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

2Principles for Identifying and Assessing Conflicts of Interest

Relationships between physicians and biomedical researchers on the one hand and pharmaceutical, medical device, and biotechnology companies on the other hand are widespread and have produced important benefits, particularly in the development of new tests and treatments. At the same time, these relationships have also created significant risks that the financial goals of industry may conflict with the professional goals of medicine. The goals of for-profit medical companies are to produce products that improve health and, at the same time, to ensure a financial return to shareholders. The primary goals of medicine include improving health by providing beneficial care to patients, conducting valid research, and offering excellent medical education. In pursuing those goals, individual professionals, health care institutions, and research organizations have obligations to put patient interests first, carry out unbiased research, critically appraise information, and serve as role models of professional behavior for students. The problem of conflict of interest arises because in some circumstances in modern medicine these goals and obligations are at risk of being compromised by the undue pursuit of financial gain or other secondary interests.

Medicine today faces many difficult challenges, including, among others, high costs of treatment and associated pressures to cut costs, lack of availability of health insurance, and persistent medical errors. In comparison, the problem of conflict of interest may seem less significant. However, none of the other challenges can be adequately met if conflicts of interest are not well managed. For example, patients and the public need to be able to trust that the high costs of health care and health insurance arise from the provision of services that are beneficial, necessary, appropriately priced, and not inappropriately driven by the financial interests of physicians, other health care providers, or medical product companies. Failure to deal with the problem of conflict of interest can undermine efforts to address the other serious challenges that medical professionals and researchers face today.

This chapter develops a conceptual framework for identifying and assessing conflicts of interests.1 In addition to defining the concept of conflict of interest and clarifying some common misunderstandings about its applications, the chapter presents principles to guide the formulation and implementation of conflict of interest policies. The principles take the form of (1) statements of the purposes of conflict of interest policies, (2) criteria for assessing the content of these policies, and (3) criteria for evaluating the implementation of policies. The principles do not directly yield decisions in particular cases or even rules that could be directly enforced, nor do they determine in advance the relative importance of all the values involved in making decisions. In applying them to particular policies and individual cases, there is no substitute for judicious practical judgment sensitive to the institutional context. However, the principles provide an essential framework for formulating and implementing any conflict of interest policy. They focus attention on the most important factors that should be considered when professionals and institutions make decisions and policies regarding conflicts of interest, select the agents who should be responsible for implementing and enforcing those policies, and choose the methods that they will use to regulate conflicts of interest.


Although conflict of interest policies are now widespread in many areas of medicine, the meaning and purposes of these policies are not always clearly understood. The term “conflict of interest” is used in many different and often inconsistent ways. Nonetheless, institutional and public policies on conflicts of interest need to define what the policies cover and what they do not cover.

The definition that the committee adopted is consistent with the core meaning of the concept as it is used in many institutional policies. It is, however, formulated to clarify key elements that are sometimes obscured in discussions of those policies.

A conflict of interest is 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.

To avoid common misunderstandings of the concept that can lead to misplaced and ultimately ineffective or counterproductive policies, the committee stresses the importance of each of the three main elements of a conflict of interest: the primary interest, the secondary interest, and the conflict itself.

The primary interest that conflict of interest policies seek to protect varies according to the purpose of a professional activity. Primary interests include promoting and protecting the integrity of research, the welfare of patients, and the quality of medical education. Physicians and medical researchers accept the primacy of these interests when they act in their professional roles. Physicians and researchers exercise judgment and discretion in their work. Patients, the public, research participants, medical students, residents, and fellows need to trust physicians and researchers to act and make judgments in ways that are consistent with these primary interests. These primary interests are sometimes stated as ends or goals (e.g., promoting patient welfare), as obligations (e.g., the physician’s obligation to promote patient welfare), or as rights (e.g., the patient’s right to have the doctor promote his or her welfare). The committee uses the term primary “interests” to encompass all of these values, however they are stated. Whatever the primary interests are, the point of regulating conflicts of interest is to try to ensure that secondary interests do not subvert physicians’ and researchers’ decisions and actions regarding those primary interests and do not undermine trust in their clinical or scientific judgment. Furthermore, medical institutions—including medical schools, research institutes, professional societies, scientific journals, patient advocacy organizations, or government health agencies—should also keep these primary interests paramount, as discussed further in Chapter 8.

To be sure, identification of the exact primary interest in specific situations may sometimes be challenging, and primary interests sometimes conflict with each other. For example, in public health emergencies or under conditions of dire resource scarcity, physicians may have fundamental obligations to the population as a whole that may compete with their obligations to individual patients. In clinical research, the welfare of the participants in a study and the study’s successful completion may be in conflict. Nonetheless, it is almost always clear that a primary interest should take precedence over a secondary interest.

The second main element of a conflict of interest is the secondary interest. Secondary interests may include not only financial gain but also the desire for professional advancement, recognition for personal achievement, and favors to friends and family or to students and colleagues. Conflict of interest policies typically and reasonably focus on financial gain and financial relationships. The reason is not that financial gains are necessarily more corrupting than the other interests but that they are relatively more objective, fungible, and quantifiable. A financial interest therefore tends to be more effectively and fairly regulated than other secondary interests. Furthermore, for-profit companies exert influence primarily through their financial relationships with physicians and researchers. They cannot bestow professional rewards such as prestigious scientific prizes that may also lead to conflicts of interest.

Most secondary interests, including financial interests, are—within limits—legitimate and even desirable goals. The secondary interests are objectionable only when they have greater weight than the primary interest in professional decision making. For example, for a researcher or a teacher, financial interests should be subordinate to presenting scientific evidence in an unbiased manner in publications and presentations.

A financial interest does not have to be great for the influence to be undue. Indeed, social science research suggests that gifts of small value may influence decisions (see Appendix D). It also suggests that influence may operate without an individual being conscious of it. When a secondary interest has inappropriate weight in a decision and distorts the pursuit of a primary interest, it is exerting undue influence.

The third key element of the definition is the conflict itself. It is not an occurrence in which primary interests are necessarily compromised but, rather, a set of circumstances or relationships that create or increase the risk that the primary interests will be neglected as a result of the pursuit of secondary interests. A conflict of interest exists whether or not a particular individual or institution is actually influenced by the secondary interest. The claim that a conflict of interest exists is based on common experience and social science research. Both experience and research indicate that under certain conditions there is a risk that professional judgment may be influenced more by secondary interests than by primary interests.

Some of these elements of a conflict of interest refer to degrees or quantities (e.g., more or less influence), but they are not directly quantifiable. What counts as undue is a matter of judgment and depends on the context. It is not a numerical probability but a judgment in a particular situation about whether a risk is undue or inappropriate. The standards for making such a judgment should be transparent and clearly specified in actual policies rather than in vague statements that professionals should avoid “undue influence.” Subsequent chapters examine what situations or relationships may be considered inappropriate in research, patient care, medical education, and practice guideline development. Appendix C offers perspectives on conflicts of interest in other professions.

Conflicts of interest should be distinguished from other closely related conflicts. Not all conflicts in medicine are conflicts between a primary and a secondary interest. A conflict of obligation arises when an individual or institution has duties that require different actions but only one of these actions can be taken in the given circumstance. Dilemmas in medical ethics often take this form, that is, the need to make hard choices between two values, neither one of which is clearly superior to the other. A common example is maintaining the confidentiality of a patient with a contagious disease, which may conflict with preventing that patient from harming someone else. There is no conflict of interest in this example because both interests have plausible claims to be considered primary. Conflicts of obligation are essentially conflicts among different primary interests. Both obligations or interests are legitimate, often equally so, and it cannot be said in advance which one should take priority.

Conflicts of commitment are closer to conflicts of interest. They often involve a conflict between what institutions view as employees’ primary responsibilities to the institution and the employees’ outside commitments, such as voluntary community service, participation in a political campaign, or teaching or conducting research for another institution. Like conflicts between primary interests, conflicts of commitment involve two perfectly respectable activities (indeed, in some cases, identical activities, except that they are conducted at different institutions). Also, like conflicts of interest, the institution can legitimately claim in advance that one activity takes priority over the other if they come into conflict in any way. The concern is not usually about the risk of undue influence over specific decisions (e.g., the prescribing of a particular medication or the reporting of research findings). Rather, the concern is about time and effort, for example, whether individuals are devoting sufficient attention to their responsibilities within their own primary institution. Conflicts of interest and conflicts of commitment are sometimes covered in the same institutional policy; but the circumstances, risks, and evaluative frameworks are sufficiently different that they warrant separate consideration. Nevertheless, it makes sense for the policies to be covered in the same documents and information resources and to be administered by the same officials and committees.


Institutions, professional organizations, and governments establish policies to address the problem of conflict of interest on behalf of the public. Conflict of interest policies are attempts to ensure that professional decisions are made on the basis of primary interests and not secondary interests. (See the discussion of the policies of other professions in Appendix C.) As discussed further in Chapter 9, such policies work best when they are preventive and corrective rather than punitive. To the extent that they are effective, they serve two overarching purposes: maintaining the integrity of professional judgment and sustaining public confidence in that judgment. That professionals should promote these purposes constitutes the fundamental principle underlying any respectable conflict of interest policy.

First, the most obvious way in which the integrity of professional judgment can be compromised is through bias. Other practices can also undermine that integrity when they violate standards of professional conduct, such as the failure to publish research findings in a timely manner, the failure to treat students and postdoctoral fellows fairly, and a lack of openness with patients. Conflict of interest policies seek to minimize the influence of secondary interests in all these practices. They most significantly guard against the risk that financial interests will have excessive weight in decisions about the conduct of research, teaching, the provision of patient care, and the development of practice guidelines.

Such policies do not assume that any particular professional will necessarily let financial gain influence his or her judgment, nor do they imply that the individual researcher or physician is an unethical person. They assume only that under some conditions a risk exists that the decisions may be unduly influenced by considerations that should be irrelevant. Nonetheless, physicians and researchers are sometimes offended by assertions that they have conflicts of interest, believing that such assertions impugn their ethical integrity.

To avoid what they believe to be the negative connotations of “conflict of interest,” some institutions use such phrases as “relationships with industry” or “financial relationships” to describe not only relationships that may be evaluated for the presence of potential conflicts but also relationships that are judged to be conflicts of interest. This less direct language has the effect of obscuring the serious risks that conflicts pose. Such language is not necessary if it is recognized that the determination that an individual or institution has a conflict of interest is a judgment about the situation and not about the professional who happens to be in that situation.

The second purpose of conflict of interest policies—to help sustain public confidence in professional judgment—is less appreciated but no less important. Here the goal is to minimize conditions that would cause reasonable individuals to suspect that professional judgment has been improperly influenced by secondary interests, whether or not it has. The public includes not only patients and research subjects but also editorial writers and journalists, officials at nonprofit foundations, public officials, and other opinion leaders. When or if the public and public officials distrust physicians, researchers, or educators, they are likely to seek greater government regulation, withhold funding, and take other steps that could jeopardize future programs of patient care, research, or education.

When a physician, researcher, or educator acts in ways that lead to distrust, the consequences may affect colleagues, patients, students, and the institution or profession as a whole. Similarly, institutional practices can be the source of distrust, and the effects of such distrust may be even more widespread and damaging than distrust of an individual. Physicians retain a high standing with the public compared with the standing of many other professional groups; but physicians should be vigilant, because once public confidence is undermined, it may be difficult to restore.

As discussed in Appendix C, other professions—law, accounting, engineering, and architecture—have also recognized the importance of conflict of interest policies and ethical codes to promote objectivity in decision making and sustain public confidence. In some recent cases, most notably, accounting, failure to adhere to these codes has led to increased government regulation.


Individuals accused of having a conflict of interest often say that they would never let financial interests influence their decisions. This objection to conflict of interest policies misses the point. Because (as noted above) the conflict is a set of circumstances or conditions involving a risk rather than a specific individual decision, the existence of a conflict of interest does not imply that any individual is improperly motivated. Nevertheless, an individual professional might still object that it is not fair to generalize in this way. He or she may want to say: “Look at my actual decisions and consider my distinguished reputation.” However, conflict of interest policies are by their nature designed to avoid the need to investigate individual cases in this way. For at least two reasons, such policies do not focus on the motives in a particular case.

First, reliably ascertaining or inferring motives in this context is usually impossible for those assessing whether a relationship constitutes a conflict of interest. Generally, medical research, patient care, and education involve multiple considerations and many small judgments and decisions that are impractical to review; and even if they were reviewed, they would likely not yield a clear picture of the underlying motivation. Thus, readers of journal articles, medical students, patients, and conflict of interest committees are not in a good position to judge whether secondary interests motivated a decision. The motives behind institutional decisions are usually even more opaque.

Second, any thorough effort to determine motivation in a particular case would be improperly intrusive and highly time-consuming. Fair hearings could not be held and reliable conclusions could not be reached without risking violation of the rights of privacy of the many individuals who might be involved and without distracting many people from other important work.

Sometimes another, closely related objection to a claim that an individual has a conflict of interest is raised. This objection accepts that motives should not be considered but denies the relevance of the conditions under which decisions are made: “Judge my decision—the results of the research, the content of the lecture, the prescription of the drug—and not my financial interests.” Here again the problem is that many people affected by professional decisions are not in a position to judge the validity of those decisions. In addition, those who are competent to judge may not be able to do so until after the damage has occurred. Furthermore, the argument for judging outcomes ignores one of the two main purposes of conflict of interest policies: the maintenance of public confidence. Even valid decisions and research may not be widely accepted as such if they occur under conditions in which secondary interests are prominent. Moreover, many decisions in research and clinical care are close calls. Plausible reasons can be cited for each of several alternative choices. The decisive factor in whether a judgment or an action is accepted as valid may turn on whether a researcher or a clinician can be trusted to be acting for the sake of scientific truth and the best interests of patients.

Because it is both intrusive and usually impracticable to investigate motives and because the competent and timely appraisal of decisions is often difficult, it may be tempting to conclude that patients, the public, and researchers simply need to trust physicians. Trust is important, but generalized trust and reliance that medical professionals act in accord with primary professional interests may be difficult to maintain in the face of evidence that this trust is sometimes abused. Furthermore, creating trust in medical professionals who conduct research or develop practice guidelines is hard because they have little or no contact with many of the people who are affected by their decisions and who have only limited knowledge with which to evaluate the decisions. Trust is necessary and desirable, but it must be based on reasonable expectations. Those who rely on professionals must have good reason to trust their decisions. In short, they need assurance that the professionals are trustworthy. Policies designed to reduce conflicts of interest and mitigate their impact provide an important foundation for public confidence in medical professionals and institutions.


Some conflict of interest policies refer to actual or perceived conflicts of interest and state that professionals should avoid “even the appearance of a conflict of interest.” That requirement may lead to confusion. All conflicts of interest involve perceptions or appearances because they are specified from the perspective of people who do not have sufficient information with which to assess the actual motives of a decision maker and the effects of those motives on the decisions themselves.

Policies that contrast actual and perceived conflicts of interest give rise to two problems. First, the contrast suggests that there is no conflict (only an appearance of a conflict) unless the decision maker actually favors secondary interests over primary interests. The implication, then, is that conflict of interest policies should treat a perceived conflict as less serious than an actual conflict. However, when a professional’s judgment is actually distorted by the acceptance of a gift or the prospect of influencing a stock in which the professional has an interest, the violation is no longer principally a conflict of interest but becomes a different kind of offense, one that may involve malpractice, scientific misconduct, or kickbacks. Those violations call for the use of procedures quite different from those on which conflict of interest policies should concentrate.

Second, the creation of a category of perceived conflicts, as distinct from actual conflicts, opens the door to overly broad and excessively subjective rules. If perceived conflicts are treated as different from the other (so-called actual) conflicts that the policy regulates, conduct that is perfectly proper can be unfairly called into question. With a loose notion of the perception or the appearance of a conflict of interest, circumstances that are suspicious only to uninformed people or predisposed reporters can be the basis of indiscriminate charges of conflicts of interest. Charges of conflicts of interest should be limited to circumstances specified by policies that are objectively grounded in past experience and reasonably interpreted on the basis of relevant and accessible information.


Conflicts are not binary; that is, they are not simply either present or absent. They can be more or less severe. The severity of a conflict depends on (1) the likelihood that professional decisions made under the relevant circumstances would be unduly influenced by a secondary interest and (2) the seriousness of the harm or wrong that could result from such influence. As discussed later in this chapter, the criterion of proportionality in conflict of interest policies provides that the expected benefits of a relationship may be considered, and conflicts of interest may be allowed to continue if those benefits outweigh the risks and safeguards that are instituted.

Table 2-1 lists the questions that need to be asked when the severity of a conflict of interest is assessed in particular cases. These questions express criteria or principles that identify the most important factors to be considered in formulating policies and making decisions about conflicts of interest. Assessments of the likelihood of undue influence and the seriousness of the consequences usually reflect general judgments about situations—on the basis of experience—rather than evaluations of the character of the individual in question. The individual’s behavior in similar situations in the past might, however, be taken into account. The next two sections discuss the criteria in more depth.

TABLE 2-1. Criteria for Assessing the Severity of Conflicts of Interest.


Criteria for Assessing the Severity of Conflicts of Interest.

Assessing the Likelihood of Undue Influence

In assessing the likelihood of undue influence, it is reasonable to assume that the greater that the value of the secondary interest is (e.g., the greater that the size of the financial gain is), the more probable is its influence. Thus, equity or other ownership interests in a small biotechnology company have great potential for an increase in value on the basis of the results of a clinical trial (as well as the potential for no value). Large fees for serving on a company advisory board are more valuable than occasional small honoraria for talks. Although absolute value is important, the secondary interest should generally be measured in relation to the typical income for the relevant class of professionals or in relation to the value of a research project, institutional budget, or medical practice.

However, the monetary value of a secondary interest is not the only appropriate measure of its potential impact. The economic value of pens, inexpensive meals, and other nominal gifts or relationships is low; but as explained in Appendix D, small gifts may help to create and sustain relationships, for example, between a physician and a pharmaceutical company and its representatives. The influence of such gifts and relationships may be subtle and the individual receiving such gifts may not even be conscious of their influence. It may therefore be necessary to manage or prohibit conflicts of interest even when the value of the secondary interest, as measured only by monetary value, is low and the likelihood of harm or wrong in a single instance is low.

Other aspects of relationships besides their dollar value may also increase their general value and therefore the risk of a conflict of interest. For example, payments that augment the income of an individual professional may create more concern than those that exclusively support the academic activities of a whole institution. A consulting arrangement that increases a researcher’s income will tend to create more concern than one in which payments are made to the institution, department, or research group as a whole and disbursed under institutional oversight. Similarly, a research contract that is reviewed by a university for consistency with policies on data access, sponsor review, and publication rights will generally create less risk of a conflict of interest than a consulting arrangement that does not receive such review or that is reported only in very general terms (e.g., as involving payments over $20,000 when the actual amount is $200,000).

A second factor affecting the likelihood of undue influence is the scope of a relationship, which refers to its duration and depth. Longer and closer associations increase the scope and therefore the risk. Examples of such associations include a multiyear consulting agreement, a continuing leadership position as a member of a company board, or the weekly or monthly provision of free lunches at a physician’s office. Likewise, long-term funding of a university or commercial continuing medical education program has more potential for influence than a one-time grant. Similarly, serving on a company’s scientific advisory board, which more intimately ties the professional to the company over time, is more likely to affect the professional’s judgment than accepting a fee for speaking about a company-sponsored research project.

The extent of discretion, that is, how much latitude a professional enjoys in making important decisions, is also pertinent. Even though some of their judgments are subject to various kinds of review, the principal investigator in a clinical trial exercises considerable discretion over innumerable, wide-ranging, and often hidden decisions, for example, decisions regarding the eligibility of patients to enter the clinical trial, determinations of clinical end points, ascribing of adverse events to the study intervention, the type of statistical analyses to be used, and the reporting of the results. This discretion is often limited by an independent oversight body, for example, a data and safety monitoring board, an independent panel that adjudicates adverse events, a medical monitor of adverse events, or an external auditor for data collection at individual research sites. Such oversight is usually required for any clinical trials whose results will be presented to the Food and Drug Administration for regulatory approval of a drug or medical device. In assessing such limits on discretion, it is also important to consider the independence of the judgment of the members of any oversight body. Furthermore, the more closely that the research and data analysis methods follow standard methods, the less room there is for improper influence. Similarly, the more conventional the subject matter of educational presentations, the less scope there is for bias that is not easily detected.

Authority and discretion often vary by role. Principal investigators can influence multiple dimensions of a research project, whereas laboratory technicians or research assistants have less scope for influence in most situations. Deans and chancellors, through their power to control appointments, promotions, salaries, and space, wield great power, although they are typically several steps removed from conducting research projects or teaching courses. At the other extreme, most administrative staff members have little power to influence a university’s research or teaching mission.

Assessing the Seriousness of Possible Harm

The starting point in assessing the seriousness of possible harm from a conflict of interest is to identify or specify the value of the primary interest. This report concentrates on the primary goals of medicine, particularly patient care, research, and medical education. Assessing the severity of a conflict requires an examination of the specific primary goal or goals at risk in a particular situation.

A second consideration is the scope of the consequences. The greater that the scope is, the more serious is the potential for harm. Conflicts of interest that may affect multiple patient care decisions have a large scope. For example, practice guidelines that set standards of care and criteria for insurance coverage may affect millions of patients. The results of a clinical trial for a common condition can affect how thousands of physicians prescribe a specific medication. Results from a pivotal trial of a novel type of therapy that may dramatically alter patient care are likely to have a larger scope than other trials that will influence care only at the margins. Thus, conflicts of interest in clinical trials deserve special attention because of the potentially large scope of their effects.

A conflict of interest may also have negative effects on an individual’s colleagues or institution. Such effects need to be taken into account even if they do not occur frequently. A pharmaceutical or medical device company’s sponsorship of a research project could raise questions about the work of other researchers in the institution and weaken their ability to raise funds from other sources. A professor’s close connections with a company not only could raise doubts about the objectivity of his class materials and presentations, but these connections could also have negative effects on the careers of his teaching assistants and the collegial culture of the institution. In view of such possible consequences, the fact that an individual has a right to engage in an activity should not be allowed to obscure the equally important fact that his or her actions may affect the rights of colleagues and students. The claim of an individual right by a professional does not preclude the possibility that this right may be regulated.

Finally, the seriousness of the possible harm depends in part on the extent of accountability. In general, a conflict of interest is more serious when the level of accountability of the physician, researcher, or educator to his or her peers, institution, licensing board, or similar entity is less extensive. If accountability for decisions is bolstered by an independent review of those decisions by colleagues or other authorities, there is generally less potential for harm and less cause for concern. However, the reviewers must be and must be viewed as being effective and independent and must have no conflicts of interest of their own. Accountability is also greater to the extent that sanctions for serious violations of policies are significant and imposed in a timely fashion, and it is further enhanced if the results of the disciplinary proceedings are regularly disclosed.

In summary, an overall assessment of whether a financial relationship constitutes a conflict of interest and, if so, how severe it is and how it should be managed depends on several considerations: the importance of the financial or other relationship for furthering primary medical values, the likelihood and seriousness of possible harm to those primary values, and the availability of measures that can reduce the likelihood or severity of harm. Chapter 3 discusses such measures and also the procedures applied by universities and other institutions to identify, limit, and manage conflicts of interest.


The discussion above focused on several questions and factors that should be considered in assessing the severity of a conflict of interest in financial relationships. They are intended to provide guidance for the formulation of the content of policies for controlling conflict of interest, for example, the specification of the information needed from individuals that will be sufficient to evaluate financial relationships, assess the severity of conflicts of interest, and guide responses to identified conflicts. Additional criteria are needed to evaluate the implementation or actual operation of the policies (Table 2-2). Even if policies are well formulated, they must also be well administered.

TABLE 2-2. Criteria for Evaluating Conflict of Interest Policies.


Criteria for Evaluating Conflict of Interest Policies.


First, the criterion of proportionality calls for policies to be efficient and effective in addressing serious conflicts of interest in a preventive or a corrective way. Complicated rules and elaborate procedures can become merely bureaucratic obstacles unless they are implemented and regularly reviewed with the goals of the policy in mind. Do the policies actually address the most important and common conflicts? Are the policies practical; that is, can they actually be effectively implemented at an acceptable cost? Are the policies administered in a way that appropriately considers the likelihood of bias, the seriousness of the harm, and the potential benefits of the conflicting secondary interest (as noted above)? Do the policies and their application unnecessarily interfere with the conduct of legitimate research, teaching, and clinical practice? Do the anticipated benefits of the policies outweigh their various costs, such as administrative burdens, and any negative consequences? The effectiveness of a specific policy can be judged only after that policy has been in use for a period of time. Insofar as experience and evaluations have raised questions about the effectiveness of similar policies already adopted, however, these questions can guide the design and implementation of new policies. Finally, whether policies can achieve their overall aims will also depend on their congruence with other criteria, such as fairness and transparency, that contribute to effectiveness and that are also important for their own sake.

The criterion of proportionality should also be applied in individual situations when an assessment is made of whether a financial relationship constitutes a conflict of interest and, if there is a conflict, how it should be handled. For example, when a researcher’s financial relationship with a company is evaluated, its expected benefits as well as its risks should be considered. As discussed in Chapters 1 and 4, industry support for well-designed and scientifically meritorious research tends to advance a primary goal of generating valid scientific knowledge. This may sometimes mean that an institution should allow an individual with a conflict of interest to participate in an activity because the expected benefits exceed the risks and because the risks have been lowered to an acceptable level. For example, an academic medical center may allow a scientist who holds the patent on a promising discovery to participate in developing a product and designing an early-stage clinical trial to evaluate an intervention because his or her involvement may be necessary to ensure that the product is safely and correctly administered. (These kinds of situations, which should be exceptional, are discussed further in Chapters 4 and 7.)


Just as disclosure is usually necessary—even if it is insufficient—in dealing with conflicts of interest, so too is transparency necessary in administering conflict of interest policies. Transparent policies are readily available in clear and simple language, together with explanations and essential information about their application. They are also available not only to those who are subject to them (e.g., researchers, authors of journal articles, or members of practice guidelines panels) but also to other stakeholders, including the public. Transparency is essential to determine whether conflict of interest policies are reasonable and if they are being implemented fairly.

Conflict of interest policies may require the public disclosure of financial and some other relationships under certain circumstances, as described in Chapter 3. These disclosure policies reflect the institutions’ ethical and sometimes legal responsibility to disseminate relevant information to appropriate parties. In addition, the values of transparency are also served when institutions explain their judgments in certain cases, for example, when they allow an investigator with a financial stake in the outcome of a study with human participants to conduct that research (see Chapter 4).

Rights of privacy and protection of confidentiality place some limits on how much information an institution discloses and to whom. For example, physicians have a countervailing privacy interest when it is proposed that their financial relationships (and perhaps those of their family members) be disclosed to the public, as noted in the discussion in Appendix F of public disclosure of personal information reported to academic medical centers and other institutions. Disclosures beyond the institution can be limited to the minimum amount of identifiable personal information that is needed to carry out policy goals. For some purposes, reporting aggregate or deidentified information to the public is sufficient.

Transparency can also help improve conflict of interest policies across institutions. Information about the way that one institution has handled a particular case or type of case can enable other institutions to learn about more and less effective practices and adjust their own policies and behaviors accordingly.


Accountable individuals and institutions explain and take responsibility for their conduct and decisions. Thus, just as a physician explains the rationale for clinical decisions to patients and researchers explain the rationale for research and research procedures, so too will leaders of accountable institutions explain their policies and their application to the individuals who are directly affected and respond to questions and suggestions.

Taking responsibility for the consequences of individual or institutional actions and decisions may involve offering apologies or compensation to those harmed by these actions and acknowledging the appropriateness of penalties when a representative of the institution has acted improperly or illegally. To demonstrate that it is accountable, an institution not only will develop explicit conflict of interest policies and procedures for implementing its policies but also will devise ways to communicate how they are applied in practice. Institutional leaders will be prepared to explain how judgments about conflicts of interest are reasonably consistent across similar cases and why, for example, they determined that it was sufficient to require only the disclosure of a relationship in one case but appropriate to manage or prohibit the relationship in another case. Finally, institutional leaders will be ready to respond to questions about their own interests and impartiality. As discussed in Chapter 8, leaders should establish procedures for dealing with the conflicts that their own institutions may have.

Public engagement is often important for accountability. For example, accountability is generally enhanced if public representatives serve on institutional panels that review individual relationships that may present conflicts of interest. To cite a somewhat parallel situation, federal regulations require institutional review boards to include at least one member not affiliated with the institution. Also, as part of a commitment to openness and accountability, organizations may invite public comment on their conflict of interest policies and may take seriously suggestions for revisions. Public participation can enhance the credibility and trustworthiness of decisions about individual cases as well as more general policies.

A final aspect of accountability is a commitment to improving conflict of interest policies and their implementation. Setting benchmarks for performance and tracking outcomes can stimulate quality improvement activities, as has been demonstrated with other activities in health care organizations.


The formal principle of fairness requires similar treatment for those in relevantly similar situations and different treatment for those in relevantly different situations. This principle has at least two implications for the application of conflict of interest policies.

First, these policies should apply to all employees or members of an institution who make significant decisions for the institution or who have substantial influence over these decisions. In academic medical centers, these decisions may involve medical education, medical research, or clinical care. Thus, residents, fellows, faculty, members of institutional committees (e.g., institutional review boards, formulary committees, and device purchasing committees), and senior institutional officials are all subject to conflict of interest policies and procedures. Although medical students do not usually have an influence over decisions that are made, they too should be expected to follow conflict of interest rules, which are among the important professional norms they are learning as they prepare for their future careers. At the same time, to be fair, conflict of interest policies and procedures may reasonably differ for people in different roles. For a medical student or resident, the policy issue might be accepting mugs, pens, and lunches from companies. For a senior leader in the institution, the issue might be serving on the board of directors of a company manufacturing medical products and receiving personal compensation for this position. In some cases, the policy response might be to prohibit a practice overall, whereas in other instances management of the conflict could be an option, depending on the specifics of the situation, as assessed by the standards listed in Table 2-1.

Second, fairness requires that individuals in different institutions who are in situations that are similar in all ethically relevant ways be treated similarly. Otherwise, the ethical basis for policies may be called into question and conflict of interest policies and decisions may be regarded as arbitrary. Although conflict of interest regulations for U.S. Public Health Service grantees and policies recommended by the Association of American Medical Colleges allow institutions discretion in setting and implementing policies to take account of local circumstances, it is important to justify such variation in ways that are understandable by and plausible to affected individuals, oversight agencies, and the public.


The purposes of conflict of interest policies are expressed in the principles that hold that professionals should act to protect the primary interests of medical practice, education, and research and to maintain public confidence in the integrity of those activities. The problem of conflict of interest is more complex than is often appreciated. As a result, both critics and defenders of conflict of interest policies sometimes misunderstand or misapply them.

A conflict of interest is not an actual occurrence of bias or a corrupt decision but, rather, a set of circumstances that past experience and other evidence have shown poses a risk that primary interests may be compromised by secondary interests. The existence of a conflict of interest does not imply that any individual is improperly motivated. To avoid these and similar mistakes and to provide guidance for formulating and applying such policies, a framework for analyzing conflicts of interest is desirable.

This chapter has presented principles for assessing conflicts of interest and evaluating policies designed to deal with such conflicts. Conflicts should be assessed by considering various factors that determine their likelihood and seriousness. Likelihood depends on the value of the secondary interest, the scope of the relationship between the professionals and the commercial interests, and the extent of discretion that the professionals have. Seriousness depends on the value of the primary interest, the scope of the consequences that affect it, and the extent of accountability of the professionals. Conflict of interest policies should be evaluated by considering their effectiveness, transparency, accountability, and fairness.

A better understanding of the nature of conflicts of interest and the clearer and fairer formulation of rules can support greater confidence in the medical profession and thereby enable physicians, educators, and investigators to concentrate on their primary missions of treating patients, teaching students, and conducting research. With robust conflict of interest policies in place, they can continue to carry out their respective activities not in wary confrontation but in beneficial cooperation with the representatives of industry.



The discussion in this chapter draws on work by Thompson (1993) and Emanuel and Thompson (2008). The committee also consulted other definitions and frameworks, including those of Davis (1998), AAMC (2001), Davis and Stark (2001), NIH (2004), Moore et al. (2005), Lurie (2007), Sage (2007), AAMC-AAU (2008), and Beauchamp and Childress (2009).

Copyright © 2009, National Academy of Sciences.
Bookshelf ID: NBK22937


  • 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...