What Drives the High Cost of Health Care in the United States?

A recent post by Jay Nawrocki discussed a CMS study of the variations in per capita spending for health care in the United States. Variations in spending were attributed to gender, age, income and relative prosperity of geographic area where a patient lived. What drives these differences?

Other studies of health care spending have examined geographic variation within the United States in a bit more detail. A wide disparity among regions of the country in per capita spending has been well documented. For example,  the Dartmouth Atlas Health Care Project has repeatedly found that geographic variation in health care spending is largely explained by local capacity; that is, the number of physicians, hospital beds, MRI scanners per capita has more bearing on the number of doctor visits, hospitalization days, and scans per capita than demographics or the severity of patients’ illnesses.  This phenomenon is called supply-sensitive care.

When regional differences are examined, at the end of life, Medicare beneficiaries in “high-spending areas”  spend more time in the hospital, see doctors more often,  and undergo more procedures than those in low-spending areas. Earlier this year, Dartmouth released updates to the previous reports.  In  2007, chronically ill Medicare beneficiaries spent fewer days in the hospital, and they were less likely to die in the hospital than in 2003. Still, wide variation remained. Medicare beneficiaries in or near New York City were far more likely to die in the hospital — 45.8 percent in Manhattan, compared to 19.0 percent in Fort Lauderdale or 12.0 percent in Minot, North Dakota.

Among chronically ill beneficiaries patients who were hospitalized, a stay in intensive care was much more likely in areas with greater hospital capacity — around 30  percent in New Brunswick, New Jersey, Los Angeles and Miami, compared to Portland, Oregon (11.2 percent).  And in the last six months of life, beneficiaries in Manhattan spent nearly 21 days in the hospital, and nearby residents of New Jersey or Long Island spent about 18 days, while patients in the Midwest and the West spent about seven. Patients in high-cost areas also were more likely to see ten or more doctors during the last six months of life.

To some extent, the increased utilization in areas of increased capacity may be attributable to the understandable inclination to use available resources to care for people who are sick.  However, there was no evidence that patients  who received more services had better health outcomes.  Consumers in high-cost and low-cost areas expressed similar levels of satisfaction, so it is unlikely that beneficiaries oin lower-utilization areas believe they lack sufficient access to services.

In fact, a recent Government Accountability Office (GAO) study shows that few Medicare beneficiaries have difficulty obtaining access to needed services. GAO found that beneficiaries’ utilization of services was greater and  continued to grow at a higher rate than average in some locations, most notably in metropolitan areas east of the Mississippi river. In this GAO study, the potentially overserved areas were compared with areas with similar populations with respect to age,  racial composition, number of hospital beds, physicians and specialists per 1,000 population.    GAO found that the differences in spending and utilization arose from differences in patterns of medical practice between  physicians in the potentially overserved areas and others.  Certain services were especially susceptible to overuse, according to this study. Advanced imaging services were 16.1 percent more prevalent, and electrocardiograms, 30.1 percent more frequent, per 1,000 beneficiaries in the potentially overserved areas.

Dartmouth research also compared spending  and growth in spending among and within different regions. The national rate of growth in healthcare spending between 1992 and 2006 was 3.5 percent per year. But the 5.0 percent annual growth in Miami was double the 2.4 percent annual growth in San Francisco. The Dartmouth Atlas estimates that if spending nationwide grew at the rate in San Francisco,  Medicare would save $1.42 trillion between 2008 and 2023. So, how is health care in San Francisco different from that in Manhattan?

It’s not the availability of technology. There is no reason to believe that San Francisco lacks the medical equipment or technical sophistication necessary to provide quality care for patients. And the data compared are fee-for-service payments.  Apparently, differences in the decision making processes of both the organizations and the individual physicians are critical.And the problem is, in part, learning how those processes differ among regions and different tpes of hospitals, i.e., government-owned and operated, not-for-profit and for-profit hospitals.