NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Institute of Medicine (US) Committee on Conflict of Interest in Medical Research, Education, and Practice; Lo B, Field MJ, editors. Conflict of Interest in Medical Research, Education, and Practice. Washington (DC): National Academies Press (US); 2009.

Cover of Conflict of Interest in Medical Research, Education, and Practice

Conflict of Interest in Medical Research, Education, and Practice.

Show details


Patients and the public benefit from constructive collaboration between academic medicine and pharmaceutical, medical device, and biotechnology companies. At the same time, medical leaders, public officials, public interest groups, and others have raised concerns about the risks associated with the extensive financial ties that link industry with the individuals and institutions that carry out medical research, medical education, patient care, and practice guideline development. The risks are that individual and institutional financial interests may unduly influence professional judgments involving these primary institutional missions. Such conflicts of interest threaten the integrity of scientific investigations, the objectivity of medical education, the quality of patient care, and the public’s trust in medicine.

The benefits of collaboration with industry are most evident in biomedical research. New medications and medical devices have significantly improved outcomes for people with a range of serious and common diseases, including—among many others—coronary artery disease, congestive heart failure, hypercholesterolemia, several types of cancers, and peptic ulcer disease. Such successful products result from a long, complex, and often unpredictable process of translating basic science discoveries into new preventive, diagnostic, or therapeutic products and services. The basic discoveries often come from the laboratories of university and government scientists; but their development into actual products available to clinicians and patients usually depends on the technical, production, and financial resources of pharmaceutical, medical device, or biotechnology companies. It is estimated that it takes an average of 15 years and more than $800 million to discover and develop a new drug, and only about 10 percent of the drugs that enter clinical testing are actually approved for marketing (DiMasi et al., 2003, 2004; FDA, 2004a). Chapter 4 and Appendix E further examine the nature and value of university-industry collaboration in medical research.

With the benefits of research collaboration and the expansion of financial relationships in other areas have also come conflicts of interest and evidence of bias. For example, in clinical research, unfavorable results in some major industry-sponsored trials have been withheld from publication, thus distorting the totality of the findings included in the scientific literature. These trials involved drugs commonly prescribed for arthritis, depression, and elevated cholesterol levels, among other medications (Wright et al., 2001; Gibson, 2004; Whittington et al., 2004; Kastelein et al., 2008). Not publishing negative results undermines evidence-based medicine and puts millions of patients at risk for using ineffective or unsafe drugs. One striking case involves the withholding of negative findings from pediatric clinical trials of the effects of selective serotonin reuptake inhibitors on depression (Healy, 2006; Turner et al., 2008). Findings were withheld so frequently that although one meta-analysis of the published literature (ACN, 2004) concluded that these drugs were safe and effective, another meta-analysis (Whittington et al., 2004) that took into account unpublished as well as published data concluded the opposite: that the risks outweigh the benefits for all but one drug in this class of antidepressants. A recent analysis found that more than half of the trials used to support Food and Drug Administration approval for the marketing of a drug or medical device had not been published within 5 years after approval (Lee et al., 2008). In addition, litigation has revealed documents that link bias in publications to financial relationships with pharmaceutical manufacturers (Steinman et al., 2006; Psaty and Kronmal, 2008; Ross et al., 2008). As discussed in Chapter 4, the statistical associations involving industry sponsorship do not prove causality, but they do raise serious concerns about undue industry influence and have prompted a range of responses, including the creation of publication protections in university-industry research contracts and the issuance of regulations and other requirements that the results of clinical trials be reported in clinical trial registries.

In medical education, it is particularly troublesome when a faculty member is a promotional speaker for a pharmaceutical, medical device, or biotechnology company or agrees to be listed as an author for a ghostwritten publication. This is because faculty members are expected to present unbiased information and objective assessments of the scientific literature and to help medical students, residents, and fellows develop life-long habits of exercising independent judgment and critically evaluating scientific evidence. They are also expected to serve as role models of professionalism. These expectations may be undermined by some financial relationships between faculty and industry and by failures to disclose such relationships.

In clinical care, patients need to trust that their physicians’ recommendations are not distorted by commercial interests. Such trust may contribute to the healing process and to patients’ sense of well-being. Some financial relationships between physicians and industry raise concerns about the risk of bias in clinical decisions. For example, companies have paid some physicians large but generally undisclosed amounts to give talks to other physicians, whose prescribing practices were then tracked by company sales representatives (Elliott, 2006; Carlat, 2007). Drug samples and other gifts to physicians by company sales representatives are major marketing tools that evidence suggests influence prescribing choices (see Chapter 5). Furthermore, during the last decade, several federal prosecutions alleging that companies made illegal payments to physicians to induce them to use the companies’ drugs or medical devices have led to settlements in which the companies agreed to modify various marketing practices and, in some cases, to post publicly their payments to physicians (see Chapter 6). The prevalence of illegal payments is not known.

Another area of concern is clinical practice guidelines. Clinical practice guidelines influence patient care, quality and performance standards, and reimbursement for health care professionals and institutions. If a risk exists that guidelines are biased or may be viewed as biased in favor of the products of the companies that sponsored the guideline development process or companies that have financial relationships with experts involved in the process, then patients may be harmed and users’ trust in the guidelines may be undermined. Evaluating the potential for such bias is often difficult, however, because many entities that develop practice guidelines do not have clear conflict of interest policies for this activity, do not disclose their funding sources, and do not reveal the relevant financial relationships or conflicts of interest for the experts responsible for developing a set of guidelines. A review of clinical practice guidelines that do include information on financial relationships of the participants suggests that conflicts of interest are common (for examples, see Chapter 7 and guidelines posted on the website for the National Guideline Clearinghouse).

Conflicting interests are, to some degree, both ubiquitous and difficult to avoid. For example, regardless of how they are paid for their services (e.g., on a fee-for-service or a capitated basis), physicians will face some incentives that may at times conflict with their professional responsibility to provide care that best serves their patients’ interests. Medical school faculty may face conflicts in the time and energy that they devote to each element of their academic responsibilities—research, teaching, and clinical care.

Many conflicts are unavoidable features of multifaceted professional roles and obligations. Others are optional, for example, the creation of a consulting or a speaking agreement with a pharmaceutical, medical device, or biotechnology company. These kinds of financial relationships with industry are the focus of this report.

As explained further in Chapter 2, this report specifically defines a conflict of interest as existing when an individual or institution has a secondary interest (e.g., an ownership interest in a start-up biotechnology company) that creates a risk of undue influence on decisions or actions affecting a primary interest (e.g., the conduct of objective and trustworthy medical research). This definition frames a conflict of interest in terms of the risk of such undue influence and not the actual occurrence of bias.

Some argue that concerns about conflicts of interest are overstated and that policy responses have been excessive, inconsistent, and unduly burdensome on physicians and researchers (see, e.g., Stossel [2005, 2007], Duvall [2006], Borgert [2007], and Bailey [2008]). According to that viewpoint, problems related to conflicts of interests are rare. Thus, the vast majority of scientists, educators, and clinicians should not be subject to onerous conflict of interest rules and regulations because of a few miscreants. The argument continues that burdensome rules and regulations stifle valuable collaborations between industry and academia. Moreover, allegations of conflict of interest inappropriately call into question the motives and integrity of individual scientists and clinicians, because a financial relationship related to one’s research, teaching, or clinical practice does not prove the actual presence of bias in decisions or judgments. Consequently, it would be better to focus on detecting and minimizing bias rather than on disclosing, limiting, or managing financial relationships with industry. Furthermore, some of the intended beneficiaries of conflict of interest policies—for example, research participants—do not seem to be concerned about the financial interests of the investigators (see, e.g., Hampson et al. [2006] and Weinfurt et al. [2006a] and the further discussions in Chapters 3 and 4). Another criticism is that the focus on conflicts of interest related to financial ties with industry distracts attention from other threats to objectivity and public trust, such as career ambitions, a desire for recognition, intellectual bias, personal ties, and physician payment methods.

As discussed in Chapter 2, many objections to conflict of interest policies are based on misunderstandings of their purpose and nature. If they are correctly explained, the policies should not be seen as impugning anyone’s motives. They are, in fact, a way of avoiding intrusive investigations into people’s motives. They also protect against bias or distrust when other methods (e.g., assessments of actual bias after the fact) are not feasible or sufficient. Although other secondary interests may inappropriately influence professional decisions and additional safeguards are necessary to protect against bias from such interests, financial interests are more readily identified and regulated.

Opposition to conflict of interest policies often focuses on what might be lost with further restrictions on ties to industry. For example, eliminating industry support for accredited continuing medical education might result in increases in the fees that physicians must pay for such education, a reduction in the number of accredited courses, and a drop in income for institutions that provide continuing medical education. To cite another example, if universities insist on contract terms that restrict a company’s ability to withhold or censor research findings, then companies might move more research contracts elsewhere (e.g., to contract research organizations or overseas research centers that do not have such restrictions). Similarly, some faculty members may leave a university if that university restricts faculty members’ financial ties with industry. Such losses (costs) tend to be immediate, easily identifiable, and tangible.

In contrast, the costs of conflicts of interest and the benefits of mitigating or eliminating them tend to be less tangible, less immediate, and more diffuse. Eliminating direct industry funding of continuing medical education, for example, could increase evidence-based physician prescribing practices, which over time could reduce wasteful health care spending and improve the quality of patient care, but demonstrating such causal relationships could be difficult or impossible. Another benefit of dealing with conflicts of interest that is even harder to define and document but that is significant could be the maintenance of public trust in medical professionals and institutions. Indeed, the maintenance of trust is a major objective of conflict of interest policies across a broad range of professions, in addition to medicine (see Appendix C).

Research suggests that people are generally not good at making trade-offs between costs and benefits that are immediate and tangible and those that are less immediate and less tangible (for a review, see Rick and Loewenstein [2008]). People tend to put a disproportionate emphasis on costs and benefits that are immediate and tangible. For example, the impact of a single, free drug company-sponsored lunch on a physician’s prescribing practices or on public trust may be small to insignificant, but the cumulative consequences of many lunches to many physicians may be great. The human tendency to overweight the immediate and tangible compared with the delayed and intangible thus complicates efforts to understand and respond to conflicts of interest.


This Institute of Medicine (IOM) report examines the extent of financial relationships with industry and conflicts of interest in medical research, education, and practice and in the development of clinical practice guidelines. It reviews policies that have been adopted or proposed to avoid or manage these conflicts and recommends steps that can be taken to improve the design, implementation, and evaluation of these policies. The report builds on the analyses and recommendations of other groups. It is different, however, in its focus on conflict of interest across the spectrum of medicine and in its identification of overarching issues and strategies that can be used to limit the negative effects of conflicts of interest while preserving the benefits of collaboration with industry, particularly in moving discoveries from basic science into improved patient care. The report has several broad messages.

  1. The goal of conflict of interest policies in medicine is to protect the integrity of professional judgment and to preserve public trust rather than to try to remediate problems with bias or mistrust after they occur.
    In all aspects of medicine, judgments must inevitably be made, and reasonable people will disagree over some judgments. Both science and medicine depend on public trust that judgments are made in good faith and are not unduly influenced by the financial interests of professionals or the institutions with which they are affiliated. Well-formulated and well-explained conflict of interest policies can help identify individual and institutional relationships that could reasonably be questioned and allow judgments to be made prospectively about whether particular relationships should be eliminated, permitted, or managed.
    It is prudent to require physicians and medical researchers to avoid or manage situations that offer a significant possibility of bias rather than to wait to investigate allegations of bias or misconduct until after they occur. Investigations performed to uncover bias after the fact can be difficult, time-consuming, and heavily burdensome for all involved. Furthermore, when bias occurs in clinical research, medical education, or practice guideline development, it can harm research participants or patients, waste scarce resources, and damage individual and institutional reputations, including the reputations of those whose relationships with industry are appropriately structured and disclosed and serve the public good. If trust is eroded by continuing revelations of withheld negative research findings, promotional relationships disguised as consulting services, and similarly troublesome situations, it may be hard to restore.
  2. Disclosure of individual and institutional financial relationships is a critical but limited first step in the process of identifying and responding to conflicts of interest.
    Institutions that carry out medical research, medical education, patient care, and practice guideline development depend on individuals’ disclosure of their financial relationships with industry. Without such disclosure, institutions will lack the information they need to identify and assess conflicts of interest and determine what additional steps—such as eliminating or managing the conflicting interest—may be necessary. Disclosure by institutions is likewise important because institutions may also have financial relationships that create conflicts of interest. The disclosures need to be sufficiently specific and comprehensive to allow an initial assessment of the risk of undue influence. At the same time, the harmonization of disclosure requirements and procedures can reduce administrative burdens for researchers and physicians who must make multiple disclosures to different institutions for different purposes.
    Disclosure does not resolve or eliminate conflicts of interest. Institutions must also evaluate and act upon the disclosed information. Actions might include the elimination of a relationship, further disclosure (e.g., to research participants, patients, or the public), or other types of management (e.g., restricting the participation of a researcher with a conflict of interest in the enrollment of study participants or analysis of study data).
  3. Conflict of interest guidelines and policies can be strengthened by engaging physicians, researchers, and medical institutions in developing policies and consensus standards.
    For conflict of interest policies to be truly effective, buy-in from physicians and researchers will be important, so that they regard conflict of interest policies as a means to help them fulfill their professional responsibilities and not as externally imposed nuisances. Furthermore, if those who are subject to conflict of interest policies participate in policy development, they may suggest how the policies can be framed to avoid unintended adverse consequences and undue administrative burdens. In several areas in which substantial policy variation or disagreement exists and greater agreement is needed, the report proposes the creation of consensus development panels with a broad range of participants, including consumer representatives. Two areas that are ripe for consensus building involve the standardization of information that physicians and researchers are required to disclose (Chapter 3) and the development of a new system of financing continuing medical education (Chapter 5).
  4. A range of organizations—public and private—can promote the adoption and implementation of conflict of interest policies and help create a culture of accountability that sustains professional norms and promotes public confidence in professional judgments.
    Institutions that carry out medical research, medical education, clinical care, and practice guideline development have the primary responsibility for addressing conflicts of interests in these activities. These institutions do not, however, act in isolation. Rather, they interact with many other organizations—including academic and trade membership associations, accreditation and certification bodies, patient advocacy groups, health plans, and federal and state agencies—that have a stake in reducing the severity of individual and institutional conflicts of interest. As discussed in Chapter 9, these organizations can create incentives to encourage institutions to adopt and implement policies that are consistent with the recommendations of this committee and other organizations, such as the Association of American Medical Colleges, the Association of American Universities, and the International Committee of Medical Journal Editors. Such incentives would encourage and reinforce professional responsibility and promote public trust.
  5. Research on conflicts of interest and conflict of interest policies can provide a stronger evidence base for policy design and implementation.
    The current evidence base for conflict of interest policies is not strong. A program of research on conflicts of interest and conflict of interest policies could provide policy makers with a better evidence base and a basis for understanding the nature and consequences of conflicts of interest in different situations. It could likewise guide policy makers as they revise policies and procedures to make them more effective and less burdensome.
  6. If medical institutions do not act voluntarily to strengthen their conflict of interest policies and procedures, the pressure for external regulation is likely to increase.
    The continuing publicity about conflicts of interest in medicine and the failure of individuals and institutions to adhere to conflict of interest policies has prompted calls for government regulation. Indeed, this report recommends some areas for government action, but it also emphasizes that risks as well as the potential benefits of regulation should be considered.

Origins of the Study

This study grew out of discussions within the IOM about the threats to objectivity and public trust in biomedical research and medicine created by conflicts of interest related to certain types of financial relationships between industry and researchers based in universities and federal agencies. Consideration of the topic was further stimulated by inquiries from groups outside the IOM about whether the IOM would examine conflicts of interest and industry ties as they might affect the publication of research and the development of clinical practice guidelines. In response, the IOM proposed a broad-ranging study that would examine conflicts of interest across medical research, medical education, clinical practice, and practice guideline development.

The IOM appointed a 17-member committee to oversee the study and develop the study report. (See Appendix A for more information about study-related activities.) Consistent with its charge, the committee

  • examined financial relationships with industry and conflicts of interest in medical research, education, practice, and practice guideline development and
  • developed analyses and recommendations to inform the design and implementation of policies for the identification and management of conflicts of interest in these contexts without damaging constructive collaborations with industry.

To address this broad charge, the committee consciously adopted a crosscutting perspective and tried to view medicine as a complex system with many interacting components and interested parties. It drafted its report for a diverse audience of academic, scientific, professional, medical institution, industry, consumer, news media, and government leaders. Their understanding of the hazards of conflicts of interest and the elements of effective, balanced policies aimed at preventing conflicts of interest from occurring is essential.

During the course of its work, the committee searched for and assessed empirical evidence relevant to its charge, and it read and heard a wide range of views. The analyses and recommendations in this report reflect the committee’s conscientious effort to understand and take these views into account. The committee also examined how conflicts of interest are handled in other professions (see Appendix C).

Focus and Concepts

Given the breadth of its charge, the committee focused on conflicts of interest involving physicians, biomedical researchers, and senior institutional officials. These individuals have been at the center of most controversies about conflicts of interest and most proposals for policy change. Many of the conclusions and recommendations presented in this report will, however, be generally relevant to nursing, pharmacy, dentistry, and other health professions and to other health researchers. In some cases, institutional policies may extend beyond researchers, professionals, and senior officials. For example, professional society policies governing members of a panel developing clinical practice guidelines will cover all members, including consumers, patients, and the representatives of health insurers.

This report generally uses the term institutions to refer to academic medical centers; professional societies; patient or consumer groups; and other entities that carry out medical research, provide medical education and clinical care, or develop clinical practice guidelines.1 The report also distinguishes (particularly in Chapter 9) supporting organizations, such as accrediting agencies and state licensure boards, that may create incentives for institutions to adopt and implement effective and credible conflict of interest policies or codes of conduct and for individuals to follow these policies or codes. Some entities, such as medical journals, cross these definitional boundaries and are covered by recommendations related to both institutions and organizations.

Reflecting the discussions that led to this study and the emphasis of much research, press coverage, and public and professional debate, this report emphasizes financial interests and relationships involving pharmaceutical, medical device, and biotechnology companies that make—or that are developing—medical products used in patient care. (For convenience, the report sometimes refers to these companies as “industry” or “medical product companies,” although some start-up biotechnology and other companies may not yet have products approved for marketing.) Other interests, such as the desire for public recognition, may also threaten objectivity and public trust, but financial interests are the central focus of conflict of interest debates and policies.

Notwithstanding the prominence of medical product companies in discussions of conflicts of interest in medicine, the committee recognized that significant conflicts of interest in medical research, education, and practice can be created by financial relationships involving many other kinds of companies. These include health insurers; prescription drug and other benefit management companies; law firms; investment companies; and suppliers of food, office supplies, and other nonmedical goods and services. Much of the discussion in this report about the adoption of policies and the disclosure of information should be relevant to financial relationships involving these other commercial entities.

The committee also understood that serious conflicts of interest may arise from the way in which physicians are paid for their clinical services and from physician ownership interests in hospitals, diagnostic centers, and facilities. The IOM did not plan this study to investigate these issues, but they are briefly discussed in Chapter 6.

Although the analyses presented in this report build on a series of reports on responsible research and integrity in science issued by the IOM and the National Research Council, those earlier reports did not examine conflict of interest in depth. Nonetheless, they provide useful perspectives. In particular, the reports Integrity in Scientific Research (IOM/NRC, 2002) and Responsible Research (IOM, 2003) underscore the importance of creating organizational and social environments that support and encourage responsible and ethical behavior by individuals and institutions. This report also builds on recommendations made in other reports that called for the undertaking of more and better comparative effectiveness studies and other steps needed to build and communicate the evidence base for clinical practice (see, e.g., previous IOM studies [1991 see, e.g., previous IOM studies [2007]). One recommendation of this report (Recommendation 9.2) is that the evidence base for conflict of interest policies needs to be strengthened to help policy makers identify effective policies and avoid unwanted consequences.


Concerns about conflicts of interest have a long history; and the responses to these conflicts have evolved as relationships with industry in medical research, education, and practice have grown more frequent and more complex. They have also evolved as different responses to such relationships have been tried and found to be in need of modification.

The following brief review indicates, first, that both government regulations and voluntary codes of conduct often follow the discovery of instances of questionable or inappropriate relationships and conduct. This is similar to the pattern in other areas, such as the oversight of research involving human participants.2 Second, government scrutiny of conflicts of interest may stimulate private, voluntary efforts by academic and other institutions to deal with problems and avoid regulation. Third, when these efforts are found to be wanting and government acts, legislators and administrators may still delegate to regulated institutions many of the details of policy development, implementation, and monitoring.

Expanding Relationships Between Industry and Medicine

Relationships between physicians, medical researchers, and medical schools and companies that produce medical products have a long history, as have efforts to encourage such relationships. For example, in the early 1920s, Eli Lilly worked with researchers at the University of Toronto to manufacture insulin in quantities adequate for research and then clinical use; the university also granted royalty-free patents to other companies to expand the drug’s availability worldwide (Thayer, 2005). In 1925, the National Research Council (which the National Academy of Sciences established at the request of President Woodrow Wilson to organize scientific research) created a short-lived National Research Fund that raised money from private companies to support research in academic institutions (Swann, 1988).

The mixing of product marketing and medical information for physicians likewise has a lengthy history (see, e.g., Podolsky and Greene [2008]). More than a century ago, a review in the Chicago Medical Recorder of Merck’s Manual of the Materia Medica (now the Merck Manual of Diagnosis and Therapy) observed: “[a]lthough this little book is gotten out by a manufacturing firm and with some view towards its advertising value, it nonetheless is of such merit that it is deserving of mention in this column” (quoted by Lane and Berkow [1999, p. 112]). Then, as now, recognition of the value of industry contributions can coexist with unease about commercial motivations and potential bias.

Professional societies and the medical products industry also have longstanding relationships, for example, industry advertising in journals sponsored by medical societies. As early as the late 1940s and early 1950s, the American Medical Association (AMA) began to market information from its new physician database to pharmaceutical companies and to commission studies of the effectiveness of different marketing techniques, the results of which were sent to pharmaceutical and device companies—along with pamphlets promoting advertising in the Journal of the American Medical Association (Greene, 2007). This AMA business has generated some controversy and is discussed further in Chapter 6 (see also Steinbrook [2006]).

Biomedical research saw a marked expansion of government funding after World War II. By 1965, spending by the National Institutes of Health (NIH) and other federal agencies accounted for almost two-thirds of the total funding for biomedical research, whereas it was about 7 percent in 1940 (Ginzberg and Dutka, 1989). Then, in the late 1970s, the balance began to shift toward commercial funding. By the turn of the 21st century, the share of health research and development funding accounted for by industry reached 55 to 60 percent (see Chapter 4). New relationships and collaborations between universities and industry during the late 1970s and 1980s were stimulated by a combination of economic conditions, pressures on the federal budget, scientific discoveries, needs for expertise outside universities, and other factors, including legislative incentives for universities to develop discoveries commercially. A Congressional Research Service report noted that another factor in universities’ pursuit of industry funding was a “desire to lessen the regulations associated with the expenditure of Federal dollars” (Johnson, 1982, p. 2).

Industry has also become a major source of funding for medical education, particularly continuing medical education. Between 1998 and 2007, the share of continuing medical education provider income accounted for by commercial sources, excluding advertising and exhibits, grew from 34 to 48 percent, with higher rates for some providers, such as for-profit education and communication companies and medical schools (ACCME, 2008a). Through their support for professional society journals and meetings, pharmaceutical and medical device companies are also important sources of income for professional societies, often accounting for 30 to 50 percent or more of the total income of professional societies (see Chapter 8).

Growing Concerns About Relationships with Industry

As they have evolved, relationships between industry and medicine have brought many benefits, primarily in biomedical research. They have also raised concerns that such relationships can—if they are not properly managed—threaten the objectivity of medical research, education, and practice and undermine public trust in critical American institutions.

Table 1-1 lists some notable events in the emergence of relationships with industry and conflict of interest as a concern in medicine. They include congressional hearings in the 1980s that posed questions about whether conflicts of interest were reducing openness in universities and biasing the advice given to policy makers. A Congressional Research Service report on the commercialization of academic biotechnology research observed that “the credibility of university scientists associated with industry has fallen into question” (Johnson, 1982, p. 5). An article in Science from the same period titled The Academic-Industrial Complex (Culliton, 1982) summarized the ethical concerns that these relationships presented to university administrators and faculty:

TABLE 1-1. Timeline of Selected Events Relevant to the Evolution of Conflict of Interest Principles, Policies, and Practices.


Timeline of Selected Events Relevant to the Evolution of Conflict of Interest Principles, Policies, and Practices.

How can universities preserve open communication and independence in the direction of basic research while also meeting obligations to industry? Is it acceptable for one corporation to dominate research in an entire department? Are there adverse consequences in terms of collaboration among faculty in various departments if one group must worry about protecting corporate rights to licenses? Will patent and licensing provisions delay scientific publication? Should corporate sponsorship be subject to peer review? Under what conditions may a faculty member have an equity position in industry? Do such ties compromise loyalty to university teaching and research? Will graduate students be compromised or poorly served? Will extensive corporate ties erode public confidence in university faculty as disinterested seekers of truth? (Culliton, 1982, p. 961)

Concerns about conflict of interest beyond the research context were also growing during the 1970s and 1980s. Some concerns related to questions about commercial bias in scientific publications. Others focused on physician referral of patients to specialty centers in which they had a financial interest and on the prevalence of company-provided gifts, lavish entertainment, marketing activities that were disguised as scientific information, and other relationships in both community and academic medical settings.

Evolving Public and Private Responses to Concerns About Conflict of Interest

In the early 1960s, in recognition of the importance of outside advice on complex scientific and policy questions from objective experts, the fed eral government (through a presidential memo) established policies to limit conflicts of interest among special government employees serving as advisory committee members and consultants. In the academic community, the American Association of University Professors (AAUP) and the American Council on Education (ACE) issued a joint statement, On Preventing Conflicts of Interest in Government-Sponsored Research at Universities (AAUP/ACE, 1965). The statement spoke of the importance of university-industry relationships but stressed the need to protect the integrity of educational institutions in the face of ties between these institutions and both government and industry. It called for universities to advise government research agencies about the steps they were taking to avoid problems. According to McNeil and Roberts (1991), this statement forestalled government regulation and led to the adoption of policies by most major research universities of “very general guidelines” on conflict of interest that relied on faculty-initiated disclosure (p. 149). By 1967, a number of universities, including Yale, Harvard, Stanford, Michigan, Chicago, Minnesota, and California, had adopted conflict of interest policies that had been approved by the Federal Office of Science and Technology (Wellman, 1967).

A few years after AAUP and ACE issued their statement and after some incidents that raised concerns about bias and conflict of interest, the National Academy of Sciences approved a letter, On Potential Sources of Bias, which it issued in 1971. The letter asked members of the organization’s scientific study committees to describe financial and other factors that in their judgment “others may deem prejudicial” (quoted in Parascandola [2007]). According to Parascandola, “[s]cientists universally opposed the policy, however, for a range of reasons—while some argued that all experienced and knowledgeable experts were inherently conflicted, others were offended at the suggestion that any expert could be biased” (p. 3774).

Such negatives responses to conflict of interest policies continue. Nonetheless, the adoption of policies has expanded as the scope and complexity of relationships with industry have increased and instances of questionable or illegal behavior have accumulated—with the attendant negative publicity.

In 1984, the Association of American Universities declined to propose conflict of interest policies for its members, but it did undertake a survey of university policies (OTA, 1984; McNeil and Roberts, 1991). It found that 19 of the 46 responding institutions relied on faculty members to determine whether they had a possible conflict of interest and then to initiate disclosure; 26 institutions had a university-initiated, annual disclosure process (reported in Maatz [1992]). In addition, 21 schools had policies on faculty equity or managerial ties to industry that required disclosure and approval.

In what appears to be the first policy of its sort, the editor of the New England Journal of Medicine announced in 1984 that the journal would ask authors to disclose their relationships with companies that could be affected by their published findings (Relman, 1984). By 1990, the Association of American Medical Colleges had issued for its members guidelines on dealing with conflicts of interest, and AMA had provided guidance to physicians on accepting gifts from industry.

Congressional concerns about financial relationships between physicians or researchers and commercial entities have led to legislation on several occasions and also to threats of legislation. As early as 1972, the U.S. Congress prohibited companies from offering and physicians and others from accepting overt or covert payments or other rewards in return for the referral of patients or ordering of services paid for by Medicare or Medicaid. Beginning in 1989, the Congress also enacted a series of restrictions (known as the “Stark laws,” after their sponsor) on self-referral arrangements, which occur when physicians refer patients to specialty hospitals, imaging centers, or other facilities in which they have a financial interest. Also, in 1989, congressional hearings and other pressures prompted NIH to issue draft guidelines on conflict of interest for its grantees. The agency then withdrew these guidelines after criticism that they were too restrictive and would “devastate productive relationships between university researchers and industry, deny scientists outlets for their discoveries at the bench and interfere with the technology transfer” (Mazzaschi, 1990, p. 137). The U.S. Public Health Service eventually issued regulations in 1995 (see Appendix B).

In recent years, members of Congress have raised questions about industry support for continuing medical education, industry payments to physicians, and faculty member disclosure of such payments. As discussed in later chapters, members of Congress have proposed legislation that addresses some of these questions. Some proposals would require companies to report consulting and other payments to physicians, and other proposals promote alternatives to pharmaceutical company sales representatives as sources of information for physicians about medications.3 A few states have adopted policies requiring companies to disclose certain payments to physicians, and some states have created alternative education programs for physicians and other prescribers of medications.

In the 1990s, social science research techniques and findings began to influence understandings of the relationships between physicians and industry. For example, in an analysis of marketing literature and interactions between physicians and industry representatives, Roughead and colleagues (1998) noted that “[r]eciprocity is one of the norms by which society abides.… The provision of gifts by sales personnel encourages an automatic response of indebtedness on the part of the receiver who will then look for ways to make repayment” (p. 307). Other research has documented the importance of unconscious bias (see Appendix D).

Since 2000, a number of private and public groups have issued reports on conflict of interest in aspects of medical research, education, or practice. Table 1-2 lists some of the more prominent reports, several of which are discussed in later chapters of this report. Most reports have focused on research. Most have recognized the value of legitimate and properly designed research, educational, and technical relationships; but several have recommended some restrictions on other types of relationships and the more effective implementation of policies. In addition, the Pharmaceutical Research and Manufacturers of America (PhRMA) revised its voluntary Code on Interactions with Healthcare Professionals (effective January 2009) to more strongly discourage noninformational gifts, such as providing tickets to sporting events and token consulting arrangements (PhRMA, 2008). The Advanced Medical Technology Association has also revised its Code of Ethics for medical device manufacturers (effective July 2009) to include generally similar provisions (AdvaMed, 2008). (Other countries also have industry codes on relationships between the pharmaceutical industry and physicians [Jost, 2009].)

TABLE 1-2. Selected Reports on Conflict of Interest Released Since 2000.


Selected Reports on Conflict of Interest Released Since 2000.

The recommendations in the reports listed in Table 1-2 are often similar (but not entirely consistent) in calling for more accountability and openness and more effective implementation. The policies of particular institutions vary, and some individuals may be subject to multiple policies that apply to their different roles and activities. To the extent that the adoption and implementation of policy recommendations have been evaluated, the results are mixed, as discussed in Chapter 3.

Evolution of Other Strategies to Limit Bias in Medical Research, Education, and Practice

At the same time that policy makers, universities, professional groups, and others were responding to concerns about conflict of interest, methodologists, statisticians, and scientists were working to develop and refine methods for designing and conducting research and analyzing data in ways that limit bias—whatever the source—during all stages of scientific investigation, from protocol design through the reporting of the results (see Chapter 4). In addition, academic medical centers have instituted education on evidence-based medicine to instruct future physicians on how to evaluate critically the evidence presented in (or absent from) journal articles, practice guidelines, and other sources of clinical information and advice (see, e.g., Bennett et al. [1987] and EBM Working Group [1992]). Others have worked to shift methods for the development of clinical practice guidelines away from unsystematic expert opinion and consensus processes toward formal, objective procedures for identifying and reviewing the relevant evidence and linking the strength and quality of the evidence to recommendations (see Chapter 7). These techniques and strategies work together with conflict of interest policies to reduce the risk of bias and maintain public trust in medical research, education, and practice.


Chapter 2 sets forth a normative and conceptual framework for the report, including definitions and the criteria used to assess the potential benefits and harms created by financial relationships. Chapter 3 presents an overview of conflict of interest policies and what is known about their impact.

Chapters 4 through 7 are devoted to examinations of industry relationships and conflicts of interest in the domains of medical research, medical education, clinical practice, and practice guideline development, respectively. Chapter 8 discusses the importance of policies on conflicts that arise at the level of the institution. Finally, Chapter 9 discusses the role that accrediting and other supporting organizations can play in promoting the adoption and implementation of conflict of interest policies by the institutions that are on the front lines of medical education, research, and practice. Several appendixes provide additional background about the report or topics mentioned in the report.



For the purposes of this report, the committee distinguished companies that produce commercial medical products from other mostly noncommercial medical institutions (and their personnel) that these companies seek to influence. (Some providers of continuing medical education are for-profit concerns.) The committee recognized that commercial companies conduct or sponsor research and may undertake activities with educational value.


In general, this report follows the practice of recent IOM reports in referring to research participants rather than research subjects (IOM, 2001, 2003, 2004; NBAC, 2001). This report uses the latter terminology when quoting and sometimes when referring to reports that employ that terminology.


Examples of legislation that was proposed but not enacted by the 110th Congress (2007-2008) include S. 2029 (Physician Payments Sunshine Act of 2007), S. 3343 (Medicare Imaging Disclosure Sunshine Act of 2008), and H.R. 6752 (Independent Drug Education and Outreach Act of 2008). The first proposal has been revised and reintroduced in the 111th Congress (Grassley, 2009).

Copyright © 2009, National Academy of Sciences.
Bookshelf ID: NBK22932


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (2.2M)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...