By John Reichard, CQ HealthBeat Editor
December 1, 2009 -- One of the more intriguing aspects of U.S. health care is how much it varies by region of the country. Because of evidence that regions with lower levels of treatment in the Medicare program do not have lower quality care, policy analysts see great hope for lowering national spending growth by bringing levels of service down in some parts of the country.
Much of the excitement stems from looking at variations in Medicare spending. However, a new study by the Medicare Payment Advisory Commission (MedPAC) shows that those regional variations aren't as great as analysts thought — if one looks at actual differences in levels of the use of health care services as distinct from spending variations.
"Regional variation in service use is not equivalent to regional variation in Medicare spending," the study emphasized. "The two should not be confused."
The commission's conclusion: "Although service use varies less than spending, the amount of services provided to beneficiaries with similar resource needs still varies substantially." The study says that "we found an approximately 30 percent difference" when comparing areas at the 90th percentile of service use to those at the 10th percentile.
Spending itself isn't an accurate measure of service variation because it reflects other factors such as regional differences in payment rates and differences in the health status of Medicare beneficiaries. Payment levels, for example, can vary because one area has many teaching hospitals, which are paid higher rates, and others don't; or because of regional differences in wages.
In their study, MedPAC analysts adjusted for differences in payment rates and the health status of patients. It found that "regional variation in the use of Medicare services reflects only differences in the volume and intensity of services that beneficiaries with comparable health status receive," the commission said. The "intensity" of service refers to the difference, for example, between a sophisticated type of imaging such as an MRI scan and a simple x-ray.
Looking at the most extreme variation, the study found a nearly two-fold difference between Miami-Dade County, the area with the greatest service use, and the area with the least service use, non-metropolitan Hawaii. The low service use in Hawaii may reflect lower levels of hospital, skilled nursing facility and hospice care, the study said.
The way doctors practice medicine may account for the high levels of service use in Miami, as well as "beneficiaries' predilection for care." But another factor was also at work in Miami-Dade County: per capita spending on durable medical equipment, which includes such items as wheelchairs and hospital beds, and home health care "were both more than seven times the national average." These patterns "raise concerns about fraud and abuse by some providers."
The study also looked at geographic variations in the growth of service use as distinct from levels of service use. "Regions that have high levels of service use are not always the regions with high growth rates," it found. MedPAC researchers also uncovered variations in service use not only between regions, but between parts of the same state and even parts of the same metropolitan region.
Pioneering research at Dartmouth College in Hanover, N.H., found Medicare spending variations wider than the service level variations reported Tuesday by MedPAC.
But Dartmouth researcher Elliott Fisher said in an interview that follow-up work at the college on levels of health care use is consistent with the MedPAC findings.
"The MedPAC study is highly consistent with our own research findings," he said. "They found substantial variations in service use across U.S. regions after taking account of differences in illness and payment adjustments—as did our published research.
"Our research went further, however, and examined the kinds of additional care provided in regions with high service use. In the high spending regions, the additional spending was largely due to greater use of the hospital—for similar patients—greater use of medical specialists, and more frequent diagnostic tests and procedures."
Fisher added that "we also found that the outcomes and quality were just as good—if not better—in the lower spending regions, after accounting for any initial differences in health.
"The implications for health care reform remain unchanged: if all regions could avoid unnecessary hospital stays and treatments as effectively as the lower spending regions, Medicare spending could be reduced by 20 percent or more."