Posted 09 January 2015
By Alexander Gaffney, RAC
Conflicts of interest among members of the US Food and Drug Administration's (FDA) influential advisory committees don't have much of a practical impact on the outcome of voting decisions, new research argues.
The paper comes as several research papers in recent years have made the opposite case—that advisory committee members with stated conflicts of interest do affect how a committee makes a decision. For example, a September 2014 paper in the journal Milbank Quarterly found "a pro-sponsor voting bias among advisory committee members who have exclusive financial relationships with the sponsoring firm but not among members who have nonexclusive financial relationships."
A 2006 paper in the Journal of the American Medical Association (JAMA), meanwhile, found only a "weak relationship between certain types of conflicts and voting behaviors."Even still, the authors noted, "for each one person with any conflict, there was a 10% greater likelihood that the meeting would favor the index drug."
The fear, then, is that such conflicts will cause a committee to vote in favor of a product that it might not otherwise support after an objective review of the evidence. Such conflicts can also taint the public's view of the process as fair or impartial. For example, a recent investigation by The Wall Street Journalfound that one cardiologist sitting on a medical device advisory committee had received at least $100,000 from a company whose product he had reviewed while on the committee. “Undisclosed conflicts raise questions about the decision-making capacity of the committees and whether the public can have confidence in their recommendations,” Joseph Ross, an associate professor at Yale School of Medicine, told the Journal in reference to their investigation.
Not Much Impact?
But a new paper published by George Mason University School of Law scholars Joseph Golek and James Cooper makes a more nuanced argument: Conflicts of interest might affect individual votes, but they rarely affect the outcome of how a committee as a whole is expected to vote.
Indeed, expectations are at the heart of Golek and Cooper's paper, FDA Advisory Committees: Conflicts of Interest and Voting Relative to Benchmarks. The duo adopts a "wisdom of the masses" approach by comparing advisory committee decisions with stock price of companies with products before the committee.
In theory, Golek and Cooper say, life science investors—who have access to almost the same information seen by advisory committee members—should have a pretty good idea of how a drug product would fare before a committee. Drugs with major safety or efficacy issues, for example, are much more likely to be rejected by a committee, whereas safe and effective drugs for a life-threatening condition are likely to be recommended.
The researchers found that investors were rarely surprised by the way conflicted advisory committees voted. "Only 42% (13) of the approval decisions that surprised the market come from conflicted ACs, despite conflicted ACs comprising the majority of the sample," the researchers found.
In other words, in the majority of cases, the advisory committees ultimately voted in the way investors believed they would, and the presence of a conflict of interest had little bearing on the ultimate outcome. "Conflicted advisory committees are also associated with a minority of market surprises involving rejections (43%)," Golek and Cooper noted.
A High-Level Look at Conflicts of Interest
Their research also looked at advisory committee decisions on a macro level, comparing overall advisory committee approval rates to overall FDA approval rates. Because advisory committee recommendations do not need to be followed by FDA, the researchers posited, if conflicted committees voted "incorrectly" their approval rates would be much higher than FDA's eventual rates. Instead, Cooper and Golek found that FDA approval rates did not differ much from advisory committee approval rates.
FDA and Advisory Committee Approval Rates by Presence of Conflict
| ||AdComm Approval||FDA Approval|
|Conflict in Favor of Drug||80.0||90.0|
|Conflict not in Favor of Drug||67.7||73.6|
Consistent with Golek and Cooper's prior research, they said their paper found "little evidence that conflicts lead to biases in favor of drug company interests," they wrote.
The analysis, however, does have some notable shortcomings. For example, there's no analysis on whether FDA decided the same way as the advisory committee on individual drugs (i.e. a drug might have been rejected by the committee but approved by FDA). Prior research has indicated FDA only takes an advisory committee's advice into account around 70% of the time. A one-to-one analysis, then, might reveal whether FDA approved and rejected the same drugs called for by its advisory committees.
In addition, influence can be felt on an advisory committee in more ways than a simple "yes" or "no." Recommendations made by advisory committees can often be nuanced, and reflect recommendations for safety warnings or further study on a particular point. Discussion at advisory committee meetings, too, can often raise nuanced points of interest to FDA regulators.
Neither does the study track which drugs were later recalled for potential safety issues. While the FDA decision at the time might be seen as "correct," the drug's market record might be seen as a better indicator of whether the drug was, in fact, properly approved.
FDA Advisory Committees: Conflicts of Interest and Voting Relative to Benchmarks