Am Fam Physician. 2014;89(3):161-167
Related letter: Clarification Regarding AFP's Conflict of Interest Policy.
Author disclosure: No relevant financial affiliations.
Medical journals go to great lengths to ensure that authors disclose potential conflicts of interest. In research journals, financial ties between authors and the makers of drugs and medical devices are inevitable because more than half of research is funded by the pharmaceutical industry.1 However, just about all medical journals permit authors of review articles and editorials to also have such ties. At American Family Physician (AFP), we do not. Our 25-year-old policy is sometimes questioned or misunderstood because it is rare among medical journals. However, I think it is an important way to guard against untoward influence and intentional or unintentional bias in our clinical review articles.2
Our conflict of interest policy states: “To avoid bias or the perception of bias, AFP will not consider manuscripts sponsored directly or indirectly by a pharmaceutical company, medical education company, or other commercial entity, or those written by an author who has a financial relationship with or interest in any commercial entity that may have an interest in the subject matter of the article.”3 Note that this policy is not restricted to specific drugs or devices to which the author may have ties; it covers the topic of the article and a commercial company's potential interest in that topic. Because our review articles are developed for the continuing medical education (CME) of approximately 175,000 readers, we take seriously our obligation to ensure unbiased content. However, we recognize that conflicts of interest are only one potential source of bias, and that intellectual bias (e.g., strongly held opinions, often specialty specific) can also unduly influence the presentation.4
We realize that many physicians see nothing wrong with drug company–sponsored talks or CME articles. So, to better explain the reasoning behind our conflict of interest policy, the accompanying table outlines some of the arguments typically given to justify disclosure as a sufficient safeguard against bias, and the reasons we think disclosure is not enough.2,5–17 The Institute of Medicine,18 the American Medical Association,19 and the Association of American Medical Colleges20 have all recently issued reports calling for elimination or limitation of industry-sponsored CME. For similar reasons, the medical editors of AFP believe that CME should not have any ties to makers of drugs or medical devices. In addition, AFP does not permit placement of advertising next to related clinical content, does not place advertisements within review articles, and does not accept drug company–sponsored supplements.
|Arguments for why disclosure is sufficient||Counterarguments|
|Arguments from industry, professional medical organizations, medical journals, etc.|
|Disclosure alerts readers to the potential for bias and allows them to evaluate the article in that light. Also, studies show that physicians do not perceive industry-sponsored CME activities to be biased.5||Readers are generally not trained to evaluate content for subtle bias or recognize “spin” in clinical presentations (e.g., word choice, framing of content, topic selection). Most physicians are not familiar with every drug in a company's portfolio, so disclosure of company ties is often perfunctory and uninformative. Promotion of a drug often starts years before the drug comes on the market, by promoting a disease, a mechanism, or problems with current therapies. It is not possible to detect bias for a drug that has not been marketed yet.|
|The social science literature suggests that physicians may have difficulty detecting bias in CME.6|
|It is unfair to put the burden on readers, who generally expect content to be unbiased.|
|Disclosure complies with the principle of transparency, and indicates compliance with proper protocol.||Disclosure may convey a false sense of security about the content: “If they're honest enough to disclose their ties to industry, the presentation must also be honest.” Disclosure may paradoxically increase an audience's trust in the information provided, and the likelihood of bias.7–9|
|Disclosure gives CME providers the opportunity to address and “resolve” the conflict of interest via peer review, adherence to “best evidence,” or some other mechanism specified by the Accreditation Council for Continuing Medical Education.10||Peer reviewers are not trained to recognize marketing messages, and may not always spot biases in the presentation (especially if they are subtle and based on such things as word choice, framing of content, or topic selection). Presentations on the prevalence or severity of a condition, “emerging” therapies with a novel mechanism of action, or problems with competing therapies are unlikely to provoke suspicion of bias.|
|Biased presentations can, and usually do, incorporate best evidence, but may be spun to highlight the advantages of targeted drugs or the disadvantages of competing therapies.11|
|Once an article has been written, it can be difficult to alter its slant and tone, or to remove subtle bias, much less to change the choice of topic itself.12 Once biased information enters the medical literature, it can be difficult to alter its distorting effect on the evidence base.2,13|
|Arguments from speakers/authors providing CME|
|“The clinical views I present are mine and mine alone—the drug/device company did not tell me what to say.”||In many cases, the drug/device company does tell their speakers and authors what to say and write. The company also provides them with slides, and writes or reviews and edits the presentation.14|
|Even if this is not the case, the presenter has been courted and selected for a speaker's bureau or authorship because his or her views align well with the company's marketing messages. Speakers who stray too far from a company-friendly viewpoint do not stay on a speaker's bureau for long.15,16|
|“I have ties to so many companies, I can't be biased in favor of any one of them.”||Such presenters are often referred to as “key opinion leaders” who have been courted, groomed, and well paid by drug/device manufacturers.15 Their views align well with multiple companies' marketing messages (e.g., an expert on chronic obstructive pulmonary disease who has ties to the manufacturers of every major medication for the disease). Any presentation elevates the apparent importance of the topic, and each company shares in the downstream benefit of increased diagnosis and treatment. Also, discourse is limited to competing branded drugs. Older drugs, less marketed drugs, generics, and nonpharmacologic therapies may not be discussed.|
|“My presentation doesn't even mention a specific drug or device, so there can't possibly be any bias.”||Presenters are often used to highlight the importance of a certain disease/clinical condition, making it seem more prevalent or severe than previously thought. It has been said that “Pharma doesn't hire doctors to sell drugs; that's a drug rep's job. Pharma hires physicians to sell diseases.”17 For instance, how often was restless legs syndrome discussed before the first drug approved to treat it was being developed? Restless legs syndrome seemed to suddenly become a scourge and epidemic. And now that generic colchicine is no longer available, presentations to physicians on gout, funded by the maker of the 50-fold more expensive brand-name version, have markedly increased.|
|Presenters may focus solely on diagnosis, not treatment; however, this can lead to increased diagnosis and, ultimately, increased treatment.|
|Presenters may focus solely on pathophysiology, not treatment; however, a presentation on incretin physiology and dipeptidyl-peptidase-4 inhibition can lay the groundwork for promoting diabetes treatments that are based on these mechanisms of action.|
EDITOR'S NOTE: Dr. Siwek is editor of AFP.
AFP's conflict of interest policy is independent and different from the American Academy of Family Physicians' conflict of interest policy.