Science is very competitive ... in large part because of the many initiatives undertaken by the federal and state governments that have been described at this meeting. These efforts have sent signals to academic researchers and their institutions that it is acceptable, indeed encouraged, for them to market their discoveries, and many have taken up the challenge with considerable vigor.
There has been a dramatic growth in recent years of various collaborative arrangements between academic scientists and their institutions and private industry, and the trend shows no sign of abating. A 1991 survey found approximately 12,000 university faculty, over 22,000 post-docs, and just under 17,000 graduate students involved in university-industry research centers at more than 200 universities. A 1994 study of 210 life science companies found that 56% support research in universities or teaching hospitals, a 10% increase since 1984. Moreover, 69% of those surveyed planned to increase their investments in university research over the next three years.
The benefits of University-Industry collaborations are numerous for industry, universities, and the larger society. For universities, benefits include a new source of funding, access to raw materials and instrumentation, the transfer of research to product development, training, education, and future employment opportunities for their students. For industry, the benefits can mean access to first-rate scientists, inexpensive labor in the form of post-docs and students, advantageous licensing agreements, reduced costs of conducting new R&D programs, and an improved public image. The larger society also gains from the increase in knowledge, new products, and a general boost to the economy.
The concerns about such research are also numerous, centered primarily on real or perceived conflicts of interest. Some critics have charged that objectivity and disinterestedness among scientists are being severely challenged by an "entrepreneurial atmosphere that has begun to alter the ethos of science." Although the promise of financial gain may be only one of several factors that could precipitate a conflict of interest, for many, disinterest and objectivity are nearly invariably compromised, in appearance or in fact, when the financial stakes for a researcher are too great. It is important to realize that the injection of large sums of private money into the research enterprise in recent years has become very problematic for science.
A conflict of interest refers to situations in which financial, professional, ideological, or other personal considerations may compromise, or have the appearance of compromising, a scientist’s professional judgment in conducting, interpreting, evaluating or reporting research, or when asked to provide independent expert advice. Such conflict could unduly affect the choice of research protocol, the collection, analysis, interpretation, and reporting of data, the use of certain instruments or statistical methods, the recruitment of research subjects, the evaluation of products, the communication of results, and the proffering of advice. There are at least two features of such conflicts that spark concern. One is their capacity to compromise independent and impartial professional judgment, and the other is that they can undermine the trust relationship between scientists and their clients or students, between scientists and their colleagues, and between scientists and the public.
There is a strong expectation, both inside and outside of science, that scientists will maintain the highest standards of objectivity and freedom from bias in their work. Whether in the conduct of their own research, in their role as impartial advisors on matters of public policy, or as reviewers of the work of other researchers, scientists’ independent professional judgment is a highly valued resource and it is what is expected of them.
A conflict of interest, however, may bias their judgment, thereby making it less reliable. For example, if researchers stand to profit from research on a drug owned by a company in which they own stock, then the outcome or direction of their work may be affected in a way that skews their research design towards a favorable outcome. Or, in another instance, close financial ties to industry may unduly influence the research priorities of scientists.
A 1991 survey found that nearly two-thirds of almost 500 university-industry research centers indicated that companies had moderate or strong influence over their research agendas. This is not necessarily bad. But the concern here is that industry support will prompt universities and faculty to pursue research with a more likely commercial payoff than more fundamental basic research with less certain applications. This concern was recently voiced by Harold Varmus, Director of NIH, in noting with some alarm that "the intense commercial interest in gene therapy is prompting a stampede into clinical trials and pressure for quick results -- before the basic science has been worked out."
By virtue of their special knowledge, scientists are particularly well qualified and sought after to point out opportunities and dangers associated with highly technical matters. But where do we look for a trusted opinion on the safety and efficacy of a proposed genetic therapy when, as Science magazine reported, "Today, most prominent geneticists have links with the genomics industry." Where can journal editors go to solicit reviews when, as an editor for Lancet recently lamented, that he "often can’t find anyone who doesn’t have a financial interest" in the drug or therapy under review. And the director of FDA’s Center for Biologics Evaluation and Research recently complained that "conflict of interest is a real problem"; it’s already hard to find academics or even NIH scientists who are free of ties to industry to serve on advisory panels. In securing knowledge, society relies on scientists, and it cannot readily dispense with that reliance except at considerable cost. The loss to the public of disinterested, independent advice would be a high price to pay.
In assessing the likelihood that professional judgment will be unduly influenced, or appear to be influenced, by competing interests, one should consider the following:
(1) the value of the competing interest; the greater its value, the more probable its influence; (2) the nature of the relationship that produces the conflict; longer and closer relationships increase the potential problem; and (3) the extent of discretion on the part of the scientist; the less authority the researcher has in a particular situation, the less latitude for improper influence, and the more routine and standard professional practices are, the less room there is for exercising judgment, and hence for the intrusion of improper influence.
My definition of conflict of interest referred not only to actual compromises of professional judgment, but to the appearance of compromising behavior as well. One may ask, why be concerned about appearance if one is worried about maintaining a scientist’s independence and impartiality? It is, after all, finding oneself in an actual conflict that can impair judgment, not how it appears to others. Well, the reason for concern about appearances is the threat posed not to decisions, but to trust and the relationships it nourishes.
Science advances and is sustained by a presumption of honesty and collegial trust. And to the outside world, the credibility of science rests on its presumed independence and disinterestedness. The general public’s robust support of science and technology and its willingness to grant researchers considerable autonomy in performing their work are premised on a good-faith understanding that they will fulfill their social responsibilities as an impartial and reliable source of expert advice.
One way that scientists earn both public and collegial trust is through the traditional norm of openness, where data are shared and findings are open to close scrutiny by colleagues. But such openness appears to be impeded in many cases of university-industry partnerships.
A 1991 survey of university-industry research centers found that in more than half the cases companies can delay publication of research findings and more than a third could require that information be deleted from a publication. These findings led the authors of the study to conclude that "to elicit industry participation..., universities have weakened their norms regarding full, public disclosure of findings." And a study of 106 biotechnology companies found that 41% derived at least one trade secret from their collaborations with universities, with smaller firms reporting a significantly higher proportion than larger companies.
There is the risk of eroding public trust and confidence if economic self-interest, at least as reflected by these data, is perceived as substituting for or adversely affecting scientific integrity. Perceptions matter, and reports of conflict based on appearances can undermine trust in ways that may never be adequately restored. As the American Medical Association warned, "the mere perception of conflict of interest may be enough to cast significant doubt on an exemplary research program." Consequently, apparent conflicts should be evaluated with the same diligence as actual conflicts.
At this juncture, let me be clear that I do not believe that these problems are inevitable. Nor do I believe that a conflict of interest is intrinsically wrong or problematic. It is not inconceivable, for example, that a researcher’s financial or other interest will be compatible with accepted norms of research. Moreover, there may be circumstances where a conflict of interest may be more acceptable than avoiding or escaping the conflict because there is no other reasonably available or better alternative for securing benefits than from the relationship that gives rise to the conflict. Consequently, in responding to conflicts of interest, we should be careful to avoid the kind of attitude that a frustrated scientist once attributed to the U.S. Environmental Protection Agency. He remarked that EPA’s regulatory philosophy was so restrictive that it would likely one day ban life because it inevitably leads to death. We must think of conflicts of interest as a situation that a researcher may encounter, and that may or may not lead to adverse consequences. The aim is not necessarily to eliminate conflicts of interests, but rather to manage them in order to maximize the benefits of such a relationship and minimize adverse effects.
A researcher can avoid the suspicions or negative influences associated with a conflict of interest, or its appearance, by refusing to enter into a particular relationship. But however much care one takes, no one can always foresee how relationships will ultimately evolve, so one cannot in every instance be expected to take precautions necessary to avoid all conflicts of interest. Once in such a situation, one can always escape by ending the relationship or by divesting oneself of the compromising interest. However, in some cases these options may be neither practical nor advisable, especially if the relationship in question produces benefits not otherwise available.
Acceptance of a conflict situation requires that outsiders know about it. To operate with a conflict of interest is of concern for reasons noted earlier; to do so without alerting others adds fuel to those concerns because it fosters a deception. Disclosing one’s conflicting interests, while not ending the conflict, does end the deception of allowing one’s judgment to appear more independent and reliable than it may be, which responds to the concerns raised earlier regarding trust. And in theory, disclosure gives others the opportunity to monitor and appreciate the possibilities of bias affecting the researcher’s judgment so that they can assess the reliability of the information they receive.
That is why more and more scientific and medical journals are requiring authors to disclose financial and other conflicts of interest that might bias their work. And that is why professional associations are increasingly including disclosure provisions in their code of ethics for their members.
But disclosure also has its weaknesses. First, it only reveals that a problem may exist; it
does not provide guidance for solving it. Second, it places the burden on the recipient of the information to decide to what extent the disclosed conflict situation has an effect on the research. Yet, the receiver may not know enough to evaluate the information usefully. And, under such circumstances, disclosing the information may simply act to increase the levels of anxiety among patients or readers who may, without real merit, end up distrusting the source of the information.
Nevertheless, disclosure is the linchpin of the federal government’s conflict of interest policies adopted by NSF and PHS. Those policies, which went into effect in October 1995, focus only on financial conflicts of interest in part because of anxiety about the influence of money on researchers’ judgments, and in part because they are measurable and more easily managed than other types of conflicting interests. They establish a monetary threshold -- above $10,000 for the investigator, spouse and dependent children combined in any 12-month period or more than a 5% ownership interest in any single entity -- that would trigger a disclosure and review, but not ban the relationship. The NSF regulations exclude institutions that employ less than fifty persons and the PHS rules exempt any ownership interests in an institution if the latter applies for funding under the Small Business Innovation Research Program. These provisions are especially relevant to smaller companies, which are more likely than larger ones to offer equity interests in their company to academic researchers in return for their expertise. The policies place primary responsibility for implementing the rules on individual investigators and their institutions. This has led to some variance in university policies. For example, some institutions now forbid a researcher to have any equity interest in a company sponsoring that faculty member’s research, while other institutions have no such restrictions. It remains to be seen how effective these rules are in managing conflicts and in maintaining public confidence in efforts to promote university-industry research collaborations.
In closing, let me say that while I believe that researchers, their institutions, and the companies with which they partner must be sensitive to the very real concerns expressed about the effects of conflicts of interest, they need not always be on the defensive. It troubles me that more often than not, a stigma is attached to being in a conflict of interest situation. But a conflict of interest should not automatically impugn one’s motives or character, or cast a dark shadow over a professional relationship. Such situations do not necessarily lead to bad outcomes. And it is very possible that a conflict of interest situation, if managed well, will produce benefits that outweigh any potential adverse consequences.
So the message I want to leave you with is this: if we want to be effective in managing conflicts of interest, then ethical guidelines and public and institutional policies should target the circumstances of the conflict and how to manage it rather than cast doubts on the integrity of the relationship of those involved. And all of us in the research community, whether in academe, industry, government, or non-profits such as AAAS, need to do a better job of educating the public and our friends in the media about what we are doing to earn their trust.