Commentary is a regular feature of the Crimson editorial page that provides a forum for opinion from members of the Harvard community. Those interested in contributing pieces should contact the editorial chairman.

QUALITY OF CARE is moving toward center stage among issues in national health care policy. With American health care now consuming nearly 11 percent of our gross national product--far less than in many industrialized countries with equally competent health care systems--purchasers of care have steadily increased the pressure to contain costs.

Cost-cutting has begun to slow the growth in health expenditures, largely by reducing the use of hospitals. But even the most aggressive advocates of cost-containment are beginning to worry. Which cost reductions are safe to make, and which stand a significant chance of harming patients?

Meanwhile, strong evidence has emerged from health services researchers that, even without the pressure of cost-containment, there is reason to be concerned about the quality of at least some segments of the American health care system. Beginning a decade ago, Professor John Wennberg, now of Dartmouth Medical School, has systematically demonstrated levels of variation in the use of health care services that defy rational explanations.

Within the state of Maine, for example, Wennberg finds that, "by the time women reach 70 years of age in one hospital market, the likelihood they have undergone a hysterectomy is 20 percent, while in another market it is 70 percent." In Iowa, the probability of a prostatectomy in a man before the age of 85 years ranges from a low of 15 percent to a high of more than 60 percent in different hospital markets. Wennberg asserts that these enormous differences are based on "practical style," not on true differences in the morbidity of the populations.

Whatever the reasons for such variation, such observations leave a disconcerting sense that quality of care may also be variable.

THE USE OF DIAGNOSIS RELATED GROUPS (DRGs) to set reimbursement levels for hospitals has further stimulated interest in potential uneveness in quality. Under the DRG system, hospitals of similar types in areas of similar economic structure are paid a "flat rate" for care of patients with specific diagnoses. Each patient is made to fit into one of 467 DRG categories (for example, DRG 235 is "Fracture of the Femur"), and the hospital will be paid only a certain fixed fee for the care of that patient if the federal government is picking up the bill.

This fixed payment scheme has led to increased exploration and debate over differences in care patterns among hospitals. For specific DRG's, the actual resources used (days in the hospital, laboratory tests, consulations, etc.) have been found to vary enormously--often several-fold. "High cost" hospitals have been predictably criticized as wasteful, and predictably defended by others on the grounds that their patients are truly sicker in ways ignored by the DRG scheme.

The debate on cost-containment thus shades inexorably into a debate on health care quality. But, it is a debate with more heat than light. Central to the continuing confusion and concern is the lack of any standardized methods in health care to measure the quality of performance of the health care system, or even a component of the system, such as a hospital. With the new marketplace--now purchaser-dominant, not professional-dominant--the demand for systematic information on quality has outgrown the state of the art for measuring quality.

Not that quality measurement has never been done. Hospitals, for example, have for decades undergone the repeated scrutiny of the Joint Commission on Accreditation of Hospitals (JCAH), and hospitals which fail to meet JCAH standards have long been forced by public and private pressures to reform.

On the whole, however, JCAH standards have not been formulated so as to measure either the outcomes of hospital care, nor even the technical processes of care. Instead, audits like those of the JCAH have tended largely to be structural; that is, they focus on assessment of the resources assembled in the hospital to deliver care, not on the care, itself.

The demand has now grown for measures of hospital performance which more thoroughly reflect the actual processes and results of care. The demand for outcome-based measurement of quality has been spurred on by recent investigations showing some apparently systematic differences among hospitals in the mortality rate of certain operations, such as coronary artery bypass surgery.

ON APRIL 17, 1985, the federal government issued final rules and regulations permitting Peer Review Organizations (who are responsible for monitoring utilization and quality of providers of Medicare services) to "disclose to the public PRO interpretations and generalizations on the quality of health care that identify a particular institution." Shortly thereafter, the Health Care Financing Administration released data identifying hospitals with death rates among Medicare patients which were signficantly above or below standard (i.e., rates adjusted for DRGs and a few other measures of population morbidity).

The publication of hospital-specific mortality rates has provoked both interest and alarm. Hospital representatives faced with such potentially inflammatory hard numbers express concern that the consumer may not be able at all to tell the differences between a bad hospital (with high death rates due to poor care) and a good one (with high death rates due to very complex patient cases). The daylight of data has starkly shown the poverty of measurement methods.

Whatever the validity of the rates actually published by HCFA, the genie is unlikely ever to return to the bottle. The combination of cost-control efforts, evidence on extreme variations in practice patterns, the growing strength of purchasers, and the apparent oversupply of both hospital beds and physicians--perhaps 200,000 American hospital beds will be closed between 1980 and 1990--has made the demand for information on quality of care now irresistable.

The crude measures of death rates and simple indices of complications mark only the first salvos in what will undoubtedly by a prolonged battle to construct, test, reconstruct and retest measures of the quality of the components of American health care. Some snake oil will be sold to those impatient for good methods of measurement, but, in the longer run, the serious efforts of health services researchers and managers are more likely than not to yield sound and credible measurement tools for the quality of care. Armed with those tools--but not without them--we may be able to defend the value of a health care system which today is squeezed between the excesses of its past and the market-driven constraints of its future.

Donald M. Berwick is vice president for quality-of-care measurement at the Harvard Community Health Plan, and an associate professor of pediatrics at the Medical School.