Looking for some trouble
The Food and Drug Administration is always playing catch-up. During the past 12 months, more than 3.5 billion drug prescriptions were filled in the United States--prescriptions for approved drugs that, despite testing, may still have dangerous side effects for some people. To detect dangers, the agency largely relies on physicians, consumers, and drug companies to send in reports of problems. And this just isn't enough firepower.
Syed Ahmad, a scientist in the agency's Office of Drug Safety, notes these reports may represent as few as 1 out of every 100 "adverse events" like bad reactions, hospitalizations, and deaths. What's more, says Sandra Kweder, deputy director of FDA's Office of New Drugs, the agency has only about 75 scientists to review these reports.
Real world. This becomes a problem because no drug is completely safe, and there is no foolproof pre-marketing testing. Trials, for sound scientific reasons, are done on people who don't have other diseases or take other medications. But in the real world, that drug will more likely be taken by people who have several health problems, take several medications, or are older or younger than the trial groups. So new problems often crop up.
To spot them, the FDA relies on a system called MedWatch. If something bad happens to a patient after taking a drug, a doctor or patient can voluntarily fill out a form and send it in to MedWatch. Drug companies are required to do so. The reports go into a database, and then scientists look for patterns. Did the bad event happen soon after taking the drug? Did the problem stop when the drug was stopped? A "Yes" answer suggests a problem and can prompt the FDA to strengthen the warnings on a drug or recall it entirely, as it did recently with the statin Baycol, which was causing muscle wasting and weakness.
But that's not good enough, says Sidney Wolfe, director of the health research group at Public Citizen: "If only 1 in 10 adverse events gets reported, then it takes too long to trigger action." The low rate means it takes a long time to gather enough reports to show a clear pattern. One reason for the small response: Busy doctors have little incentive to fill out complex forms. "Countries in Europe often do better, because they use local agencies to push local doctors," says Rajesh Balkrishnan, a drug-outcomes researcher at the University of Texas School of Public Health. Incentives in the United States would be a good thing, says Wolfe: "If we got twice as many reports, then the pattern would appear twice as quickly." And fewer people would get hurt. -Josh Fischman
With Amanda Spake
This story appears in the December 13, 2004 print edition of U.S. News & World Report.
advertisement
