlaxoSmith-Kline, and others had received grant money and honoraria from the company.
The study outcome, according to the editorialists David Chadwick, MD, and Michael Privitera, MD, was predictable on the basis of earlier trials and was probably not worth demonstrating. However, assessing the effectiveness of the two drugs in controlling seizures would have been helpful, but not to GlaxoSmithKline.
“It is more likely than not that a study adequately designed and powered for seizure outcomes would have favored topiramate,” Drs. Chadwick and Privitera wrote. “When a trial is sponsored by industry, subtle biases in its design may be invisible to all but the experienced.”
PERVASIVE CONFLICTS OF INTEREST
Not long ago, a scientist who had such blatant conflicts of interest would have been unable to get into a peer-reviewed journal. When Jerome P. Kassirer, MD, became Editor of the New England Journal of Medicine in 1991, the journal refused to publish papers by researchers who had a financial conflict of interest. That policy was abandoned in 2002, however, because virtually all the authors who submitted papers reported some sort of conflict of interest.
Such conflicts “reduce professional dignity and integrity, denigrate the profession, and erode trust in the profession's practitioners, researchers, and institutions,” Dr. Kassirer declared in his 2004 book, On the Take: How Medicine's Complicity With Big Business Can Endanger Your Health (Oxford University Press).
Today, medical journals “have devolved into information laundering operations for the pharmaceutical industry,” according to Richard Horton, Editor of the Lancet. If research favorable to a certain drug appears in a respected, peer-reviewed journal, the manufacturer can cite the facts in its ads and order reprints to distribute to doctors. (Both ads and reprints are very profitable for journals.)
RESPONSE BY JOURNALS
Many journals have responded to conflicts of interest by demanding that authors report all money they have received from the drug companies.
“Disclosure mitigates a conflict but doesn't eliminate it,” said James L. Bernat, Professor of Neurology at Dartmouth Medical School and a former Chair of the AAN Ethics, Law, and Humanities Committee (1993–2003). “If people are aware that the author is an employee of a pharmaceutical company, or was paid a stipend as a consultant or lecturer, then they can exercise some degree of skepticism. That's why I've always been a strong advocate for disclosure. Even though it only mitigates the problem, at least it's a step in the right direction.”
Neurology, for example, requires authors to divulge if they have received financial compensation for their authorship. Even Drs. Chadwick and Privitera, authors of the editorial criticizing the bias produced by conflicts of interest, had to disclose that they have received payments and research support from GlaxoSmithKline, the company whose research they were questioning, as well as from other pharmaceutical companies.
While this may seem ironic, Dr. Chadwick contends that accepting money from pharmaceutical companies is unavoidable these days. “One has to find a method of working with them in a sensible way,” he said. “The key is that the data collection and analysis has to be in the hands of the independent academic, not the industry, which treats information as a valuable commodity and guards it jealously. Many of the people working in drug development are entirely reputable scientists, but the marketing people exploit those clinicians who are willing to go anywhere and say anything as long as the speaker's fees are large enough. Their disclosures probably aren't worth the paper they're written on.”
IS DISCLOSURE SUFFICIENT?
Is the disclosure of a conflict of interest sufficient? An elected official who accepted money to vote a certain way would be disgraced, even if he or she disclosed the arrangement. But scientists who receive generous payments from a pharmaceutical company may report their findings about that company's drugs in respected journals, as long as they disclose the arrangement.
Yet, even disclosure is too burdensome for some scientists. Charles Nemeroff, MD, Editor of the journal Neuropsychopharmacology, resigned in August because he failed to disclose that he and seven other authors of a study in the July journal about a vagus nerve stimulator for depression have financial ties to Cyberonics, the company that manufactures the device (2006; 31:1345–1355). The authors reported that the stimulator is “a promising and well-tolerated intervention that is effective in a subset of patients with treatment-resistant depression.”
However, in congressional testimony earlier this year, Sen. Charles Grassley (R-IA) said FDA reviewers opposed the device because the manufacturer had not satisfactorily demonstrated its safety and effectiveness.
In another example of the murky issue of disclosure, JAMA published an article by 13 psychiatrists claiming that pregnant women who discontinued their antidepressants during pregnancy were more prone to relapse (2006;295:499–507). Research over the past decade has found that mothers who take serotonin-selective reuptake inhibitors (SSRIs) during pregnancy have a slightly higher risk of having babies with birth defects, cardiac malformation, respiratory distress, and withdrawal syndrome.
The authors of the article failed to disclose that they were paid by pharmaceutical companies that sell SSRIs. The Wall Street Journal reported that the lead author of the report, Lee Cohen, MD, a Harvard Medical School professor and Director of the Perinatal and Reproductive Psychiatry Research Program at Massachusetts General Hospital, “is a longtime consultant to three antidepressant makers, a paid speaker for seven of them, and has his research work funded by four drug makers.”
Dr. Cohen told the Associated Press that the authors did not disclose more than 60 financial relationships with drug companies because “we did not view those associations as relevant.”
Such indifference to disclosure rules is not uncommon. Nor are more brazen attempts to circumvent the safeguards that journals put in place to prevent manipulation of research.
In an article titled “The Influence of Money on Medical Science,” JAMA editor Catherine D. DeAngelis, MD, MPH, described how the authors of one study sponsored by a pharmaceutical company and accepted by JAMA withdrew the paper rather than let it be subjected to the independent statistical analysis that the journal requires of all industry-sponsored studies (2006;296:996–998). “The paper was published elsewhere shortly thereafter and has received much media coverage,” she said.
WHAT CAN BE DONE?
What can be done in addition to demanding full disclosure to curb such egregious and widespread conflicts of interest in biomedical research? Dr. Smith, the former Editor of the BMJ, has called for more public funding of clinical trials, “particularly of large head-to-head trials of all the treatments available for treating a condition.”
But such an alternative would face daunting opposition in the pro-business political climate of the United States today. “I think it's so unlikely to happen given the American tax structure that it's not even worth discussing,” said Robert C. Griggs, MD, Editor-in-chief of Neurology. “You'd be talking about replacing private enterprise. I believe in capitalism.”
Instead, he suggested that drug companies be required to register their research studies. “That way, before they're published you can look to see if they deviated from their plan,” Dr. Griggs said. “Registration is required by our journal.”
Michael L. Goldstein, MD, Vice-President of the AAN, said the Academy's Quality Standards Subcommittees (QSS), which assess the literature on specific problems such as Parkinson disease and epilepsy, and then issue treatment guidelines, help detect bias in research.
“I've participated in QSS discussions, and the people involved understand the field and argue with each other about the meaning of studies so they can present the best information to neurologists,” he said.
But disclosure remains the primary line of defense against bias, which is why the New York Times recently called for putting “more muscle” into rules requiring disclosure of conflicts of interest. “If all leading journals agreed to punish authors who fail to reveal their conflicts by refusing to accept further manuscripts from them, a lot more authors would be inclined to fess up,” the editorial stated.
Michael A. Williams, MD, Chair of the AAN Ethics, Law, and Humanities Committee, agrees. “There has to be a consequence for failure to disclose or people will game the system,” he said. “You need ethical principals to follow, but unless rules have consequences, people won't pay any attention.”
ARTICLE IN BRIEF
✓ Editors and authors of major journal papers debate whether conflicts of interest can be avoided when most studies are funded by pharmaceutical companies.