In an ideal world, medical advances would result from an altruistic pairing of outstanding translational science with well-designed and efficient clinical trials, and the resulting conclusions could be applied for the betterment of global health – all in a not-for-profit environment.
In reality, while industry doesn’t fund research for entirely altruistic purposes, the benefits of academic-industry collaborations are incontrovertible. But do these relationships shake physician and public confidence in the validity and impartiality of the resulting data? Despite the combative connotations of the name, are conflicts of interest (COIs) always bad?
COI is pervasive in all areas of medicine, from research and publishing to teaching and clinical care. The question is not how medicine can rid itself of COI, but how it can best manage it. ASH Clinical News takes a closer look at COI – its prevalence, manifestations, and inherently controversial nature.
Research Under the Influence
The broad definition of COI is standard across all stakeholders: According to the Institute of Medicine (IOM; now the National Academy of Medicine), a COI comprises “a set of circumstances that creates a risk that professional judgment or actions regarding a primary interest will be unduly influenced by a secondary interest.”1
Similarly, the International Committee of Medical Journal Editors (ICMJE) describes COI as a situation in which “professional judgment concerning a primary interest (such as patients’ welfare or the validity of research) may be influenced by a secondary interest (such as financial gain).”2
The authors stress that a key issue is simply that the COI exists “whether or not a particular individual or institution is actually influenced by the secondary interest.”
“The point to be made about the definition, which people get confused about, is that the conflict relationship exists at all,” Robert Steinbrook, MD, a professor adjunct at Yale School of Medicine in New Haven, Connecticut, who has written extensively on the subject, told ASH Clinical News. “It’s not that your opinion or actions will necessarily be influenced one way or the other by that relationship, because that’s impossible to know.”
Dr. Steinbrook, an editor-at-large for JAMA Internal Medicine, clarified that adding the word “potential” to the phrase misses the point: “Many people will say, ‘Well, I have all this money from industry, but I’m not biased,’ or they’ll say that COI is a pejorative. When, in fact, it’s just a statement of a situation that exists that needs to be disclosed. This is a hard distinction to get people to accept.”
In his view, there’s “absolutely nothing wrong” with receiving research support from companies that are doing real research. “If you have bona fide scientific and clinical reasons to have relationships with companies, then tell people, and have the contracts written in ways that preserve your independence,” he said.
Follow the Money, Not the Bias
Even when COIs are relatively benign, they can erode readers’ confidence in research findings if they are not disclosed, Dr. Steinbrook and others agreed. According to the IOM report, disclosure of these potential secondary interests – and the ICMJE considers only financial interests, although they allow that COIs can occur for other reasons – is important for maintaining “public trust in the scientific process and the credibility of published articles.”
The ICMJE electronic COI template is the standard form in the industry, used by hundreds of peer-reviewed journals.
While the ICMJE’s definition pertains only to financial interests, the IOM report’s authors allow that these secondary interests extend beyond monetary gain. COIs can take the form of professional advancement and recognition, favors to others, or self-dealing; financial conflicts are the focus of these definitions, the authors wrote, because they are “relatively more objective, fungible, and quantifiable.”
“Although there are continuing disagreements about whether and when individuals with COI[s] should be excluded from such activities as writing opinion or review articles for medical journals or serving on federal advisory committees, the minimum standard in medicine for years has been full disclosure,” Dr. Steinbrook wrote in a 2017 editorial published in JAMA Internal Medicine.3 “That includes both financial and ‘appearance issues,’” he added. “Perceptions of COI are important.”
“[COI is] just a
a situation …
that needs to
This is a hard
get people to
—Robert Steinbrook, MD
This may seem straightforward, but undisclosed and unmanaged COIs result in a never-ending game of whack-a-mole. In May 2017, an issue of JAMA attempted to address some of the many facets of COI, exhaustively covering the topic in 23 essays and review articles. As an example of the omnipresence of COIs in medicine, two of the many topics covered in this wide-ranging issue were an examination of the COI pressures inherent in fee-for-service or volume-based payment models, and the ethical requirements to reveal COI to patients when referrals are made for diagnostic services at facilities in which the referring doctor has a financial stake.4,5
Spoiler alert: Salary-based payment systems reduce COI and might also reduce physician burnout. Patients reported having more trust in physicians who are upfront about their financial arrangements.
Two Sides of the Industry Coin
“It’s pretty clear that if you take money from anybody, you are more likely to come up with biased results regarding the product for which you are taking money,” Shannon Brownlee, MSc, senior vice president of the Lown Institute, told ASH Clinical News.
The data would seem to support her position. A 2012 Cochrane review analysis of 48 studies evaluating drugs and devices showed that industry-sponsored studies were “more favorable to the sponsor’s products than non–industry-sponsored drug and device studies, due to biases that cannot be explained by standard risk-of-bias assessment tools.”6 When the researchers updated the data in 2017 to include 27 new papers, for a total of 75 papers, the findings were similar: Industry-sponsored studies were 34 percent more likely to report favorable conclusions than non–industry-sponsored papers.7
Ms. Brownlee conceded that industry funding for research is inevitable and suggested a variety of approaches to manage the ensuing conflict. For one, she said, she’d like to see more institutions taking proactive roles in handling research contracts and creating greater separation between researchers and potential industry influence.
“Institutions can mitigate against the influence of who is paying for the trial,” she explained. “They can, for example, require that the design of the study must be investigator-controlled, the data have to be investigator-controlled and available to other investigators, and, of course, the writing should not be done by a ghostwriter.” To help maintain investigator independence, she added, institutions also can “protect researchers who come up with the wrong answers and make sure they are allowed to publish these results.”
Others, however, argue that the movement toward full transparency – or complete separation between industry and academia – has gone too far. Thomas P. Stossel, MD, a hematologist from Brigham and Women’s Hospital in Boston, Massachusetts, believes the COI disclosure campaign has overstepped its bounds and is now stifling innovation and wasting resources on bureaucracies that exist solely to manage research relationships.
“There are now-entrenched cultural
biases against industry that create
barriers to appropriate collaborations
that are in the best interest of patients
and scientific progress.”
—Michael Rosenblatt, MD and Sachin H. Jain, MD, MBA, in Harvard Business Review
“Most clinical applications of discovery are made by the pharmaceutical and medical device industries,” he wrote in a Surgery editorial last year.8 “Partnering with these industries not only affords research funding, but also a share of the market value of new discoveries.”
He cited the marked advances in health care over the last half-century as a justification for the existing system. “These improvements are due to partnerships between physicians, academics, and industry,” he wrote. But, because the “COI instigators, enablers, and enforcers” have so maligned the relationship between industry and innovators (“guilty until proven innocent”), he suggests, innovators have been hamstrung and innovation hampered, in some cases likely to fatal effect.
Dr. Stossel isn’t alone in his thinking. In a Harvard Business Review article published in June 2017, Michael Rosenblatt, MD, chief medical officer of Flagship Pioneering, and Sachin H. Jain, MD, MBA, president and chief executive officer of CareMore Health System and a former senior adviser to the administrator of the Centers for Medicare & Medicaid Services under the Obama administration, contend that COI rules are “holding back medical breakthroughs.”9
“We must recognize … that there are now-entrenched cultural biases against industry that create barriers to appropriate collaborations that are in the best interest of patients and scientific progress,” they stated.
The authors, who have worked in industry, academia, clinical medicine, and government during their respective careers, continued: “It is time for all parties to revisit those policies and replace them with rules that recognize both true conflicts and true confluences of interest. They are essential to forging the strong collaborations that are worthy of society’s trust.”
They suggest, among other things, that institutional COI policies be reworked to serve as frameworks for collaboration, rather than as tools “solely for policing.”
Academia also needs to accept COI related to careerism and other non-financial motives, they write. For example, “since many faculty members rely in part or fully on grants for their salaries, they are highly motivated to report data that supports grant applications.” This can have the unintended side effect of compelling researchers to design careers around supporting theories.
Their argument, essentially, is that one bad apple shouldn’t spoil the bunch. “Clearly, some unacceptable past industry practices [have] driven continuous efforts to strengthen policies,” they write, urging “continued prudence” to prevent unlawful and egregious behavior.
But policies that are created to punish an outlier may have downstream adverse effects, like the establishment of advisory committees that, in compliance with restrictive rules, are forced to staff themselves with “conflict-free” individuals who may not be true experts.
The medical world continues to make rules and policies aimed at managing conflicts and avoiding the perception of bias. One such change was heralded widely last year, when the National Library of Medicine (NLM) announced that COI statements would be displayed on the front page of an article’s abstract on PubMed.10
Traditionally, COIs were absent from indexed abstracts and only readers with access to the full-text article could view the disclosures and funding sources. The new policy places a COI statement at the bottom of the abstract in an expandable text box, alerting readers to any possible industry influence.
The COI field is indexed and searchable, just like other PubMed fields. A user searching “[hascois]” will find all citations that contain COI statements. Conversely, the field tag “[cois]” can also be used to restrict a search. For example: “merck[cois]” will limit the search to abstracts that include the pharmaceutical company Merck as a sponsor or that are authored by researchers with relationships to the company.
The change was “a congressionally mandated feature to improve the transparency of research,” according to NLM Technical Information Specialist Marie Collins, MS.11
Since the change was implemented, nearly 160,000 COI-containing abstracts have been indexed, but participation among publishers has been sporadic. “We rely on the publishers to include COI information when they submit the rest of the information about the citation,” NLM Librarian Sarah Helson, MLIS, told ASH Clinical News. “While some consider including COI statements in the publication a best practice, NLM does not provide any incentive to publishers to include this data.”
“It is up to the publishers to include this field,” Ms. Helson explained. “It is up to the publishers to determine if they require the authors to include COI information with their publication.”
So far, it is unclear how, why, and which publishers have opted to provide COI information to date. The lack of information prompts other questions, such as how disclosures are made, whether individuals with COIs are permitted to contribute editorial comments, and whether journal editors and publishers are under the influence of industry money.
Less attention has been paid to payments by industry entities to medical journal editors. In a recent study published in The BMJ, investigators found that industry payments to U.S.-based journal editors (identified using publicly available government data) are common, can sometimes be substantial (mean general payment = $28,136; mean research payment = $37,963), and are infrequently disclosed or readily apparent.12
The Disappearing Trust in Medicine
At the extreme end of each side of the COI argument, people sound, well, extreme. So-called “pharmascolds” insist that researchers whose work is funded by industry money are untrustworthy, while others retort that COI rules have gone too far. Is compromise possible?
Take this example to judge for yourself: A recently published, industry-funded study had two co-principal investigators (PIs). One holds the patent on the assay used in the trial and most likely stands to gain financially from the trial producing positive results. The other PI has had an illustrious career in academic medicine and has received continual funding from the National Institutes of Health for decades. She took the company’s money to run the study but has, throughout her career, avoided taking industry money for consulting, speaking, food, travel, etc. All COIs were clearly disclosed in the publication of the study findings.
Is the first PI protecting his patent? Is the second protecting her grants? Or, can we trust that both physician-scientists are, despite competing interests, mostly driven to advance medical knowledge that will save lives? —By Debra L. Beck
- Lo B, Field MJ. Conflict of interest in medical research, education, and practice. National Academic Press, 2009.
- International Committee of Medical Journal Editors. “Author Responsibilities—Conflicts of Interest.” Accessed January 7, 2018, from http://www.icmje.org/recommendations/browse/roles-and-responsibilities/author-responsibilities–conflicts-of-interest.html.
- Steinbrook R. Disclosing the conflicts of interest of US Food and Drug Administration Advisory Committee members. JAMA Intern Med. 2017;177:919.
- Larkin I, Loewenstein G. Business model-related conflict of interests in medicine. Problems and potential solutions. JAMA. 2017;317:1745-6.
- Zuger A. What do patients think about physicians’ conflicts of interest? Watching transparency evolve. JAMA. 2017;317:1747-8.
- Lundh A, Sismondo S, Lexchin J, et al. Industry sponsorship and research outcome. Cochrane Database Syst Rev. 2012;12:MR0000033.
- Lundh A, Lexchin J, Mintzes B, et al. Industry sponsorship and research outcome. Cochrane Database Syst Rev. 2017;2:MR000033.
- Stossel TP. A guide to the anti-innovation “pharmaphobia” narrative for the aspiring innovator. Surgery. 2017;161:305-7.
- Rosenblatt M, Jain S. “Conflict-of-Interest Rules Are Holding Back Medical Breakthroughs.” Harvard Business Review. Accessed January 12, 2018, from https://hbr.org/2017/06/conflict-of-interest-rules-are-holding-back-medical-breakthroughs.
- Collins M. PubMed updates March 2017. NLM Tech Bull. 2017;415:e2.
- Collins M. “PubMed: Redesigned Citation Management to Better Serve PubMed Users.” Accessed January 12, 2018, from https://www.nlm.nih.gov/bsd/disted/video/mla_2017/pubmed.html.
- Liu JJ, Bell CM, Matelski JJ, et al. Payments by US pharmaceutical and medical device manufacturers to US medical journal editors: retrospective observational study. BMJ. 2017;359: j4619.