May 29, 2015
By Shannon Brownlee, MSc
When I first heard of the problem of conflict of interest in medicine, I couldn’t believe my ears. It was 2003, and Drummond Rennie, an editor at the Journal of the American Medical Association (JAMA), was giving a lecture at a meeting on the topic, and what he was saying violated everything I thought I knew about the integrity of science, the effectiveness of peer review, and the ethical imperatives of the medical profession. I was so naïve.
Rennie argued that medical research and publishing was being systematically and demonstrably biased by the flow of cash from the pharmaceutical industry. Academic researchers were being used as stealth marketers; articles published in the medical journals were serving as advertising vehicles for product; and physicians were effectively being bought with trips to Hawaii, free lunches and trinkets. It all sounded a little crazy, a perception that was aided by the fact that Rennie himself seemed more like a nutty professor, complete with English accent and unruly white hair, than editor of one of the world’s premier medical journals.
I didn’t think much more about the topic until the New York Times Magazine asked me to write an article about conflict of interest and the medical journals. That assignment was sparked by the appearance of a story in the Times about Charles Nemeroff, then an editor at a neuroscience journal and chairman of psychiatry at Emory. Nemeroff probably holds some sort of record for the most, or maybe just the most egregious conflicts of interest. Nemeroff had co-authored for Nature Neuroscience, assessing the effectiveness of two dozen treatments for mental illness without disclosing that he held direct financial interests in three of the treatments that received high praise. When asked by a reporter from the New York Times why he failed to at least disclose his conflict of interest, Nemeroff replied, in effect, that the journal never asked.
The more I dug into conflict of interest, the more I realized that if anything, the Nemeroff case and Drummond Rennie’s lecture only touched the tip of the iceberg. I found a major study of Paxil that purported to show it was safe and effective for children, but which buried cases of children who were hospitalized, and exaggerated the positive results. It would later emerge that the lead author, Martin Keller, then chief of psychiatry at Brown University, had conflicts of interest worth tens of thousands of dollars a year. The Vioxx scandal was still unfolding, but it was already clear that Merck, the drug’s manufacturer, had hidden data showing the drug increased the risk of stroke and heart attack, and marketed the drug even more heavily after questions were raised. The academic co-authors on the main study that launched Vioxx, the VIGOR trial, had various financial ties to the company and had effectively helped bury the evidence deep in the paper showing that whatever protection the drug offered over other NSAIDS against gastrointestinal bleeds was wiped out by the increased risk of cardiovascular events. It would later turn out that crucial safety data reported to the FDA were not included in the VIGOR article.
I learned that the industry deliberately and methodically used speaking and consulting fees to recruit doctors to serve as covert marketers, or “key opinion leaders,” in drug industry marketing lingo. At that time, most research universities had flimsy policies in place to ensure that industry-funded clinical studies were conducted, analyzed, and reported without undue influence from the sponsor. I went to medical meetings and toured the exhibit halls, which could more properly be described as schwag extravaganzas, where doctors went from booth to booth stuffing booty bags full of everything from coffee mugs to expensive electronic items, all embossed with the name of a drug.
Most persuasive of all, study after study showed that medical research – the scientific information that doctors and patients rely on to guide treatment – was being systematically distorted by industry sponsorship and by the conflicts of interest of academic researchers who conducted it. Among the most important of these studies was a meta-analysis published in JAMA in 2003. The authors analyzed eight systematic reviews, which included more than 1,100 individual studies assessing the relationship between industry funding and the outcomes of the research. The bottom line: studies that are funded by industry were 3.6 times more likely to produce pro-sponsor outcomes than those that are publicly funded. As the authors of the paper drily put it, “Industry preferentially supports trial designs that favor positive results.”
And all the while, academics were trying to claim that the consulting contracts and speaking fees had no impact — they and their science were above reproach. A typical comment came from former surgeon general C. Everett Koop, after he testified before Congress that latex gloves do not cause allergic reactions. Of course, latex does cause allergic reactions, and it would later emerge that Koop had received more than half a million dollars from a manufacturer of latex gloves. When asked if the money had anything to do with his opinion of latex, Koop replied, “It never occurred to me it could be a conflict of interest.”
That a physician of Koop’s stature and accomplishment could be so blind to the ethics of taking money from a manufacturer underscores the often unconscious and insidious nature of conflicts of interest. Only in rare cases do clinical researchers deliberately alter data because they stand to gain financially. Instead, conflicts of interest affect behavior in more subtle ways. A researcher decides not to publish a negative result because it will affect her ability to get funded by a company for future studies. The company exerts control over the study design and analysis of the results, and the researcher goes along. A prominent academic agrees to put his name on a ghostwritten paper. Or the researcher already believes the treatment is effective and sees nothing wrong with being paid to tout it. Whether the recipient is aware of her own bias or not, companies use these financial relationships to further their agenda, which is to sell more of their product.
He who pays the piper calls the tune
Fast-forward to the present, and we have a few policies in place to protect the integrity of medical research, and the patients whose lives and wellbeing depend upon rigorously good science. Most universities have at least some sort of conflict of interest rules. Many academic medical centers have banned detailers, and the free samples and lunches they bring. The media has woken up to the problem and begun writing stories about conflict of interest. We have the Sunshine Act, which has exposed just how widespread (and in some cases lucrative) consulting and speaking fees are for those key opinion leaders. The major medical journals and continuing medical education programs have implemented disclosure rules. Pharmaceutical companies have pledged to register all clinical trials with the National Institutes of Health, reducing the temptation to bury negative results and exposing potentially biased changes in trial design.
Slender as many of these remedies are, it was inevitable that they would trigger a backlash. The first sign was a flurry of blogs and op-eds a few years ago by Tom Stossel, a Harvard researcher, and David Shaywitz, a consultant, both with ties to industry, who are on a crusade to discredit the entire notion of conflict of interest. They argue that any barrier to industry-academic relationships “slows the progress of potential treatments,” and anyone who dares to suggest that conflict of interest might be a problem is a “pharmascold.”
Now the New England Journal of Medicine has weighed in with three weeks devoted to a series of essays by yet another conflict of interest denialist, Lisa Rosenbaum, a young cardiologist, who argues that the pharmascolds have gone too far. Her essays are so filled with illogical holes and inconsistencies it would take too long to critique them point-by-point, but four principal fallacies stand out.
First, she claims there’s no evidence that conflict of interest has harmed patients. That can only be true if your definition of evidence involves a randomized controlled trial with one arm conducted by conflicted researchers and funded by pharma, and the other conducted by unconflicted researchers and funded by government. (Not that government funding is a perfect guarantee of ethical research. Remember Tuskegee.)
Requiring RCTs to show harm due to conflict of interest is absurd, of course. If Rosenbaum wants evidence that conflict of interest has hurt patients, she need to look no further than the recent scandals around Fen-Phen, Rezulin, Vioxx, the atypical antipsychotics, the selling of growth hormone for normal children of short stature, among the many examples of drugs and devices whose harms have been underplayed and benefits oversold with the help of biased research and academics with conflicts of interest. She might also consider the harm caused by overmedicalization, or “disease mongering,” widening the boundaries of treatable disease in order to sell a treatment.
Second, she thinks that the medical profession and the drug and device industries have a “shared mission to fight disease.” That’s a nice sentiment, and undoubtedly true for many, maybe even most individuals who work in industry. I even know some of them. But companies are also institutions, which have a very different fiduciary duty: they are beholden to shareholders. Much of the research they fund serves their marketing needs more than a lofty goal of fighting disease, and they certainly are not interested in expanding the body of medical knowledge that will contribute to improving health unless it also contributes to the bottom line.
Third, Rosenbaum claims that limiting conflicts of interest will kill the golden goose of medical innovation. Really? Does that mean the academic researchers who do not have conflicts of interest have nothing to offer the world in the way of cures? For somebody who claims to want evidence that conflicts of interest cause harm, Rosenbaum is remarkably willing to parrot a groundless position. Forgotten in such a claim are the contributions to cardiology of Bernard Lown and Andreas Gruentzig, or Percy Lavon Julian, who synthesized steroids, or Jonas Salk or the many researchers who labored to improve the lives of patients without industry backing.
That’s not to say we don’t need manufacturers to advance medicine. We do. But we also need medical evidence we can trust, which leads to the most important hole in Rosenbaum’s arguments: she simply ignores the bad science that industry has funded over the years and that academic medicine puts up with because academic researchers have been paid to conduct so much of it. How does science get distorted? Let me count the ways. There are straw-man comparators, underpowering, data dredging, post-hoc subgroup analysis, failing to report harm, publication bias, surrogate endpoints, composite endpoints, and the list goes on.
Equally important is the research that does not get done because we have handed control over the funding of clinical trials to industry, giving it extraordinary power over what we know and don’t know about what works and does not work in medicine.
Do I think pharmaceutical and device companies are run entirely by evil people? Of course not. They are working within the rules they’ve been given. Do I think researchers who have conflicts of interest are unethical and greedy? I’m sure some of them are, but most are simply kidding themselves that they can take the money, allow industry to control the agenda, and still conduct unbiased research. As physicist Richard Feynman once said, the first principle of good science is you must not fool yourself, and you are the easiest person to fool.