The U. S. federal government has already begun providing information on some medical outcomes. Peer review groups already disclose data which shows:

The overall death rate, by hospital, for different types of surgery.

The number of patients who develop postoperative infections.

The average length of hospital stay for a given condition.

How frequently each surgical procedure is performed at a given hospital.

Quality-of-Care assessment will eventually extend to all health care institutions and all providers. Patient outcome will be the key measure. Until now, physician credentials and organizational factors were the key measures. In the past, it was simply assumed that doctors who had the proper credentials did top-quality work. In the future, reviewers will be looking at actual results.

Individual departments—and individual physicians—will also be subject to outcome-based review.2 In one recent study of intensive care units (ICUs), George Washington University researcher William Knaus found that ICU death rates in the worst ICUs were 2.7 times higher than those in the best. Another recent study of physician performance revealed a 3.2 percent overall rate of medical mistakes for inpatient care. Of these medical errors, 23 percent involved unnecessary medical services, 29 percent were life threatening, but did not result in injury or death, 4.0 percent resulted in delayed recovery, 6.0 percent caused a non-life-threatening illness or injury, 6.0 percent caused a life-threatening illness or injury, and 4.4 percent caused death.

The emergence of clinical outcome measures raises two thorny questions:

Are hospital administrators the most appropriate people to administer such a quality-control program or might they be unduely influenced by the increasing pressures of cost consciousness and by narrow concerns for the viability or success of individual departments, institutions, or enterprises? The answer here seems clear: We must include consumers as well as professionals on quality review boards.

Will these new clinical outcome measures be able to account for such subtle variations as genetic, economic, and social status, level of individual responsibility, and what clinicians sometimes call "good and bad protoplasm?" Although much recent progress has been made, the best of our present clinical outcome measures leave much to be desired. The development and application of effective clinical outcome measures must be among our top medical priorities over the next decade.

Deprofessionalization. On the brighter side, some observers believe that the availability of more accurate outcome data will provide health care institutions with a greater degree of flexibility in developing more cost effective treatments. The quality control methods of the past, based on credentials and degrees, led to placing M.D.s in virtually every responsible position. Quality outcome measures should allow institutions to develop more effective ways to perform specialized techniques—e.g. using surgical technicians to perform heart bypass surgery or surgical robots to perform cataract surgery. If an institution is able to document the fact that the patient outcomes from such new procedures are equal or superior to those achieved by M.D.s while costs are substantially lower, it will be very difficult to argue that they should not be allowed to provide these services.

Thanks to Dick Fine, Jesse Hall, and Charles Inlander for their valuable contributions to this column.