My new firm, Precision Health Economics (PHE), also produced a report for the IOM’s project on regional variation in health care spending, utilization and quality. PHE’s goal was to synthesize the results across the separte investigations into the Medicare, Medicaid and commercially insured populations. Their findings are the following.
In synthesizing the population-specific studies of geographic variation in health care spending, utilization, and quality, this report finds that:
- Across the populations, variation in spending, utilization, and quality is high. Coefficients of variation on prescription drug utilization, for example, range from 0.13 to 0.62, while standard deviations of quality outcomes are as high as 0.52. While some populations vary more than others, geographic variation exists in all populations.
- Without adjusting for any predictors, coefficients of variation on input price adjusted spending range from 0.12 to 0.59. Common predictors of variation like health status and health care market characteristics reduce this variation across the populations. However, the majority of that variation remains even after adjusting for those predictors.
- Regionally-based policies would likely have a lesser impact on geographic variation if they are implemented at more aggregated levels of geography. About half of the variation in health care spending and utilization is occurring at sub-regional levels of geography. This result is particularly true for spending in the commercially insured population.
In analyzing how a measure of total (all-population) spending varies geographically and how well it predicts health care quality in the Medicare population, this report finds:
- Total spending varies geographically at a similar magnitude as the population-specific studies. It is generally more variable than commercial spending but less variable that Medicaid spending.
- Adjusting for important predictors of health care variation reduces the coefficient of variation of the measure of total spending. In particular, adjusting for health status redues variation by 25%, and adjusting for health status as well as market factors reduces variation by 30% overall. These reductions in variation are of a similar magnitude as those in the population-specific studies.
- Total spending is a somewhat worse predictor of Medicare quality than is Medicare spending.
Source:
- Precision Health Economics. Geographic Variation in Health Care Spending and Promotion of High-Value Care: Final Report. February 4, 2013.