The most striking fact about health care spending in the United States is its rapid growth over the past 25 years. In 1980, 8.7 percent of our GDP went for health care. Today, the figure is over 15 percent. Moreover, given the strong growth in GDP over this period, health care accounts for a bigger slice of a much bigger pie.

This rapid rise is mostly due to what I call activist medicine. Activist medicine means undertaking procedures that have a low probability of affecting the final outcome. Whether or not activist medicine is cost-effective is a statistical question, which can only be answered by studying the results of standard medical practices, such as a routine colonoscopy.

For example, I suspect that when I went to a doctor for lower back pain and was sent for an MRI, this was not justified. Regardless of what the MRI might show, what I needed to do was avoid lifting anything heavy for a while.

At the time, I had no quarrel with getting an MRI. Quite the contrary. If all the doctor did was send you home to rest your back, you would be disappointed. You'd be ashamed to tell people you even went to the doctor. On the other hand, being sent for an MRI means you can hold your head up. Me have Real Injury! Doctor say me have to go for MRI!

On another occasion, I was much more inclined to resist a precautionary procedure. Several years ago, I was sent for an intravenous pylogram and a cystoscopy when my doctor found microhematuria (blood in a urine sample that is visible only under a microscope). Because more than 10 percent of the healthy population has microhematuria, and because I had no other symptoms or risk factors, I questioned the need for these procedures, but my doctor vehemently insisted upon them. In fact, my views would now seem to be supported, at least by this article.

More generally, an important study by John Wennberg, Elliot Fisher, and Jonathan Skinner found variations as high as 100 percent in Medicare spending on patients with apparently similar conditions in different regions of the country. Either one region is doing too much, another region is doing too little, or both.

As David Cutler points out in Your Money or Your Life, the U.S. health care system could perform better by doing more of what is cost effective and less of what is not. The proposal below is an idea to help steer our health care system in the right direction. It elaborates on Fuchs' idea, quoted above. See also Fuchs and Alan Garber.

Medical Guidelines Commission

Cutler argues that we need to reward health care providers on the basis of effectiveness rather than on quantity of services. Ultimately, that seems like a good idea. However, the problem is that the effectiveness of activist medicine can only be judged in a statistical sense. To know whether the procedures were appropriate, you have to know the distribution of outcomes that occurred when those procedures were used in many similar cases. The issue is not whether patient A obtained relief, but whether the procedures followed in diagnosing and treating patient A were those that best balanced the probabilities of safe and successful treatment with the cost.

My proposal is that either a consortium of health insurance companies or the government establish a formal mechanism for evaluating the benefits of medical practices. Call this a Medical Guidelines Commission (henceforth the Commission).

The Commission would consist of, say, seven members drawn from the disciplines of medical research, medical practice, statistics and actuarial science, and economics. It would have a staff with similar expertise to evaluate research and to oversee grants to stimulate research where none is available.

The Commission would have two main functions:

• Collect and summarize research about the effectiveness of various medical practices. For those of us willing to work with doctors to make our own decisions, this information would be sufficient.

• Issue guidelines that could be taken as recommendations for best practices. These guidelines would serve those people who would be intimidated by statistical research and prefer to be told what to do. The guidelines might also serve as industry-wide benchmarks.

Guidelines would not be carved in stone. Patients and doctors would be free to choose treatments that differ from the guidelines. Moreover, the guidelines themselves would change over time, because medicine is a constantly-evolving art.

My hope would be that the guidelines would provide doctors with some protection against malpractice suits. A doctor who follows the guidelines would have a persuasive argument in court. The guidelines might reduce the pressure to practice purely defensive medicine.

The Commission's research could serve to replace the "efficacy" rules of the Food and Drug Administration. Today, the FDA requires new drugs to be tested not just for safety, but for how well they work (efficacy). Instead, we could return to the regime that existed prior to 1962, in which the FDA would only test drugs for safety. The Commission would report on the efficacy of drugs and choose either to include them or not include them in its guidelines. Even so, patients and doctors would be free to make choices outside the guidelines. Along these lines, Daniel Klein and Alex Tabarrok have written, "The experience with off-label prescribing and the experience of pre-1962 America suggest that initial efficacy requirements may do more harm than good. Dropping efficacy requirements is a proposal that should be taken seriously."

The guidelines might influence insurance company reimbursement policies. For example, an insurance company might reimburse most or all of the cost of procedures under the guidelines, but only 20 percent of the cost of procedures that go beyond the guidelines. Of course, the specific reimbursement rules would be negotiated between insurers and their customers, not set by the Commission.

Information and Decisions

The United States has the best health care resources in the world. Using them wisely will not be achieved by magic. Because many specialist referrals, diagnostic procedures, and referrals are discretionary, it is important to look at a lot of cases in order to assess the frequency with which outcomes are affected.

This information will be necessary regardless of which direction the United States takes with health care reform. However, I believe that the best approach to utilizing information that connects procedures to outcomes would be "bottom up," with patients and doctors making better decisions.

I see the Commission as an information-supplier rather than a decision-maker. We need to centralize the collection of data on the outcomes of typical medical procedures for many patients. I also believe that the Commission could help many people by translating the information it collects into guidelines. However, I am skeptical that centralizing more of the decision-making process would be constructive.