Every month it seems that news headlines feature a scandal involving a physician conflict of interest (COI). Three articles1–3 in this issue of Academic Medicine illustrate how the medical profession and academic health centers (AHCs) can take the initiative in addressing COIs. The findings and conclusions of these papers are on the whole consistent with the recommendations of the Institute of Medicine (IOM) report, “Conflict of Interest in Medical Research, Education, and Practice,” issued in April 2009.4,5 This consensus report was evidence-based and peer-reviewed. Its recommendations covered all aspects of medicine and applied to all physicians, researchers, and medical organizations, including AHCs and professional membership societies. The congruence between these articles and the IOM report suggests that agreement is emerging on key issues regarding COIs. This commentary will first summarize points of agreement and then discuss areas in which the IOM report goes beyond the policies and practices described in these papers.
These three papers make several important overarching points that are consistent with the IOM report.
First, loss of trust is an important risk in COIs, as Camilleri and Parke2 recognize. Organizations that carry out medical research, education, and patient care should develop COI policies, analyze specific relationships with industry that faculty or departments report, and assess the risk of undue influence over professional judgment and loss of trust. It may be necessary to manage or prohibit some significant COIs.
Second, we need better empirical evidence to improve COI policies. The available evidence regarding the extent and consequences of significant COIs is weak. We also know little about the impact—beneficial or adverse—of COI policies. These important articles begin to address this evidence gap. As Steinman and colleagues3 carefully point out, their findings do not prove that industry support of continuing medical education (CME) does not lead to bias. As they caution, other explanations for their findings are possible, and their findings may not generalize to other CME providers or to more in-depth methods of assessing bias. They suggest that “more research is needed to systematically investigate the presence and impact of bias in CME.”3 The IOM report recommends more empirical research on COIs and COI policies.
Third, education and discussion about COIs are essential. As the rich case study by Dubovsky and colleagues1 describes, COI policies are more effective if they are supported by the academic community rather than viewed as imposed externally. Trainees and faculty who are subject to COI regulations need to understand the ethical concerns raised by COIs and the arguments for and against various COI policies. As part of their mission, AHCs need to teach trainees how to critique promotional literature and to interact appropriately with industry representatives. The model of a departmental symposium at which faculty teach about COIs and respond to presentations of peer-reviewed articles by industry representatives or industry-sponsored scientists is promising. Other innovative educational efforts should also be encouraged, formally evaluated, and disseminated.
Additional Recommendations by the IOM
On several important COI issues, the IOM report goes beyond the conclusions of these articles.
Disclosing financial relationships
Disclosure of financial relationships with industry is an essential but limited first step. AHCs require faculty to disclose their relationships with industry; without such disclosure, AHCs cannot identify problematic relationships and manage or eliminate them if needed. However, disclosure is useful only if it provides sufficient information to assess whether a relationship presents an unacceptable risk of undue influence or loss of public trust. Some commonly used categories of disclosure, however, do not allow such an assessment. For example, “consulting” could be scientific consulting or helping drug and device companies increase market share. Scientific consulting usually provides value to society through improving the quality of medical research; however, much marketing consulting turns physicians into spokespersons for the company or promotes a company's products without regard to the value added to patient care.
Another problem with disclosure is that physicians often must make multiple disclosures—not just to their AHC but also to funding agencies, medical journals, and CME audiences, which often have different requirements, categories, and formats. Such variation may lead to inconsistent disclosures—for example, when multiple publications from a research team make inconsistent COI disclosures. Such inconsistency may lead to allegations of misbehavior, when actually it may be caused by journals having different disclosure policies. Standardizing disclosure would not only help those who must judge whether a disclosed relationship is a significant COI but would also reduce administrative burdens on physicians who must make disclosures to multiple organizations.
A final issue is public reporting by industry of payments to physicians. Several states require companies to publicly report the payments they make to physicians, and some drug and device companies report such payments either voluntarily or as part of settlement agreements with prosecutors.6 These public reports have revealed that some academic physicians have failed to disclose to their AHCs, as required, payments from industry. In some highly publicized cases, unreported payments totaled hundreds of thousands of dollars. In response, “Sunshine” bills have been introduced in Congress.6 The IOM recommended that Congress require drug, device, and biotechnology companies to report on a public Web site all payments to physicians, researchers, and medical institutions, including AHCs, professional societies, and providers of accredited CME. These reports should be in standardized format, allowing total payments to an individual or institution to be readily aggregated. Such public reporting would enhance transparency and accountability and allow AHCs to verify disclosures that individual faculty are required to make.
Camilleri and Parke, in their comprehensive discussion of COIs facing professional membership organizations, highlight the important role these organizations play in developing practice guidelines.2 They illustrate how financial relationships between industry and the organization or the committee members developing the guidelines may not be disclosed when the guidelines are published. The IOM report also emphasized the importance of disclosure in the context of practice guidelines but went further, because disclosure does not eliminate the risk of undue influence. The IOM committee recommended that the starting point should be to try to appoint a committee whose members have no significant COIs. If it is necessary to appoint members with conflicts in order to have needed expertise, those members should play a restricted role. They should not serve as chairpeople, draft guidelines, or vote.
Unresolved issues regarding COI may be best addressed by a consensus development process that involves multiple stakeholders, as Dubovsky and colleagues1 illustrate. Policies developed with input and collaboration from physicians and institutions—including industry—that must live with them may have fewer unanticipated adverse consequences and administrative burdens than would policies promulgated without such real-life experience. Moreover, buy-in from front-line physicians and residents helps to develop a culture of accountability. An example of a successful collaborative consensus process is the Association of American Medical Colleges and Association of American Universities report7 on COIs in human subjects research, which involved senior executives of pharmaceutical companies. Participation of public or patient representatives in such consensus activities would also be beneficial. The IOM recommended additional consensus development processes to address two other important issues: a recommendation for a standardized format for disclosure of financial relationships, and a new model for accredited CME that provides high-quality education without undue influence from industry. In these deliberations, empirical studies like the ones in this issue will be valuable.
A consensus process must grapple with the details that bedevil COI policies. Dubovsky and colleagues1 and Camilleri and Parke2 suggest that unrestricted grants from industry help to protect against undue influence. However, such policies need to specify to whom the unrestricted grants may be awarded—an individual faculty member, a department, the AHC, or a consortium of AHCs? A firewall between the sponsoring company and the recipient of the payment is likely to decrease the risk of undue influence. However, with such processes there will also be increased administrative complexity. How much separation between the funder and project is optimal? Such details remain to be worked out, and vigorous and candid discussions among various institutions and stakeholders are needed to forge effective policies that also minimize undue administrative burdens and unintended adverse consequences.
Managing institutional COI
Institutional COIs require attention in addition to COIs for individual physicians and researchers. Camilleri and Parke2 illustrate how organizations and organizational leaders may have relationships with industry that carry an unacceptable risk of undue influence or loss of trust. Such institutional COIs have received relatively little attention. To address this, the IOM report recommended that institutions establish committees to address institutional COIs. These committees should report to the board of trustees or governing body and should be composed of members who themselves have no conflicts. (This committee would be distinct from the committee that reviews conflicts for individual physicians.)
Closer collaborations between academia and industry may be desirable in some areas, particularly the development of new drugs and products, which benefit patients. The IOM committee heard testimony from leaders of AHCs that the model of individual consulting by faculty with a range of companies is suboptimal, and tighter organizational relationships would be more effective. An important challenge is to design new models of academic–industry collaboration for developing new therapies that foster the goal of improved care for patients without leading to undue influence by companies over independent scientific judgment.
Honoring the educational mission
AHCs have special professional responsibilities for fostering an appropriate learning environment for students and trainees. Accreditation agencies require medical schools and residency programs to promote the development of professional attributes and critical appraisal of evidence and provide appropriate role models and mentoring.4 The IOM report determined that it would be inconsistent for faculty members to require students and residents to think critically but then serve on speakers' bureaus that require physicians to use slides and text prepared by the company. The IOM therefore recommended that physicians not serve on such speakers' bureaus.
Taking the Lead
The articles in this issue illustrate how AHCs and professional membership organizations can take the lead in improving COI policies. Dubovsky and colleagues1 show the value of building a culture of professionalism and caution against trying to “legislate behavior with regulations that are impossible to enforce.” However, sometimes leaders need to be ahead of their faculty or members. A top-down visionary policy may help change physician attitudes, expectations, and behaviors, particularly when coupled with discussion and education. There are several reasons for leaders in academic medicine to take the lead in forging strong COI policies. Self-regulation may prevent external regulations that may impose heavy bureaucratic burdens. Physicians might develop better COI policies than legislators or regulatory agencies because they are sensitive to the potential unintended consequences of policies. The Association of American Medical Colleges has taken the lead in recommending thoughtful COI policies regarding human-participants research. These articles show that physician leaders, AHCs, and professional societies can continue to take the lead by developing, implementing, enforcing, and improving thoughtful COI policies.
Supported by National Institutes of Health (NIH) Grant Number 1 UL1 RR024131-01 from the National Center for Research Resources (NCRR), by National Institutes of Health Grant MH062246 and the NIH Roadmap for Medical Research, and by the Greenwall Foundation. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of NCRR or NIH.
1Dubovsky SL, Kaye DL, Pristach CA, DelRegno P, Pessar L, Stiles K. Can academic departments maintain industry relationships while promoting physician professionalism? Acad Med. 2010;85:68–73.
2Camilleri M, Parke II DW. Perspective: Conflict of interest and professional organizations: considerations and recommendations. Acad Med. 2010;85:85–91.
3Steinman MA, Boscardin CK, Aguayo L, Baron RB. Commercial influence and learner-perceived bias in continuing medical education. Acad Med. 2010;85:74–79.
4Lo B, Field MJ. Conflicts of Interest in Medical Research, Education, and Clinical Practice. Washington, DC: National Academies Press; 2009.
5Steinbrook R. Controlling conflict of interest—Proposals from the Institute of Medicine. N Engl J Med. 2009;360:2160–2163.
6Steinbrook R. Online disclosure of physician–industry relationships. N Engl J Med. 2009;360:325–327.
7AAMC–AAU Advisory Committee on Financial Conflicts of Interest in Clinical Research. Protecting Patients, Preserving Integrity, Advancing Health: Accelerating the Implementation of Conflict of Interest Policies in Human Subjects Research. Washington, DC: American Association of Medical Colleges; 2008.