) pulled its Vioxx off the market after a study confirmed suspicions that it was linked to an increased risk of heart attack and stroke. Then, after rival Pfizer (PFE
) announced its Cox 2 Bextra drug also showed a higher risk of heart attack and stroke in patients who had just undergone a certain kind of heart surgery, it said it planned to study its other painkiller, Celebrex, to determine if it could actually prevent heart attacks.
If the real risks of the big-selling Cox 2 drugs are still unknown, the controversy has shown that the Food & Drug Administration's current system for monitoring the long-term effects of prescription medications is inadequate. Safety studies that take place before a drug is approved typically last only weeks or months, and many new drugs that treat chronic conditions are used for years, or even decades. And the system for spotting problems after a drug is approved is based on voluntary, and sometimes spotty, reporting by physicians.
"COMPLETE DISCONNECT." What consumers need is a new system that aggressively looks for potential problems with drugs, especially the difficult-to-spot side effects that may surface only after several years of use. "The whole system is set up to identify problems that occur soon after commencing treatment," says Thomas J. Moore, health-policy analyst at George Washington University. So when it comes to the way drugs are studied, vs. how they're actually used, "there's a complete disconnect," he says.
Certainly there have been some recent nasty surprises. While consumer advocates and some medical researchers had already been suspicious about a link between some Cox 2 drugs and cardiovascular problems, Merck says the study that led to Vioxx's withdrawal showed a rise in the risk of heart attack or stroke only after 18 months. In 2002, a government-led study showed that women who were taking hormone-replacement therapy had a higher risk of breast cancer -- but that the increased risk emerged not until four years after the study was begun.
The problem with the current system is that it relies primarily on physicians to make the connection between a drug and its adverse effects. Doctors file reports with the FDA or the drugmaker when they observe a problem they believe is linked to a drug being used by their patient. But because the system is voluntary, and physicians simply don't have the time, most problems go unreported. As a result, it may take longer than necessary for the FDA to see a pattern of problems.
Just as important, if a side effect emerges only after many months, or even years, of exposure to a medicine, doctors may not realize the problem stemmed from a particular drug. After all, if a middle-aged man has a heart attack two years after starting a particular medicine, what are the chances his doctor will assume the events are linked and report it to the FDA, particularly if the heart-attack risk isn't mentioned on the medicine's label?
BETTER CHECK-UPS. One way to fix the system is to have a review mechanism in place for reevaluating a drug a few years after it's approved. Dr. Jerry Avorn, associate professor of medicine at Harvard Medical School, believes this should be done within three years after a drug is launched. At that time, physicians and others could air and discuss questions about safety and possible problems.
If a drug's safety looks questionable, the FDA can use information technology to get a preliminary handle on the problem. Groups like Kaiser Permanente, the nonprofit health-care system, or the Veterans Administration, have extensive and detailed databases on their members' experiences with a variety of drugs. Such databases can be used to cull data, controlling for factors like age, sex, and general health, to evaluate whether there's a possible safety problem with a drug.
In the case of Vioxx, Kaiser Permanente did an FDA-funded study released in August that looked at records for 1.4 million of its members. That analysis, which cost less than $100,000, found Vioxx users were at higher risk for heart attacks. While such studies aren't definitive because they lack the controls of formal clinical trials, they "lead you in the direction of the right answer," says Dr. Sharon L. Levine, associate executive director at Kaiser.
In cases where such studies or other data do ring alarm bells, then large, formal, long-term trials should be initiated, similar to the 2002 hormone-replacement study. Such trials, while expensive, are still the best way to get definitive answers about the true risks and benefits of prescription drugs. Barrett is BusinessWeek's Philadelphia bureau chief