Skip to main content

Variation in Health Spending Again

By February 28, 2012Commentary

Researchers at the Center for Studying Health System Change’s National Institute for Health Care Reform have fired the latest missive in the discussion on geographic health spending variation.  They issued a report based on studying non-elderly unionized auto workers across the country.  These workers have common benefits and the report analyzed claims in 2009 from over one million enrollees and dependents.   (CSHSC Report)   In all nineteen different regions were compared, with average per enrollee spending ranging from $4500 in Buffalo, New York, to $9000 in Lake County, Illinois.  This is obviously a very wide range.  The basic analysis of spending looked at quantity of services per person in each region as well as the relative prices for services.  The quantity of services was analyzed in terms of age and sex difference in a region and health status differences.  Price differences were analyzed for cost-of-doing business variation across geographies.

Overall, about two-thirds of variation was explained by differences in quantity of services and one-third by price differences.  As might be expected, high cost areas have both more quantity of service and higher prices and low cost ones have the opposite phenomenon.  Differences in health status and age/sex mix account for all but about 20% of the quantity of services variation.  In other words, higher cost regions have people who are sicker, with more disease, and therefore, need more services.  But the 20% represents variation that may not be justified by health care needs.   In a regard to prices, differences in the cost of doing business in various localities explains almost none of the variation, it is purely higher or lower prices.  Physician prices tend to vary little, but hospital prices have wide variation.  To verify that health status differences were real and not a reflection of variations in coding aggressiveness, the researchers compared  health status with broader measures of community health, finding that poor community health was associated with regions where auto workers had worse health status.  Interestingly, while hospital prices account for most of the price variation, such differences do not appear to be related to hospital quality or market concentration.  Lowering hospital prices and improving the health of certain communities appear to be one approach for reducing spending.

Leave a comment