For many years, the main alternative to bypass operations was the cardiac stent, a tiny mesh tube placed in the heart's arteries, where it supposedly protects against blockages. Doctors loved stents because they could put off cracking open someone's chest for a while, or forever, and because stents were much cheaper than surgery. Stent manufacturers continued to produce newer and fancier stents, including ones that were soaked in anti-inflammatory and anti-clotting drugs, and doctors continued eagerly prescribing them to their patients.
There was only one problem—no one had ever proved that stents actually helped people live longer. Researchers wanted to perform randomized controlled clinical trials—the gold standard of clinical evidence—but doctors were so convinced that stents worked that they were reluctant to enroll patients. So they kept prescribing them—and stent manufacturers kept selling them—for years without any reason beyond sheer habit.
Then, in March 2007, researchers finally managed to achieve results in a clinical trial, finding that the new, fancier stents were no better than good old-fashioned drug treatment for relieving chest pain. By April, use of such stents had dropped precipitously.
It may seem surprising that doctors were actually shoving a bit of metal into the blood vessels that feed the heart without any good evidence that it worked. But the truth is that medicine is a highly unscientific enterprise. Harvard Medical School cardiologist Thomas Lee estimates that only 30 percent of what doctors do is supported by solid evidence. The other 70 percent of the time, doctors are relying purely on personal experience and the experience of close colleagues, even to the extent of actually ignoring statistics to the contrary. Charles Phelps, a health economist at the University of Rochester, told me that the most frightening words a doctor can say are: "In my experience ..."
In response to broadly untested treatments, a movement called evidence-based medicine that wants doctors to base their medical practice on scientific studies—as opposed to personal experience—is gaining ground among medical researchers. Getting doctors to practice evidence-based medicine is almost certain to improve health care. And the next step—imposing rigorous scientific principles not just on individual treatments but on the field as a whole—would almost certainly lead to a cheaper and more streamlined health care system in general. The thing standing in the way of these reforms, however, is the medical community itself.
A GOOD EXAMPLE OF THE WAY doctors can stymie evidence-based medicine is the story of beta blockers, drugs that slow the heart's rhythm and are used to treat high blood pressure. In 1981, the Journal of the American Medical Association (JAMA) published a study showing that giving patients a particular beta blocker after one heart attack reduced the chances of a second one. The testing was so overwhelmingly conclusive that the researchers were actually able to stop running trials earlier than they'd planned. And no one disputed their results. Beta blockers are very cheap and, considering that the alternative is a heart attack, relatively side effect-free for most patients. You might expect, given all this evidence, that doctors would adopt the procedure immediately once the results were published.
In fact, it took them 25 years. A lot of doctors were concerned that beta blockers could cause side effects in people with other conditions, such as congestive heart failure and emphysema (both diseases are often seen along with heart attacks). But their reasoning may have stemmed from something besides conviction—inertia. "It was easy to find an excuse not to start [patients] on the drug," says Lee.
The effort to convince doctors of the efficacy of beta blockers was by no means passive. A major 1984 textbook mentioned the JAMA study, societies like the American College of Cardiology released guidelines detailing the practice, the government launched a federal hospital performance measurement plan to test whether doctors were prescribing beta blockers, discharge checklists were created that integrated beta blockers, and health plans created incentives for their use. Nonetheless, doctors resisted until very recently. And, of course, that represents an improvement—but it's appalling to think about the number of patients who missed out on this beneficial treatment in the meantime.
It's little mystery why doctors are resistant to developments in medical technology. First off, there's the climate of medical school, in which doctors are basically taught that they are gods when it comes to medical decisions. Nurses, social workers, and pharmacists are important, sure, but it's doctors who are in charge. This attitude makes it difficult for doctors to accept change. They bristle at guidelines and call evidence-based medicine "cookbook medicine" (as if chefs in five-star restaurants never deviated from the Joy of Cooking). And managed care has only made doctors more anxious about the loss of their autonomy. Who is some clerk (or research scientist, by extension) to tell them how to practice medicine—even if the advice makes sense?
Another, more cynical explanation is financial. When a new procedure comes out, everyone wants to be able to do it. Doctors will actually ignore guidelines in order to perform lucrative procedures. For example, some cardiologists are now trying to insert stents in the legs of people with poor circulation, despite the unproven nature of the procedure.
There are less pernicious reasons as well. Some doctors argue that randomized controlled clinical trials are too good—that they are designed in a way that doesn't replicate real-life situations. For example, patients in clinical trials are observed to make sure they take all their medication—but patients in real life, as everyone knows, don't always follow prescriptions to the letter. The difference between the results of a clinical trial and the results a doctor sees in practice is known as the "voltage drop"; and, because everyone knows about it, doctors frequently feel justified discounting studies that don't seem to jibe with what they see in their own clinics.
Finally, doctors suffer from the same overabundance of information that patients do—there is often just too much available data for anyone to make sense of it. It takes only five years for the amount of information a doctor needs to know to double. And, with so many studies coming out, it's sometimes hard to know which ones to take seriously.
BUT PERHAPS THE REASON WHY doctors are so reluctant to accept evidence-based medicine is the fact that many proponents of reform also back turning the scientific process on the doctors themselves, in the form of performance testing and the elimination of unnecessary procedures. In early October, Dartmouth Medical School's Elliot Fisher stood up at a conference at the Institute of Medicine and claimed he had a simple proposal for cutting America's $2 trillion annual health care costs by nearly one third: get doctors to order fewer tests and treatments, and hospitalize patients less often.
Fisher and those who agree with him are trying to perform evidence-based trials on the practices of doctors and hospitals; and, like the scientists who study stents and beta blockers, they are finding results that overturn the received wisdom. Fisher and his Dartmouth colleagues, for example, found that hospitals that bill Medicare more often don't necessarily have higher success rates—in fact, they sometimes have slightly higher mortality rates than hospitals that spend less. Fisher's mentor at Dartmouth, Jack Wennberg, has shown that dying patients don't benefit either from more days spent in hospital or, after a certain point, from more doctor visits.
While the one-third cut in health care costs promised by Fisher seems a bit dubious, given the fact that the poor and uninsured actually have too little health care and not too much, implementing the reforms he and his colleagues champion could go a long way toward improving our health care system. And, most importantly, the oversight they propose would make it more likely that America's doctors would pause before stuffing a bit of metal mesh into someone's arteries for no good reason at all.
Ivan Oransky, M.D., is managing editor, online, at Scientific American and clinical assistant professor of medicine at NYU. This article appeared in the March 26, 2008 issue of the magazine.