Conflict of interest: Disclosure to whom? And how?



Conflict of interest in academic science is a controversial, but most of all a highly emotional, issue in the academy. Scientists and administrators disagree vehemently about whether it is a good or bad thing, and many aver that it has no impact on research—or that it is no one’s business. The thing is that the term conflict of interest is descriptive of a state, not a quality, and its effects or impacts can only be known, sometimes tragically and always far beyond the individual involved, after-the-fact. This is why disclosure is generally considered to be a poor means of avoiding bias or harm—it is too little, too late, and we are left with not knowing who or what to trust, and may need to run around retracting articles, shutting down trials, imposing sanctions, firing people, or engaging in yet another political effort to rein in a growing ill. Physician, heal thyself: an ounce of prevention is worth a pound of cure.

At the moment, though, disclosure is the modus operandi. Even that took decades to put in place, and efforts to make disclosure, let alone enforcement, more robust have been disappointing exercises in which the whole point—the integrity of science and scientific institutions—bows to external pressure and resistance. One of the many disappointments in the recent (2009-2011), largely failed effort to strengthen conflict of interest rules and enforcement was the rejection of additional reporting to the federal government. In making my public comments at the time, I for one had called for reporting of all disclosures, patent/royalty data, equity values, contract terms, management plans, and other conflict-specific data to a central website created and maintained by the Public Health Service (HHS/PHS), the funding agent soliciting comments on proposed new rules. Some others made similar recommendations, as discussed below. In a rationalized interpretation of lessening the administrative and financial burden to institutions, the final rules instead called for each institution to create its own site and/or respond to requests for information on a case-by-case basis. Appeasement often yields the irrational.

A few weeks ago the Institute of Medicine (IOM) issued its own proposal for a central system, an elaboration of its original call for standardization of disclosure content and form in its 2009 report. For well over a decade, the IOM has taken a leadership role in convening discussion, promoting awareness, and recommending policy related to conflict of interest in medical research. This latest effort continues in that spirit, but I wonder if it too has bowed to pressure as it broaches the trickier shoals of implementation. I’ve said it before and I’ll say it again: ethics isn’t about preferences. While it is pleasant to hope that accommodation and conciliation will smooth the way, sometimes we just have to do what will get the job done rather than what we think we can get people to agree with. In the end, since the resistant resist regardless, a strategy that reacts to such pressure tends to undermine the whole. And once again, a claim of maximizing efficiency and burden reduction seems, at least on the information available, to do the opposite.

I’ll limit my remarks to these overall objections, because the details don’t matter if the high-level stuff is off.

• The IOM proposes a new non-profit in preference to what it calls a “federalized” system; this is an interesting choice of words in that much of the conflict of interest at issue, even if it also involves industry, is related to research or relationships either funded by or otherwise regulated by the government, which has existing authority to require reporting. If there is a formal requirement to report, no “acceptance” is needed; it is part of compliance.

It’s worth mentioning that during the 2009-2011 round of conflict of interest discussions, some of the same institutions making the IOM proposal offered suggestions for a central, federal system. For example, Northwestern had suggested enhancing the existing NIH Commons system, now used for grant administration, and the Association of American Universities and Association of American Medical Colleges, the latter also in the IOM proposal group, suggested NIH develop standard input formats for a system and that HHS consider maintaining it. Their initial instincts were correct.

• The system is still aimed at individuals rather than institutions and relies on their interpretations of what might influence their judgments or actions; considering the known problems with individual disclosure, including the belief that conflict of interest affects others but not oneself, what reason is there to believe that such a system would be any better than what we have now?

• The proposal leaves open the big question in conflict of interest policy: disclosure to whom? Its numerous references to “requesting parties” raises the question of whether a requesting party—say, an investigative journalist, a researcher, an interested citizen—could be denied access to the data. It is silent on this issue. But several aspects of the proposal, from the repeated reassurance that the data is “owned” by the individuals reporting (this is arguable) to the “as authorized” qualification for requesting parties and the governance structure make one wonder. It is a limited-purpose, 1:1 system.

• In putting the control in the hands of the discloser, the proposal seems to be organized backward, paradoxically making it far more burdensome than it need be. The details of data entry are unclear, so the following comments come with that caveat. But it appears that individuals might need to respond to multiple requests, or have multiple data files for multiple roles, rather than disclose all data relevant to conflict of interest (not to an individual role in an individual organization) one time to cover all possible requests, and then have the system mine data and generate outputs based on search parameters. From an information technology and information services standpoint, the proposal is inefficient, even primitive. It would soon be out of date if, as stated, types of potential “requesters” had to agree on output content and formats in advance: needs change, and those querying the system may have reason to get additional information beyond, say, that for one individual or one circumstance. Further, the proposal seems to narrow the categories of information reportable to be those covered by the Sunshine Act, a subset of conflicts of interest of interest to users.

• The proposal to create a new entity or put it inside an existing entity is both expensive and seems to misunderstand that such a system should be independent of the reporting agencies and individuals.

Taken together, the proposal appears to be a system designed for the disclosers, not the users. The whole point of disclosure on matters of public concern—and those that involve public money or public welfare—is that it be public; a closed system is not the solution to a transparency problem. The weak objections to the federal option, and the inappropriate analogies used in arguing the proposal, undermine the credibility of the proposal. While a central system is the right way to go, from the standpoint of efficiency, utility, access, and effectiveness, it makes sense for the government to house this system in a single repository from which everyone—including other government agencies—could access data on demand. It also makes sense for inputs, not outputs, to be standardized and submitted by primary institutions to which the data is disclosed, not by individuals. Such a system would simultaneously allow compliance tracking; integration with other government data such as grant awards; research on conflicts of interest at the individual and institutional level; and an important new means of potentially changing the current culture of conflict interest and restoring trust in—and support for—the scientific enterprise and scientific institutions. This last possibility should be the overarching goal of the system, and should drive its design.