Taking back the FDA

In effect, the user fee act put the FDA on the payroll of the industry it regulates. Last year, the fees came to about $300 million, which the companies recoup many times over by getting their drugs to market faster.

But while it's a small investment for drug companies, it's a lot of money for the agency, and it has drastically changed the way it operates -- creating a disproportionate emphasis on approving brand-name drugs in a hurry. Consequently, the part of the agency that reviews new drugs gets more than half its money from user fees, and it has grown rapidly. Meanwhile, the parts that monitor safety, ensure manufacturing standards, and check ads for accuracy have languished or even shrunk.

Most tellingly, the office that approves generic drugs is so small that approval time for generics is twice as long as for brand-name drugs. There is now a backlog of more than 800 generics. That delay is worth billions of dollars to the drug companies whose high prices depend on not having generic competition.

As part of the emphasis on speed, the FDA often approves brand-name drugs on the basis of less evidence than in the past. In these cases, approval may be contingent on companies conducting further safety studies after the drugs are on the market. But the companies usually don't honor that commitment. Of the roughly 1,200 such studies outstanding -- some for years -- over 70 percent haven't been started.

The FDA is strangely silent about this inexcusable dereliction. When questioned, it weakly protests that it doesn't have the authority to compel the research. In fact, it has enormous leverage, since it can withdraw drugs from the market.

The FDA also refuses to release unfavorable research results in its possession without the sponsoring company's permission. Here again, it contends not to have the authority to do so, but providing evidence of side-effects or negative results would seem to be an integral part of its job. It's no wonder that serious safety concerns about drugs such as Vioxx, Paxil, and Zyprexa have emerged very late in the day -- years after they were in widespread use.

The agency's coziness with industry is underscored by the composition of its 18 advisory committees -- outside experts who help evaluate drugs. Incredibly, many of these advisers work as consultants for drug companies. Although they are supposed to recuse themselves if there is a direct conflict of interest, the FDA regularly grants exemptions from that requirement. Of the six members of the advisory committee that in 1999 recommended approving Vioxx -- the arthritis drug pulled from the market in 2004 because it caused heart attacks -- four had received waivers from the conflict-of-interest rule.

The FDA now behaves as though the pharmaceutical industry is its user, not the public. Fortunately, the user fee law is subject to renewal every five years, and this is one of those years. Congress should let the law die this time around and substitute its own support -- which ought to be increased. Other reforms recently proposed, such as administratively separating drug approval from safety surveillance, will not mean much as long as this law is in effect.

At $300 million to $400 million a year, the equivalent of about a day in Iraq, Congress can easily afford to buy this vital agency back for the public, and it should.

Dr. Marcia Angell, a senior lecturer at Harvard Medical School, is a guest columnist.