Vast geographic differences found in drug spending under Medicare

Published in the Online First June 9 issue of the New England Journal of Medicine, and the first to explore regional drug spending under Medicare, the study also found that higher spending on drugs was not balanced by lower spending on other medical care services such as hospitalizations and visits to the doctor's office.

"As current health care reform legislation seeks to address inefficiencies in Medicare spending to get costs under control, it is vitally important to look at how spending differs regionally," said the study's lead author, Yuting Zhang, Ph.D., assistant professor of health economics at the University of Pittsburgh Graduate School of Public Health. "One of the key questions is whether Medicare patients who spend more on drugs to control chronic conditions have fewer physician visits and hospitalizations. Without examining drug expenditures, it is impossible to know whether spending in one area may substitute for spending in another."

In the study, Dr. Zhang and her colleagues, Joseph P. Newhouse, Ph.D., and Katherine Baicker, Ph.D., Harvard School of Public Health, calculated average drug and non-drug medical spending in 2007 among 306 hospital-referral regions across the U.S. as defined by previous studies. They controlled for geographic differences in health care prices, population demographics, insurance status and overall health, and they developed maps to illustrate annual drug and total medical spending by Medicare beneficiaries across each hospital-referral region.

The authors found that drug spending accounted for more than 20 percent of total medical spending, but varied substantially. For example, the highest region for drug spending under Medicare was Manhattan, N.Y. ($2,973 annually per beneficiary) and the lowest as Hudson, Fla. ($1,854 annually per beneficiary). Non-drug medical spending also varied widely and was twice as high in the highest-spending regions compared to the lowest.

The authors also found that variations in drug spending were only weakly associated with variations in non-drug medical spending.

"Spending more on drugs didn't clearly result in less spending on other medical services," said Dr. Zhang. "Although there was a weak correlation between the two types of spending, high spending in one area was not offset by low spending in the other. This data gives us valuable insight into the use of health care resources and may help guide public policy related to health care reform."

The study was funded by the University of Pittsburgh and grants from the National Institute of Mental Health and the Agency for Healthcare Research and Quality.