Deconstructing “Conflicts of Interest”: A User’s Guide

November 24, 2009

The debate within the medical profession over “conflicts of interest” (COIs) has often been shrill, and sometimes seems to be based on misunderstandings or myths about what COIs entail. In this psychiatrist’s view, it is helpful to step back from confident proclamations, acknowledge that the issues involved are complex, and aspire to some semblance of humility. Nobody has cornered the market on “the right way” to deal with COI in the realms of medical research, publication, and education.1 At the same time, as Alan Stone, MD, has noted (personal communication, August 27, 2009), ethical considerations lie at the heart of any debate on COI-in particular, the ancient dictum, “Do no harm.” Indeed, ethicist James M. DuBois has pointed out a direct connection between some types of COI and harm to the general public: “Mental health consumers are at risk when studies that involve questionable scientific and publication practices are translated into therapeutic practice.”1(p205)

The debate within the medical profession over “conflicts of interest” (COIs) has often been shrill, and sometimes seems to be based on misunderstandings or myths about what COIs entail. In this psychiatrist’s view, it is helpful to step back from confident proclamations, acknowledge that the issues involved are complex, and aspire to some semblance of humility. Nobody has cornered the market on “the right way” to deal with COI in the realms of medical research, publication, and education.1 At the same time, as Alan Stone, MD, has noted (personal communication, August 27, 2009), ethical considerations lie at the heart of any debate on COI-in particular, the ancient dictum, “Do no harm.” Indeed, ethicist James M. DuBois has pointed out a direct connection between some types of COI and harm to the general public: “Mental health consumers are at risk when studies that involve questionable scientific and publication practices are translated into therapeutic practice.”1(p205)

What follows are simply the views of this educator, editor, and sometime-researcher-views strongly held but not intended as “proclamations.” With that prologue, here is my “User’s Guide” to COI, written in the form of questions and answers.

In the health care field, what is the professionally accepted definition of COI?

There is no single, universally accepted definition of COI, although there is substantial convergence around a few general definitions of the term. Therefore, when someone is alleged to have a COI, the first order of business is to ask the person making the allegation to define COI.

But aren’t professional journals and organizations providing reasonably clear definitions of COI?

Yes, but the definitions differ in important ways and are sometimes difficult to interpret. Most definitions of COI-both outside and within the medical profession-follow 1 of 3 underlying paradigms-which we might call the “3 Ps”: perception, potential, and probability. These are founded, respectively, on judgments regarding how observers perceive the situation in question; on whether the situation has anypotential for conflict of interest; and on whether the situation is more likely than not to lead to such conflict. For example, one definition of COI from the business world emphasizes perception: “We can define a COI as a situation in which a person has a private or personal interest sufficient to appear to influence the objective exercise of his or her official duties as, say, a public official, an employee, or a professional [italics added].”2

He who proclaims will proclaim flat. -Lionel Ziprin

Note that this definition does not require any actual influence on the person’s objective exercise of duties. Neither does it require either the potential or the probability of an actual COI arising-rather, it falls under the rubric of “having the appearance of impropriety.” Of course, one might suspect that if there is such an appearance, there must also be a reasonable likelihood of COI (“Where there’s smoke, there’s fire!”). However, this line of reasoning fallaciously assumes that all perceptions of COI are necessarily accurate or objective, and are not themselves influenced by all manner of malign motives and biases.

Another widely cited definition of COI from Columbia University emphasizes the potential of some situation to compromise one’s objectivity: “The simplest working definition [of COI] states: A COI is a situation in which financial or other personal considerations have the potential to compromise or bias professional judgment and objectivity.”3

The Columbia definition of COI, like that of MacDonald and colleagues,2 does not require that any decision actually be biased, or even that such bias be likely; on the contrary, the Columbia doctrine is clear that “a COI exists whether or not decisions are affected by a personal interest; a COI implies only the potential for bias, not a likelihood [italics added].”3

Incorporating the word “potential” into the core definition of COI appears to create logical and syntactic problems for the Columbia position when it comes to defining “potential conflict of interest.”3 The Columbia position also seems internally contradictory, as when it states: “There are many varieties of conflicts of interest, and they appear in different settings and across all disciplines. While conflicts of interest apply to a “wide range of behaviors and circumstances,” they all involve the use of a person’s authority for personal and/or financial gain.”3

The word “use” implies that the situation has gone beyond merely a potential action, and has veered over into an actual abuse of power or authority. Note that in the Columbia definition, “a COI is not considered misconduct in research, since the definition for misconduct is currently limited to fabrication, falsification, and plagiarism [italics added].”3

Probability- or tendency-based definitions of COI imply that in a certain set of conditions, an individual’s judgment, more likely than not, will be influenced by some secondary interest. (“Potential” COI might entail a 2%, 30%, or 60% probability; “tendency” or “likelihood,” as I use the terms, implies a greater than 50% probability.) This is exemplified in the definition offered by Harvard ethicist Dennis Thompson,4 for whom a COI is “a set of conditions in which professional judgment concerning a primary interest (such as a patient’s welfare or the validity of research) tends to be unduly influenced by a secondary interest (such as financial gain) [italics added].”

Note that this definition avoids the term “potential” altogether. Rather, it emphasizes the likelihood or probability of a bias developing (“tends to be unduly influenced”). Thompson also includes a “quantitative specifier” for COI; that is, there must be an undue degree of influence. This is conceptually very important, because it implies that a COI may not be present if the degree of influence remains modest or moderate (ie, within some unspecified parameters of acceptability). In this view, not every situation in which bias might occur-or be perceived by others as existing-is a full-blown conflict of interest. Indeed, in theory, Thompson’s definition even allows for the possibility that in a situation in which someone stood to gain financially by compromising objectivity but was literally not “interested” in financial gain, a full-fledged conflict of interest might not exist.

One important caveat, however, applies to Thompson’s definition. As philosopher Arthur Schafer has observed, all “interests” are not created equal in terms of the physician’s moral and fiduciary obligations. Thus, as ethicist Dr Howard Brody5(p34) (citing Schafer) tartly observes: “The physician has a professional duty to care for the patient at a level of reasonable competence and fiduciary commitment. By contrast, the physician has no duty to make more money.”

Is a COI an “either/or” thing? That is, is it a binary or categorical term, like “on/off,” such that either one does or one does not have a COI? Or are there “degrees” of COI?

Thompson’s definition implies the latter; that is, that there may be “minor” or “low-level” conflicts that do not rise to the level of full-blown COIs. Indeed, this is also implied in the definition of COI provided by the International Committee of Medical Journal Editors (ICMJE) in their Uniform Requirements for Manuscripts Submitted to Biomedical Journals: “Conflict of interest exists when an author (or the author’s institution), reviewer, or editor has financial or personal relationships that inappropriately influence (bias) his or her actions (such relationships are also known as dual commitments, competing interests, or competing loyalties). These relationships vary from negligible to great potential for influencing judgment. Not all [financial or personal] relationships represent true conflict of interest [italics added].”6

That said, the ICMJE appropriately cautions that the reviewer or editor is not necessarily the best judge of whether a COI is present and, indeed, may not even be conscious of an actual or incipient COI: “the potential for conflict of interest can exist regardless of whether an individual believes that the relationship affects his or her scientific judgment. . . . [Therefore] all participants in the peer-review and publication process must disclose all relationships that could be viewed as potential conflictsof interest. Disclosure of such relationships is also important in connection with editorials and review articles, because it can be more difficult to detect bias in these types of publications than in reports of original research [italics added].”6

Thus, the ICMJE guidelines would require disclosure of putative conflicts that might not rise to the level of COIs, but which might be perceived as potential conflicts of interest. This highlights the difference between a definition of COI that invokes “inappropriate influence”; and guidelines for disclosure, which may require a lower threshold.

What are the main types of COI?

It is important to distinguish 2 types of COI: financial and non-financial. In theory-and assuming honest financial self-disclosure by authors and researchers-financial COI is “detectable” (eg, by evaluating an individual’s declared stock holdings). Non-financial COI, however, may be nearly impossible to detect if it involves, say, an editor’s intense dislike of a particular author-or a peer reviewer’s wish to thwart a competitor’s academic goals by summarily rejecting the researcher’s submission. The only real protection against this type of “stealth” COI is to submit papers to a double-blind review (neither author nor reviewer knows the identity of the other) ideally by several reviewers. Alternatively, some journals have an “open review” policy, in which reviewers actually sign their reviews. This approach, used in some BioMed Central journals, has pros and cons. On the one hand, it may reduce the likelihood of anonymous “slash and burn” reviews; on the other hand, it may constrain the reviewer’s candid assessment of a manuscript.

Some authorities also distinguish “inherent” from “induced” conflicts of interest.1 For example, some might argue that the medical research enterprise inherently involves a tension between a “selfless” dedication to benefiting mankind, and a not-so-selfless interest in advancing one’s career (see below, re: “vested interest”). Other COIs are “induced”; for example, when a researcher knowingly creates situations of financial dependency that may compromise judgment.

Is a “vested interest” the same thing as a “conflict of interest”?

No. A particular individual’s motivation virtually always contains an element of self-interest, or a vested interest in promoting the well-being of one’s family, friends, employer, religious or civic group, etc. (I am not using “vested interest” in the legal sense but rather to denote a special interest in protecting or promoting that which is to one’s personal advantage). For example, an editor may be highly motivated to present his or her journal in a positive light (it would be a peculiar editor who would seek to do otherwise!). In so doing, the editor’s “bias” may indeed work to the journal’s advantage (eg, by attracting readers, advertisers). But by itself, this does not necessarily entail a conflict of interest. That is, there is not necessarily a conflict between one’s primary professional and fiduciary obligations and one’s secondary interests-such as securing an income, promoting one’s career, and so forth. The crucial question is whether one fulfils one’s professional and fiduciary obligations ethically and responsibly. Indeed, it may be in the interest of all parties concerned (editors, writers, readers, the scientific community) to see a journal, professional organization, or company flourish.

In short, a vested interest is not necessarily a conflictof interest, although in some cases it may become one.

Does having a COI inherently mean one is behaving unethically?

No. If we accept the concept that a COI represents a set of conditions, it is not possible to ascribe moral qualities to a set of conditions. A COI is not an action or behavior; therefore, it cannot be “ethical,” “unethical,” or of any moral character at all. Indeed, in their widely cited study of industry-sponsored research in psychiatry, Perlis and colleagues7 state, “Conflict of interest need not imply unethical behavior or wrongdoing.” Similarly, DuBois1(p204) notes that COI “does not in itself imply wrongdoing.”

On the other hand, one might deliberately engage in behaviors that create a conflict of interest (see above-mentioned “induced” COI). For example, one might set oneself up as the direct financial beneficiary of some company [condition “A”], then subsequently agree, in some other setting, to render ostensibly impartial verdicts on the quality of that company’s goods or services [condition “B”]. Knowingly creating such a COI would be unethical in itself; failure to disclose condition A to those involved in condition B would greatly compound the ethical lapse. Indeed, as Carlat has argued, “choosing to participate in a potentially damaging incentive structure is, by itself, morally wrong in a way that is even “worse” than simply calling it a conflict of interest” (D. J. Carlat, MD, personal communication, August 30, 2009).

In short, what is ethical or unethical is whether or not one knowingly creates a COI, whether or not one discloses the conflict to the appropriate parties, and how one manages the conflict.

Example: A psychiatrist, Dr Smith, agrees to join a speakers’ bureau for a large pharmaceutical company [Company A] in exchange for a regular monthly stipend. She then accepts a position at a community mental health center, where Dr Smith is named chair of the Pharmacy Committee. It is the committee’s job to decide which medications to include on the center’s formulary; “candidate” medications include several drugs produced by Company A. Dr Smith does not inform her clinical supervisor at the health center of her affiliation with Company A.

Q: Does a conflict of interest exist for Dr Smith?

A: Yes, clearly.

Q: Did Dr Smith act unethically in accepting the Pharmacy Committee position?

A: That depends on whether she knew, or should have known, that drugs produced by Company A were likely to be “candidate” drugs for the formulary. If she had such knowledge but accepted the position anyway, she acted unethically to the extent that she knowingly created a COI for herself. If Dr Smith did not know and could not reasonably have foreseen that the committee would be considering drugs made by Company A, she may have acted imprudently or naively, but not necessarily unethically, in accepting the position.

Q: Having accepted the committee chair, did Dr Smith act unethically by failing to inform her clinical supervisor, and the committee, of her affiliation with Company A?

A: At the very least, Dr Smith did not fulfill her professional and ethical obligations (ie, her affiliation certainly should have been disclosed). Even if Dr Smith was unsure of what drugs might come before the Pharmacy Committee, her affiliation with Company A should have been disclosed to her supervisor and the committee. Had Dr Smith disclosed the COI to her supervisor, they might have worked out an agreement whereby Dr Smith would recuse herself from any committee decisions regarding drugs produced by Company A.

Does having a COI mean that a doctor or researcher has made a faulty decision, produced inaccurate results, or rendered a judgment that is biased or unfair?

No, not necessarily. Again, a COI is a set of conditions, not a behavioral outcome or “verdict” on the quality of one’s work. To be sure, those conditions that create a COI do, by definition, predispose one toward bias, inaccuracy, or unfair judgments, but none of these outcomes necessarily follows from the COI per se.

Example: Dr Jones is a medical researcher who has written a review paper on the efficacy of the drug “nullotropin,” and the writing has been funded by an “unrestricted writing grant” from the pharmaceutical company that produces nullotropin. Dr Jones carries out a careful review of all the placebo-controlled randomized studies of nullotropin-including several unpublished studies that found no significant effect or benefit from the drug. During the writing of the paper, Dr Jones has no contact with the underwriting company, and is subject to no “monitoring” or “feedback” from the company. His paper ultimately concludes that “nullotropin appears to be moderately effective for its indicated use, notwithstanding 2 unpublished, negative studies.” The drug company does not require “pre-authorization” or “right of refusal” with respect to Dr Jones’s work, and the paper is submitted to and accepted by a peer-reviewed journal that uses “blinded” reviews. The financial support for Dr Jones’ study is disclosed in the appropriate section of the paper.

Q: Did a COI exist with respect to Dr Jones writing and submitting the paper? If so, was it a major COI?

A: Arguably, a COI did exist, using Thompson’s definition, if Dr Jones believed (correctly or not) that his writing grant might be withheld by the company-or that the company would never again support his writing-if the outcome of his paper was unfavorable to nullotropin. If that was the case, then there existed “a set of conditions in which [Dr Jones’] professional judgment concerning a primary interest (such as a patient’s welfare or the validity of research) tends to be unduly influenced by a secondary interest (such as financial gain).” In theory-however implausibly-if Dr Jones held no such belief about the company or, for whatever reason, had no need at all to concern himself with financial support in the future (let’s posit that Dr Jones is fabulously wealthy, on the scale of Warren Buffet!), some might argue that no COI was actually present. Nevertheless, erring on the side of caution, one should assume a COI in a case such as this, and disclose the financial relationship, as Dr Jones did.

Given the many safeguards built into this vignette-for example, no contact between Dr Jones and the company during the writing and no “pre-approval” of the paper by the company-one could argue that the COI in this case was not a serious or major one. But that is often not the case. One can easily imagine (and sometimes, document) instances in which the pharmaceutical sponsor puts considerable pressure on the writer/researcher to “play ball” and produce a paper consistent with the corporate interests of the company. Recently, there have been disturbing reports in which some companies are alleged to have concealed “negative studies” of their product from the very physicians who were speaking or writing about them.8

Are studies underwritten by pharmaceutical companies inherently biased, owing to financial COI?

No, although it is easy to become cynical about such studies. For example, there has been much written about the disproportionate number of “positive” (favorable) drug studies associated with pharmaceutical company underwriting. However, there remains some question about the underlying basis for this finding. For example, one widely cited study found that industry sponsorship is disproportionately linked with positive results, in published studies of psychiatric drugs.7 This is certainly of great concern and importance. However, Perlis and colleagues7(p1959) also acknowledged the possibility that “industry sponsorship may allow larger and better-designed studies, with greater statistical power to identify significant differences if such differences exist. Indeed, the median number of subjects was larger among studies in which conflict of interest was present. Industry-funded trials would naturally examine drugs already suggested to have efficacy in earlier-stage trials.”

Similarly, a comprehensive review of pharmacoeconomic studies, carried out in the United Kingdom at the Centre for Health Economics, University of York, concluded that while there are some causes for concern, given the fact that most pharmacoeconomic studies report positive findings for the sponsor’s drug, “a more detailed analysis suggests that while the methodological quality of some published studies may be poor, the main reason for positive results is that companies only sponsor economic studies where a positive outcome is likely.”9

It has been argued that the best defense against inaccurate and misleading scientific presentations lies in the accurate, scientific scrutiny of the article’s methods, design, statistical analysis, and conclusions.10 As Klein and Glick10 aptly observe, “the important issue is whether the presentation truthfully reflects the known scientific facts and draws justifiable conclusions.” If a study or article meets this test, the issue of financial underwriting is not of great public health importance. If the study fails this test, the research is not redeemed by its independent funding source.

A number of authorities have argued in favor of transparent access to all ongoing clinical studies; pharmaceutical company disclosure of any “negative” studies; availability of the researchers’ “raw data” to journal editors10; and, as this writer once suggested, a national online “clearinghouse” of data, disclosing all financial ties between researchers and the pharmaceutical industry. Daniel J. Carlat, MD,11 has also called on journals to “require authors to disclose all financial ties to any health care–related company-whether seemingly relevant or irrelevant to the topic of the article. It would then be up to the readers to decide whether these ties represented true conflicts.”

These ideas all have points in their favor. Yet it remains true that a researcher utterly without any COI may still produce an atrocious and misleading piece of research. Conversely, a researcher or writer who has an acknowledged (or even unacknowledged) COI may produce an article or study that is accurate, meticulously designed, and scrupulously analyzed. Journals can do only so much to assess COI in submitted articles, given limited time, staff, and financial resources. As Klein and Glick10(p3026) observe, “the complex COI issue cannot be dealt with by an editorial fix.”

In summary, there is no single, universally accepted definition of COI. In this writer’s view, Thompson’s definition4-which emphasizes likelihood of COI and undue influence-is perhaps the least problematic, notwithstanding its failure to distinguish moral obligations from mere “interests.” A COI per se does not indicate wrongdoing, but whether one discloses a COI and how one manages a COI does have important moral implications. Pharmaceutical company sponsorship of a research study warrants careful scrutiny for COI, but there is no valid reason to assume that sponsored studies are inherently biased or unreliable. Ultimately, physicians need to equip themselves with the analytic skills and understanding that will allow them to distinguish real science from “junk science.”10,12

References:

References

1.

DuBois JM.

Ethics in Mental Health Research: Principles, Guidance, and Cases.

New York: Oxford University Press; 2007.

2.

MacDonald C, McDonald M, Norman W. Charitable conflicts of interest.

J Bus Ethics.

2002;39:67-74.

3.

Conflicts of Interest: responsible conduct of research. Columbia University.

http://ccnmtl.columbia.edu/projects/rcr/rcr_conflicts/foundation/index.html#1_1

. Accessed October 20, 2009.

4.

Thompson DF. Understanding financial conflicts of interest.

N Engl J Med.

1993;329:573-576.

5.

Brody H.

Hooked: Ethics, the Medical Profession, and the Pharmaceutical Industry.

Lanham, MD: Rowman & Littlefield Publishers; 2007.

6.

International Committee of Medical Journal Editors. Conflicts of Interest.

http://www.icmje.org/ethical_4conflicts.html

. Accessed October 20, 2009.

7.

Perlis RH, Perlis CS, Wu Y, et al. Industry sponsorship and financial conflict of interest in the reporting of clinical trials in psychiatry.

Am J Psychiatry.

2005;162:1957-1960.

8.

Kaplan A. Forest under fire.

Psychiatr Times.

2009;26(4):1, 7-8. http://www.psychiatrictimes.com/display/article/10168/1399086. Accessed October 20, 2009.

9.

Barbieri M, Drummond MF. Conflict of interest in industry-sponsored economic evaluations: real or imagined?

Curr Oncol Rep.

2001;3:410-413.

http://www.upf.edu/cres/_pdf/interest.pdf

. Accessed October 20, 2009.

10.

Klein DF, Glick ID. Conflict of interest, journal review, and publication policy.

Neuropsychopharmacology.

2008;33:3023-3026.

11.

Carlat DJ. Conflict of interest in psychiatry: how much disclosure is necessary?

Psychiatr Times.

2006;13(1):7-8.

12.

Ghaemi SN. Good clinical care requires understanding statistics.

Psychiatr Times.

2009;26(3):31-32.

http://www.psychiatrictimes.com/display/article/10168/1385693

. Accessed October 20, 2009.

For further reading
Dubovsky SL, Dubovsky AN. Psychotropic Drug Prescriber’s Survival Guide: Ethical Mental Health Treatment in the Age of Big Pharma. New York: WW Norton; 2007.