FDA Panels: Too Many Conflicts or Too Little Expertise?


This Op-Ed was published by Pharmalive.com.

Last week, a paper in the journal Science argued that a 2012 law that loosened conflict-of-interest restrictions on FDA advisory panels could not only allow more drugs with troubling side effects to enter the marketplace, but was actually unnecessary. The discussion, which analyzed the utility of caps placed on waivers, once again raised the thorny debate over conflicts and panel members (read more here). But Diana Zuckerman, who is president of the National Research Center for Women & Families, a think tank, and long-time FDA observer, argues this only tells part of the story…

Should FDA Advisory Committee members be allowed to have financial conflicts of interest regarding the medical product they are recommending for approval (or recommending against approval)?  The answer is not as simple as it might seem.

The Searle Civil Justice Institute of George Mason University held a Congressional briefing this week on the subject and I was invited to participate. The funding for the institute came from the late Daniel Searle, former CEO of the pharmaceutical company GD Searle. The focus  was a report entitled “FDA Advisory Committees: An Empirical Examination of Conflicts of Interest,” that is being written by Joe Golec, Professor of Finance at the University of Connecticut, and two of his colleagues at the George Mason University School of Law.

The statistics compiled for the report focus on voting patterns for all the FDA advisory committee meetings on new prescription drugs in recent years. The results indicate that the committee members who get waivers allowing them to serve on the committee despite conflicts of interest vote very similarly to the committee members who don’t get waivers. The authors conclude that this shows that conflicts of interest don’t unduly influence voting patterns on FDA advisory committees. They also conclude that the members with conflict of interest waivers tend to have more expertise, and therefore are an important addition to the committees.

I do not doubt the data that the authors presented, but I question their conclusions. The publicly available information about conflicts of interest on FDA advisory committees is too limited, and the analysis misses the flavor of advisory committee meetings, as well as the not-so-subtle nuances regarding conflicts of interest.

FDA defines conflicts of interest as financial ties during the last 12 months, so even extensive financial ties in the recent past – or even 12 months plus one day earlier — would not be included and would not require a waiver. The FDA advisory committee that met over a year ago to discuss whether Yaz oral contraceptives are too risky to stay on the market is a perfect example.  Advisory committee members with previous extensive financial ties to Bayer, which sells the pills, but who did not have waivers voted in support of Yaz (and Yasmin, Beyaz, and other contraceptives made with the hormone drosperinone). This shifted the vote to keep these pills on the market.

In addition to the many conflicted advisory committee members who don’t get waivers because their financial ties are more than a year old, voting patterns don’t tell the whole story.

I’ve been to dozens of advisory committee meetings, and I’ve seen how members with financial ties to the company or product often talk more at the meetings. They may talk more because they know more. They may talk more because they want to show the company how smart or helpful they are. Whatever the reason, their greater participation can be influential. Many advisory committee members ask no questions and make no comments at these meetings, until required to explain their votes. The advisory committee members with more direct knowledge of the products, including those with financial ties to the company or the product, can greatly influence the vote when they talk more, ask softball questions, or steer the conversation toward topics of benefit to the company. These members may have grants or consulting relationships with the company.

For committee meetings reconsidering safety issues for popular products such as Vioxx, Avandia, Yaz, osteoporosis drugs, surgical mesh, hip joints, and breast implants, advisory members who have frequently prescribed or implanted the products being reviewed are not considered biased and also do not have waivers. These potentially more knowledgeable but less objective members influence how others vote, making a comparison between members with waivers and members without waivers rather meaningless.

After attending so many advisory committee meetings, and studying 89 of these meetings in our report what is striking to me is how many of the members are not truly experts worthy of giving advice to the FDA. In fact, many of these voting members don’t understand statistics and ignore the clinical trial data unless they support their desire to get the drug on the market or keep it on the market.

Advisory committee members tend to be clinicians who want more drugs to be approved, making comments like “if this drug can help one patient, we should get it on the market.” For example, I recently went to an advisory committee meeting for Merck’s new sleeping pill, suvorexant, where the members ignored the FDA’s concerns that the data indicated many patients would have trouble driving to work the next day and could even fall asleep at the wheel (read here).

These sleeping pills had a half-life of 12 hours. Most of the advisory committee members didn’t care about that at all. Instead, they focused on the fact that people have insomnia and need help falling asleep. I think that people who take sleeping pills are mostly concerned about getting enough sleep so that they can function well the next day. A good night’s sleep doesn’t seem so beneficial if it means falling asleep while driving to work the next day.

At a meeting a few months ago, FDA advisory committee members recommended approval for a TB drug that was five times as likely to kill the patients as the current standard of care, a statistically significant difference. The sponsor, Janssen, speculated that the high death rate happened by chance in this randomized double blind clinical trial.  Amazingly, that ridiculous justification was good enough to convince most of the advisory committee members. Apparently, they didn’t understand that the entire purpose of a statistical analysis of a randomized double blind clinical trial is to determine whether or not a difference in outcome occurred by chance – and this one almost definitely didn’t.

I am very pleased that FDA Commissioner Margaret Hamburg has asked agency officials to reduce the number of advisory committee members with waivers. Unfortunately, the waivers are just the tip of the conflict of interest iceberg. FDA advisory committee members continue to have many members with financial ties to the companies and no disclosure of who they are. The media have publicly outed some of those advisory committee members, but most of the time that information is not known to the public, or the reporters covering advisory committee meetings.

Meantime, the bigger problem is that so many FDA advisory committee members don’t understand statistics or truly value or understand the results of clinical trials. FDA is supposed to make decisions based on scientific evidence that patients are more likely to be helped than harmed by a new medication or medical device. FDA approval should not be based on speculation or wishful thinking about whether a drug might “help at least one patient.” When committee members ignore the documented risks and focus on their hope for unproven benefits, thousands of patients die unnecessarily.

The number of FDA advisory committee members with conflict of interest waivers is lower than ever, but many members still show clear bias in favor of approving drugs and medical devices that are not proven safe or not proven effective. Whether those FDA advisors have financial conflicts of interest, other types of bias, or lack of interest in scientific evidence will not matter to the patients who are harmed by these medical products. Unfortunately, that will be the legacy of too many FDA advisory committee meetings.