Sunday, March 30, 2008

AAMC Policy on Institutional COI: An Assessment

In February, the Association of American Medical Colleges and the Association of American Universities jointly released a report on financial conflicts of interest in human subjects research (link is to new release which in turn has a link to the complete document):


http://www.aamc.org/newsroom/pressrel/2008/080228.htm


While you could not necessarily tell from the unwieldy title, the focus of this document is on institutional conflicts of interest. The AAMC has already issued reports on how academic centers ought to handle individual faculty conflicts of interest, and had previously issued a preliminary report on institutional conflicts. The intent of this report is apparently to complete work on the institutional COI front, and the report contains as an appendix a template for a detailed university or medical center policy to that end. The report calls for all U.S. institutions doing human subjects research to have such a policy in place within the next 2 years.


Now, to assess such a document, it is reasonable to keep in mind some important "test cases" for institutional COI policies. Two recent challenges to institutional integrity are:
  1. The 1999 Jesse Gelsinger case from U. Pennsylvania, in which a healthy 18-year-old died in a gene therapy experiment, with both the principal investigator and the University itself owning part of the biotech statup company that would have profited had the therapy worked, and later questions about the adequacy of risk disclosure
  2. The 1998 deal between Novartis and UC-Berkeley, in which the University received $25M in exchange for first rights over the next 5 years to any research discoveries from the Department of Plant and Microbial Biology, and imposing a confidentiality agreement on all department faculty
The AAMC-AAU report does a number of things that in my view are positive. It sees the COI problem related to human-subjects research as dual--there is the need to protect the subjects of the research from undue risk, and the need to protect the integrity of the research process to attempt to discover scientific truth, rather than the answer the manufacturer wants to sell a product. It recognizes that "institutional" COI can come in two forms. First, a highly placed official such as a dean may own stock in a company or otherwise be conflicted, and the problem is to make sure that that individual official's activities do not create a problem for research integrity or subject safety. Second, the university as a whole can be so dependent upon a large gift or grant, from a commercial source, as to raise questions about integrity that cannot be assigned to any one official (none of whom may actually receive any personal financial benefits from that source). Both of our test cases highlight the latter version.

The actual policies for oversight and review contained in the report and in the model policy appendix seem generally decent, and are certainly better than nothing, given that so few universities today have any sort of policy.

But there are also serious problems. The model policy offers a definition of institutional COI that is impossibly broad: "An institution may have a conflict of interest...in human subjects research whenever the financial interests of the institution...might affect--or reasonably appear to affect--institutional processes for the design, conduct, reporting, review, or oversight of human subjects research." By this definition it is hard to see what aspect of institutional activity is not a COI. The ethical point of COI is missed, which is the concern that a financial or commercial interest comes into conflict with the institution's duty to maintain research integrity and to project subjects from harm or exploitation. (Admittedly, if you read all of that page of the document, the preceding discussion makes the point much clearer; but if a policy document has a specific passage called "definition," we ought to have reason to hope that what is offered there accurately defines the term in question.)

The far more serious problem is that the policy seems toothless to address serious institutional-level COI of the sort at Berkeley and Penn, when the institution as a whole is in bed with the business interests that create the COI. The best that the policy can come up with is the idea that there might be a 7-member committee to review allegations of COI, and that 2 of the 7 might be "public" members who are not on the institution's payroll in any way (though 1 of the 2 might have another tie such as being an alum). Certainly having some sort of public involvement in institutional oversight is a good idea. But the problem remains that if, for example, the President of the University decided to sell off the Department of Orthopedic Surgery to a maker of artificial knee prostheses, we would somehow have to expect the COI committee to have the backbone to tell the President and the Trustees that this is a bad idea.

By contrast, as I briefly discussed in HOOKED, some Canadian commentators seemed to grasp what was needed when they called for each Canadian university to agree to submit such complaints and concerns to a super-committee that would function at the national level and that would not be beholden to any individual university. I see no obvious U.S. national body that could function in quite that fashion, so as a substitute, I would wonder about some sort of round-robin arrangement, in which certain sorts of institutional COI issues at University A would be referred for review to a committee based at University B, University B send its dirty laundry off to University C, and so on. (You would avoid pairing off A and B, for instance, so as to avoid any charges that the review was simply an ol' boys' club of "You scratch my back, I'll scratch yours.")

The bottom line, I would propose, is that if a University claimed that a committee that consisted mostly of those on the University payroll could adequately policy institutional COI matters involving the entire University, it would be more or less like an individual COI review process that proposed to resolve all COI at the faculty level by identifying all faculty who had indefensible commercial COIs, and locking them all together in a room and requiring that they work it all out among themselves.

No comments: