The Institute of Medicine (IOM), part of the prestigious National Academy of Sciences, recently published a report on FDA oversight of drug safety. The IOM report was sufficiently critical of the FDA to generate newspaper headlines like "FDA Told U.S. Drug System Is Broken" (Washington Post, September 23, 2006, p. A1) and "Study Condemns FDA's Handling of Drug Safety" (New York Times, September 23, 2006).

Resident Scholar John E. Calfee

One of the report's recommendations focused on FDA advisory committees. These committees (adcoms) are standing groups composed almost entirely of well-known academics and practitioners. The FDA uses adcoms a lot. Before deciding whether to approve a new cancer drug, for example, the FDA may convene a meeting of the Oncology Drug Advisory Committee to review various materials along with the views of the FDA's own staff. The FDA usually, but not always, follows the advice of these committees. Their importance should be perfectly obvious.

The FDA requires adcom members and "voting consultants" to describe their potential conflicts: consulting, grants, stock ownership, and so on, related to the manufacturer of the drug under review or manufacturers of competing drugs. If such potential conflicts are non-trivial, the FDA decides whether to exclude the panelist from the instant issue (recall that these are standing committees that can be convened on a series of matters), or to grant the panelist a waiver to participate while also making the essential details public. The FDA often grants such waivers. When asked over and over again why it does this, the FDA always replies that the search for the very best pharmaceutical research scholars inevitably turns up numerous conflicted researchers because pharmaceutical firms also want to obtain the best possible expert advice. Eliminate conflicted researchers, and you tend to eliminate the most valuable adcom members.

The IOM authors bought this story, but only part way. The IOM report said the FDA should toughen its conflict-of-interest standards so that a “substantial majority” (at least 60%) of the members have no significant sufficient conflict at all. But even that did not satisfy a lot of FDA critics, some of whom quickly denounced this part of the IOM report on the grounds that the FDA should require that 100% of adcom members, not a mere 60%, should be bereft of significant conflicts of interest.

What we have here are two competing visions of what motivates top-flight researchers and clinicians. The crack-down-on-conflicts-of-interest vision sees adcom members casting their votes with a sharp eye on where their next research grant or consulting contract is likely to come from. A tough vote for a favorite client--or against one of that client's competitors--could generate considerable good will where it counts most. The predictable result: a clear correlation between where consulting interests lie and where the committee members direct their blessings and curses.

But here's a very different vision. Think about what is in the self-interest of a typical adcom member culled from the nation's elite academic researchers and clinicians. Yes, their income comes from salary, grants (government and private), consulting, maybe royalties, and so on. But in the researchers' world, the dominant coinage is reputation among peers, from which all else eventually flows. Committee members must know that if they vote more consistently with their short-term financial interests than with their take on the science, their famously picky colleagues will catch on. That would mean impaired professional relationships and eventually a distressing erosion in awards, prestige and even income. In a world dominated by enlightened self-interest (not to mention ethical sensibilities), correlations between votes and financial interests will be coincidental, and will tend to wash out over time.

The redoubtable Sidney Wolfe and his colleagues at the activist institution Public Citizen recently did us all a favor by researching four years of voting patterns in a total of 221 meetings of 16 advisory committees (the counts included voting consultants as well as committee members). The results were published in the April 26, 2006 issue of the Journal of the American Medical Association.

One finding was that only 28% of voters had conflicts. More importantly, after deleting all votes by members with a stated conflict, not a single recommendation would have changed, although majorities would often have changed. Subsequently, however, FDA staff posted a simple re-analysis of the JAMA data. They pointed out the simple fact that if a member's interest was in a competitor rather than the manufacturer of the drug under review, a vote for approval was contrary to personal interest and vice versa for a vote against approval. After reclassifying the data to reflect this information, the voting actually tended to be against members' interests. And by the way, let us give credit where it is due: Public Citizen submitted its study to a high-profile journal even though many of the results undercut that group’s own views. Peer reputation is important for them, too.

Until new and compelling evidence arrives, I think we can dismiss practically all worries about conflicts of interest in FDA advisory committees. Yes, we need transparency, a very useful tool. But any attempt to abolish conflicts of interest would jettison good talent with no perceptible gain in integrity, tying the FDA’s hands as it goes about the crucial task of marshalling the best scientific advice.

John E. Calfee is a resident scholar at AEI.