Information, Please

Sounds crazy, but one way to arrest the spiraling cost of health care would be to figure out what treatments actually work

Health-care spending is currently rising at a rate of around eight percent a year, and a large proportion of the health care that is delivered in this country simply drives up costs without improving health or increasing longevity.

Previous efforts to address these problems have neglected an element so basic that its wisdom has largely been ignored—namely, reliable information, the lack of which hampers the efficiency of the health-care market and prevents doctors and hospitals from learning what constitutes high-quality care. It has been more than a decade since the manifesto of "evidence-based medicine" was published in the Journal of the American Medical Association, yet much medical care is still based more on intuition than on science. Doctors don't know, for example, whether regular mammograms for women in their forties save lives, and they ignore evidence that Celebrex, a widely prescribed and much ballyhooed prescription painkiller, is no more effective than ibuprofen, which costs one tenth as much. When common medical practices are put to the test, as only a small fraction have been, many turn out to be either ineffective or excessively risky—at least for some patients. In one well-known example a clinical trial that ended in 2002 showed that hormone-replacement medication, which generated $1.2 billion in sales in 2000 and which postmenopausal women have been taking for more than thirty years, does not lower the risk of heart attack, as previously believed; in fact, it appears to raise it.

As a result of such uncertainty health-care consumers—patients, employers, and insurers—cannot choose the best care at the most reasonable price. As health-care buyers, insurers—whose costs keep rising—have little data with which to judge the effectiveness of all the hospitalizations, tests, office visits, and procedures they are paying for, and thus no rational basis for adjusting reimbursement according to the quality of care.

More information would increase both the quality of health care and the efficiency of the system—if only someone would gather the data. Most insurers, with the exception of Medicare and a couple of the largest private companies, don't compile the records that might help them improve health-care quality, largely because they lack the expertise and manpower to do so. The National Institutes of Health, which have traditionally focused on finding cures rather than on improving existing medical practices, set aside less than one thirtieth of their combined budgets—amounting to $800 million annually—for such research. The Agency for Health Care Policy and Research—the one government agency that was devoted to gathering outcomes data—saw its budget slashed in 1995 after lobbying by a determined and influential group of doctors who didn't like the results of the agency's analysis.

Here's what's needed: a Clinical Sciences Institute, which would serve as a clearinghouse for outcomes data. The institute would need the political insulation of the Federal Reserve Board; it could be funded by the federal government either alone or in concert with private insurers. Its first task would be to gather information on patient outcomes from hospitals and doctors in order to determine who is delivering the safest and most effective care. It would also pull together all available evidence on current medical practices. If it couldn't determine the effectiveness of a treatment or a test using existing data, the institute would fund a clinical trial to get the answer. It should initiate clinical trials like the one that found that hormone-replacement therapy increases the risk of heart attack, and head-to-head comparisons of drugs like Celebrex and Vioxx with their cheaper substitutes.

With real evidence in hand, doctors could begin to base standards of care more on science than on intuition. Insurers would have a rational basis for deciding which drugs and procedures to cover—and they could adjust reimbursement to hospitals and doctors according to who did the best job. Patients would know which doctors and hospitals to choose. The danger, of course, is that weighing costs against the effectiveness of treatments and drugs would raise the specter of medical rationing. Polls consistently show that Americans will reject any plan that threatens to deny them potentially lifesaving treatment for reasons of cost. But patients might feel differently about all the medical care they think they want—and deserve—if they had access to clear and accurate information showing that the most expensive drugs weren't necessarily the best and that many treatments and tests posed unseen risks without many proven benefits.

Put into practice, the information provided by the Clinical Sciences Institute would improve the quality of care, and might even save billions of dollars that are now wasted on unnecessary care. Here's just one example: A recent study found that carotid endarterectomy, a $15,000 surgery performed on about 150,000 patients a year to prevent stroke, turns out to pose a greater risk of triggering a stroke than preventing one in two thirds of patients. If surgeons stopped performing this procedure on the wrong patients, limiting it to the one third it actually benefits, they could save the U.S. health-care system $1.5 billion a year.