There is little agreement about how conflicts of interest should be disclosed to potential research participants or how that process should be accomplished. Our review of the conflict of interest disclosure policies of academic medical centers documents the current state of affairs and identifies the variety of approaches taken by their IRBs and COICs as they respond to the call to disclose conflicts of interest to potential research participants. About half of the institutions' IRB or COIC policies made explicit reference to disclosing conflicts of interest to potential research participants. This lack of guidance at many institutions is troubling, given (1) the challenges of disclosing conflicts of interest to research participants and (2) evidence that suggests that potential research participants want to know about financial interests.8
It should be noted, however, that some of these institutions might not have explicit guidance because they do not endorse disclosure to potential research participants as an acceptable management strategy. Of those institutions that explicitly addressed disclosure to potential participants, most guidance was found in the informed consent document template. Just over half of the explicit instances of guidance included verbatim language that investigators could use in designing their informed consent documents. Where guidance exists, there is considerable variability among institutions concerning the amount and type of information that should be disclosed.
With regard to the detail found in disclosures, our data suggest that most institutions that explicitly discuss disclosure of conflicts do not recommend or require a great deal of information beyond the name of the sponsor. Even statements that include information about the nature of the relationship are brief, mentioning only that a sponsor supports the research. In terms of the three goals of disclosure, it is possible that the lack of detail required for disclosures is consistent with the goal of protecting against legal liability without affecting meaningfully the research enterprise.
In contrast, if the goal were to allow people to make informed decisions about participation, it seems reasonable that more information would have to be provided regarding the nature of the financial relationships and perhaps even the way in which funds are allocated. More important, it may be necessary to inform potential research participants of the possible consequences of these relationships. A small minority of institutions made statements indicating that the investigator or institution could benefit financially from the results of the study. Without such statements, it is not clear whether or how people understand information about financial relationships in clinical research and the implications of such relationships for their desire to participate. With regard to the goal of deterring conflicts of interest, it is highly questionable whether the sparse detail required by most institutions would serve as a disincentive for investigators to hold substantial financial conflicts of interest. Exceptions to this majority include institutions requiring that potential research participants be told that the investigator might benefit financially and those that encourage discussion between the investigator and the potential participant.
Thus, it appears that the majority of the institutions' policies are most consistent with the goal of avoiding legal liability. Whether or not this is the intent of those framing the guidelines, the resulting template language appears to serve this legalistic goal better than the goals of ensuring an informed decision or of deterrence. Whether the recommended language would actually protect individuals and institutions from liability of course remains to be seen.
An alternative interpretation of the findings, however, is that institutions are interested in fostering more informed decision making but are unsure about what or how to disclose with respect to conflicts of interest. This uncertainty is understandable, because there are few published data to inform investigators and administrators about (1) the information that is viewed as “material” by different parties involved in research, and (2) how best to communicate that information to potential research participants (see Kim et al8
for one empirical study). Our review sets the stage for this line of research by identifying several models for disclosure to submit to empirical testing. For example, it would be interesting to know whether a simple disclosure stating that the investigator could benefit financially from the results of the study (Example 6) is better understood and more satisfying to potential research participants than a long description of the nature and possible consequences of the financial relationship. Our current work is focusing on such issues.
Our data also present an opportunity to examine conformity with the 2001 AAMC recommendation11
that disclosure language “should include an explanation of the fact that the financial interest in question has been reviewed by the [conflict of interest] committee, approved subject to committee oversight, and determined by both the committee and the IRB not to pose any additional risk to the welfare of research subjects or the integrity of the research” (p. 18). Only two of the disclosure templates (Examples 5 and 6) included such language. This small number might result from the fact that the AAMC recommendation assumes that a well-defined process is in place to make determinations regarding the risk of conflicts of interest to research participants or to the integrity of research. The failure of almost every institution to follow the AAMC recommendation in their policies could be a reflection that such a process is not yet well-established at many institutions. Our findings also add to the picture of conflict of interest management in academic medical centers recently presented in a report by the AAMC.4
The authors of the report note that, although progress is being made in many academic medical centers in the management of conflicts of interest, there are still areas requiring attention. Better compliance with AAMC guidelines might be accomplished by incorporating the guidelines into IRB accreditation criteria.
Our conclusions should be qualified by several considerations. First and foremost, our findings apply only to academic medical centers. The vast majority of industry-sponsored clinical research is conducted outside of academic medical centers, so the generalizability of our findings is not clear. Second, while we made every effort to identify the relevant policy statements and materials at each institution, additional information regarding disclosure of conflicts of interest may have been contained in unidentified documents. It is interesting to note, however, that our failure to obtain conflict of interest policies from every institution is itself telling with regard to the AAMC guideline11
that the “policy and the related information should be readily accessible to covered individuals and to the public: in addition to conventional means of communication, the policy should be placed on the institution's website, if one exists” (p. 16). Even for the policies we found, considerable effort was required to identify them. Third, some institutions' documented policies might not have kept pace with changes in practice. We are now conducting interviews with IRB and COIC chairpersons to understand current practices. Still, the review of documented policies is important in that, compared to undocumented, informal practices, documented policies should provide a baseline of assurance that conflicts of interest will be managed consistently and effectively. It is conceivable that accreditation of human research protection programs can and will play a role in establishing such a baseline in the future. Fourth, our focus on template language in the informed consent documents should not imply that consent forms are the only or even the most important aspect of the informed consent process. Especially given the complexity of the information involved, greater attention is required for other aspects of the consent process that include giving potential research participants adequate and understandable information that is relevant to the research. The AAMC guidelines11
for example, recommend explicit statements that encourage participant-investigator dialogue (see Examples 9 and 10).