More Health Care Is Not Better

July 16, 2008

healthcarecosts.jpgSurprising results from the Datmouth Atlas of Health Care

Commentary by Zackary Berger MD PhD, PGY-3, Health Care Policy Section Editor

What explains the large variation in health care costs across the country? You would expect that the regions with the highest health care expenditures have the sickest patients, or have the highest prevalence of chronic health conditions.

But you would be wrong. Health care expenditures are highest where health care supply is the highest. As the number of hospital beds, doctors, and medical technologies (“supply-limited health care”) increases, the use of these goods increases as well. “A built hospital bed is a filled hospital bed,” as health services researcher Milton Roemer said.

Both these answers are provided by the Dartmouth Atlas of Health Care, a project of the Darmouth Institute for Health Policy and Clinical Practice. The Atlas is a collection of tables, organized by geographic region and hospital, including information about rates of procedures, Medicare costs, and types of procedures which generated these costs. A nearly infinite number of tables can be generated by a visit to the Atlas’s web site, but the information in the Health Care Atlas 2008 is more than enough to start with.

To give one example, this table titled Medicare Spending Report presents the hospital care intensity (HCI) index for various New York hospitals. The index is a measure of health care spending for Medicare patients during the last two years of life, taking into account the number of days spent in the hospital and the intensity of physician services provided in the hospital. Notice that the highest HCI (for NYU Medical Center) is greater than the lowest HCI (Kings County Hospital) by more than a factor of two.

I mentioned above that sick patients, or high prevalence of chronic health conditions, do not explain the significant variance in HCI. What about quality of hospitals? Perhaps quality care costs more. However, even within our own campus, Bellevue has a significantly lower HCI than Tisch. It doesn’t necessarily follow that the care at Tisch is better than Bellevue. (It appears that the VA hospitals are not included in the Darmouth databases.)

The assumption that cost means quality is disproven by another recent report from the Darmouth Institute, Tracking the Care of Patients with Severe Chronic Illness. Among the findings of this report was a comparison of the cost of care given at five “top hospitals” (as ranked by US News and World Report), including such perennials as Johns Hopkins University Hospital, St. Mary’s Hospital of the Mayo Clinic, the Cleveland Clinic, Massachusetts General Hospital, and the University of California hospital in Los Angeles. The cost of care, even among this group of ‘best” hospitals, varied by a factor of two between the highest and the lowest spenders. The Mayo Clinic and the Cleveland Clinic, concludes the report, spend less per capita than their peers among academic medical centers, while still receiving high marks on accepted measures of health-care quality.

To sum up the Dartmouth findings in the center’s own words: “The patterns among academic medical centers are consistent[…]: the intensity of care for supply-sensitive services is driven by the subliminal influence of available resources on clinical decision-making.” The proper reaction to these phenomena is not blame – how could the administrators of hospital X squander so much money! – but appreciation of the perverted incentives created by our health-care system, which pays for procedures and tests but not (or not as much) for diagnosis and therapy.