The U.S. Food and Drug Administration (FDA) has the responsibility to determine if newly developed medical products are safe and effective. Whether it is a prescription medication, a medication sold over the counter, a medical device, a vaccine, or another type of biologic, the product can be marketed for general sale in the United States only if it has FDA approval.

FDA advisory committees are the most visible part of the FDA approval process. They meet in public to review the most controversial and cutting-edge medical products, examining applications for FDA approval. Committee members discuss the strengths and weaknesses of the studies and their enthusiasm or concerns about the medical product under review. At recent FDA advisory committee meetings on controversial drugs and medical devices such as Vioxx®, silicone implants, and antidepressants, the media have provided the Congress and the general public with a glimpse of the approval process.

Questions have arisen about committee members' financial ties to the companies submitting applications, their commitment to scientific scrutiny, the independence and objectivity of the deliberative process, and inconsistencies between the panel members' expressed concerns and their approval recommendations.

This report describes the results of a study conducted by the National Research Center (NRC) for Women & Families, providing the first objective analysis of the key role of FDA advisory committees as part of the FDA approval process. The purpose of this report is to better understand the strengths and weaknesses of the FDA's advisory committee process for FDA's two largest centers, the Center for Drug Evaluation and Research (CDER) and the Center for Devices and Radiological Health (CDRH).

The study analyzes the voting patterns and committee discussions of a random sample of 6 of 16 drug advisory committees and 5 of 18 medical device advisory panels:

Drug Committees
Antiviral Drugs
Arthritis Drugs
Dermatologic and Ophthalmic Drugs
Gastrointestinal Drugs
Pulmonary and Allergy Drugs
Reproductive Health Drugs

Medical Device Panels
Immunology Devices
Microbiology Devices
Obstetrics and Gynecology Devices
Ophthalmic Devices
Radiological Devices

Data for these advisory committees were collected from the FDA Web site, based on transcripts of advisory committee meetings from January 1998 through December 2005. In that time, the 11 randomly selected advisory committees considered 89 prescription drugs and medical devices, including arthritis medications, LASIK devices, erectile dysfunction drugs, and devices to improve the accuracy of mammograms. There were 866 committee member votes.

Findings

As described by FDA officials, its advisory committees meet only to discuss the most controversial or innovative products, or products whose data are not clear-cut. The public might expect, therefore, that many of the drugs and devices reviewed by advisory committees would not be recommended for approval. The data indicate that this is true for some advisory committees, but not others. Overall, the 11 randomly selected advisory committees recommended approval for 79% of the 89 products reviewed between 1998 and 2005. The device advisory panels were even more likely to vote for approval than the drug advisory committees, recommending approval 82% of the time compared to 76% for drugs.

Despite the controversies surrounding many of these products, the votes for or against approval were rarely close. On the contrary, committee members agreed unanimously for 66% of the drugs and 75% of the medical devices that they recommended for approval.

Drug and Device Approval Recommendations

A review of the meeting transcripts indicates that advisory committee members frequently expressed strong concerns about the safety or the efficacy of the drug or device under review. However, those concerns were not necessarily reflected in their recommendations for approval.

There were many examples of committee members who strongly criticized the studies or the medical products under review, and then recommended approval anyway. FDA officials at the meetings almost never expressed concerns about the disconnect between the committee members' explicitly expressed doubts about safety and effectiveness and their votes in favor of approval.

Of the 50 drug committee voting sessions in the study, 38 (76%) recommended approval of the drug. Most of the votes were unanimous, and almost all (93%) of those unanimous votes recommended approval.

Some of the committees were much more likely to recommend approval than others. The percentage of drugs they recommended for approval ranged from 50% for reproductive health drugs to 100% for arthritis drugs. The percentage of individual votes cast to recommend approval ranged from 50% for reproductive drugs to 98% for arthritis drugs.

What happens after the meetings are over? Of the 38 drugs recommended for approval by the drug advisory committees, all were subsequently approved by the FDA except one drug whose application was withdrawn before FDA made its decision. The FDA also approved four (36%) of the 11 drugs that the drug advisory committees voted against, including products that were opposed by almost all the committee members.

Of the 39 device panel voting sessions studied, 32 (82%) recommended approval of the device. Most of the votes were unanimous, with almost all (92%) of those unanimous votes recommending approval.

The percentage of devices that were recommended for approval ranged from 67% for microbiology devices to 88% for ophthalmic devices. The number of panel member votes cast to recommend approval ranged from 57% for microbiology devices to 91% for radiological devices. Three of the five randomly selected devices panels - the Radiological Devices Panel, the Immunology Devices Panel, and the Microbiology Devices Panel - had unanimous support for approval whenever they recommended approval during the eight years of the study.

Almost all (94%) the devices recommended for approval were subsequently approved by the FDA, and close to half (43%) of the devices that were not recommended for approval obtained FDA approval anyway.

Overall, the study found:

Implications and Conclusions

The findings suggest that when the FDA schedules meetings for several of its advisory committees, the outcome is almost certainly going to be FDA approval for the products under review. In most cases the advisory committee will recommend approval, but even products that are not recommended for approval are frequently approved by the FDA. Even lopsided votes against approval apparently do not have much weight, since the FDA subsequently approved many of those products.

Although FDA officials describe the advisory committees as providing diverse perspectives and expertise, the large number of unanimous or nearly unanimous votes suggests that either the data are exceptionally convincing or that the committee members are reluctant to disagree with their colleagues or believe that the FDA wants the advisory committee members to come to consensus.

By combining information from the NRC study with studies of conflicts of interest on FDA advisory committees, it is possible to understand how a few committee members with conflicts of interest can have a disproportionate impact on approval recommendations. NRC's analysis of meeting transcripts indicates that many committee members' votes seem inconsistent with their concerns about the safety or efficacy of the drug or medical device under review. These transcripts clearly illustrate the pressures that committee members describe to conform to their colleagues or to be able to vote 'yes' even if it means changing the wording of the question so that they can do so in good conscience. The report includes examples of committee members directly trying to influence the views or votes of other committee members.

If the FDA is relying on advisory committees to help determine the conditions of approval, one would expect that FDA officials would provide explicit oral instructions about the types of conditions that the FDA is willing to impose, and that the FDA would impose most of the conditions and then enforce them. That is not the case, however.

Committee members frequently recommended unenforceable or vaguely worded conditions of approval and expressed their intention to recommend approval for products that they did not believe were proven safe or effective. Their candor suggests that they would welcome guidance from the FDA officials present, to make sure their recommendations were appropriate. Nevertheless, during committee discussions FDA officials showed remarkably little interest in providing oral guidance regarding the criteria for approval, or the realities of approval conditions to advisory committee members during the eight years of the study. Conditions of approval imposed by the FDA often did not reflect the conditions recommended by the advisory committees. Conditions that were imposed were rarely enforced.

Overall, the findings indicate that committee members, intentionally or unintentionally, move toward a consensus that often seems inconsistent with their differing views or perspectives in making decisions that may have life-or-death consequences for millions of Americans. Voting for approval contingent upon conditions is a popular compromise, but the FDA does not impose most of the specified conditions on the companies when it grants approval. The Committees' tendency toward approval seems to reflect the FDA's goals; in fact, the FDA appears to be even more geared toward approval than the advisory committees. The FDA approved almost all the prescription drugs and devices recommended by the advisory committee, and also frequently approved products that were opposed by the committee members.

Whatever the reasons, many of today's FDA drug and device advisory committees are rubber stamps for approval almost every time they meet. Moreover, even when an overwhelming majority recommend "nonapproval," there is a good chance that FDA officials will approve the product anyway. Approval is even more likely for medical devices than it is for drugs.

Recommendations

If the FDA wants to restore confidence in the FDA, and restore the independence that FDA advisory committees were intended to provide, it is essential that the FDA make changes in the policies and process governing its advisory committees. The following recommendations are based on the assumption that the Congress and the FDA are committed to that end:

1. The FDA should stop granting conflict-of-interest waivers for committee members, except under very restricted conditions.

2. The FDA should provide explicit and specific oral guidance whenever needed during advisory committee meetings regarding appropriate criteria for safety and effectiveness, and appropriate criteria for conditions of approval.

3. The FDA should demand more from advisory committee members, and then be more responsive to their concerns.