The Institute of Medicine's recent study of the prevalence of error in medicine is discouraging but not surprising. Those of us who have been physicians for a long time can certainly concur with the study's main conclusions: that simply too many injuries and deaths are caused by a health care system that is supposed to make us better, and that the annual deaths due to medical intervention are a national scandal.

The basic problem lies in the combination of the dictum that "to err is human" and the extremely complicated nature of medical interventions. One study concluded that the patients in an average intensive care unit (ICU) received an average of 178 "medical activities" each day, and that the error rate of 1.7 errors per day meant that the ICU personnel were functioning at a 99 percent level of perfection.

That sounds pretty good, but unfortunately a failure rate of even one percent is higher than that tolerated in many industries. Efficiency expert W. E. Demming points out that in aviation, a 99.9 percent perfection ratewould allow seven unsafe plane landings per day at LaGuardia.

The problem of error is intensified by the general culture of medicine, a culture that demands perfection but in which mistakes are not only inevitable but also inadmissible, where "near-misses" are hidden and not analyzed, where errors are associated not with quality improvement but with legal punishment.

Here's the crux of the problem: The threat of malpractice litigation almost eliminates the opportunity to disclose error, learn from mistakes and receive support from colleagues and even patients when inevitable errors occur.

When I was a young surgeon, it was our custom to convene weekly morbidity and mortality conferences, which were some of the best training opportunities afforded me and my surgical colleagues. As in similar conferences today, we discussed deaths and critically ill patients, but we also voluntarily discussed errors. We labeled each mistake as an error in judgment, in technique, a procedural error, etc. In other words, we had institutionalized the process of dealing with errors and were thus able to act to reduce them.

But sometime in the 1940s, the hospital lawyers warned us to stop keeping records of such conferences, because the admission of even a small error could lead to a lawsuit. And then the lawyers advised us not even to attend regularly scheduled meetings that could be the subject of a legal inquiry.

Voluntary reporting of small errors before they become large errors is the most important element in improving patient safety. We will not be able to deal with error and institutionalize it unless we modify the current medical malpractice system, which doesn't work for either the patient or physician.

True, the Institute of Medicine recommends protecting voluntary reporting of errors from the process of legal discovery, but I'm afraid this mildly orded recommendation will get left by the wayside because of its political problems. Any change in the malpractice mess must win legislative approval, and it is almost impossible to get state and federal legislatures, packed as they are with lawyers, to do anything that might limit the legal profession--unless we the people hold their feet to the fire.

And we the people must also then realize that the measures we need to take to reduce error will have side effects. One side effect of needed medical malpractice changes is that we may need to forgo our right to file a lawsuit against any health care professional whenever we want.

Error prevention has other side effects. For example, those who are concerned by errors caused by fatigued interns and residents working long hours without sleep will applaud the recent decision by the National Labor Relations Board to allow doctors-in-training to unionize. Health care workers often complain rightfully about the stress they feel in their jobs. But we need to resist efforts to remove stress altogether. Goldilocks had it right when it comes to stress; it is important to get it "just right," not too much, but not too little either.

Errors tend to happen in times of panic or boredom. The unionization of interns and residents might lead to shorter shifts, and if patient care is not to be lessened, that means we'll need more interns and residents. But it also means we'll end up with more doctors in medical specialties that already are overcrowded, driving up the cost of health care.

That's probably the biggest side effect: the short-term cost of reducing medical error, even if the long-term benefits to society will outweigh the direct costs. We've learned that safety costs money, often a lot of money, whether in education and retraining in safety procedures or in redesigning equipment so that standardization will eliminate operator confusion, or in staffing error-reporting agencies, or simply in setting up the redundancy in equipment and procedures that has done so much to make airplanes and automobiles safer.

We will have some hefty costs and some tough side effects from our efforts to reduce error in medicine, but the cure will be worth it.

The writer was surgeon general from 1981 to 1989.