A Drug on the Market.(The Talk of the Town)(Avandia)

Last month, a fierce and costly battle erupted over the diabetes drug Avandia after an article in The New England Journal of Medicine suggested that the drug raised the risk of a heart attack by forty-three per cent. In the publicity storm that ensued, the number of new Avandia prescriptions shrank by twenty-one per cent, and investors lopped more than twelve billion dollars off the market capitalization of the drug's maker, Glaxo-Smith-Kline. Then came a furious backlash. The article's lead author, Steven Nissen, was accused of being a publicity-seeking crusader with a conflict of interest (Nissen had previously received research support from the maker of one of Avandia's competitors). GlaxoSmithKline dismissed Nissen's work, which was a meta-analysis of forty-two other studies, and published interim results from its own long-term study of Avandia's safety, which it claimed proved the drug to be no more dangerous than its competitors. There were complaints about the "tabloid" hype that journals attach to their stories, and the British medical journal The Lancet said that "alarmist headlines and confident declarations help nobody."

This kind of brouhaha, with volleys of personal attacks and fights for the biggest headline, doesn't look much like science. But it's all too typical of the way we measure the safety and efficacy of drugs. The U.S. has no rational system for "post-market surveillance"--the evaluation of drugs after they've been approved. Instead, oversight is left to a motley collection of altruists, academics, lawyers, self-publicists, and drug companies, who make their own arbitrary decisions about which drugs to study, how to evaluate them, and what risks to look for. Somehow, the truth is expected to rise to the surface from among all these competing interests and random decisions.

One might expect the Food and Drug Administration to bring order and rationality to this system. But the way the F.D.A. is configured and run prevents it from doing so. Before a drug has been approved, the F.D.A. has tremendous leverage over pharmaceutical companies, and can require them to do the studies that it deems necessary. As soon as the agency actually approves a drug for sale, though, its authority is markedly diminished. The agency can recommend that the manufacturer of a drug already on the market conduct studies, but it can't, with a few exceptions, force the company to do so. Furthermore, it can't fine companies that don't follow its recommendations, and it can't limit their advertising or sanction them in any real way. As a result, most post-market studies promised by drug companies have never been started, and, of those which have, nearly three-quarters remain incomplete.

Instead of relying on drugmakers to test their own products, the agency could, in principle, run or commission its own trials. But it lacks both the money and the infrastructure to do so. While there is an office devoted to "surveillance and epidemiology," which is theoretically responsible ...