The Sources of Cost Difference in Health Insurance Plans: A Decomposition Analysis

Abstract

The goal of this paper is to develop methods to identify and quantify the importance of the factors contributing to the wide differences in health care expenditures among employer-provided health plans. Our hope is that understanding the reasons for cost differences will direct more focused attention to the ways that costs can be controlled. This work is intended as an important first step in that direction. We consider eight plans which vary in average expenditure for individuals filing claims from a low of $1,645 to a high ov $2,484. We present a statistically consistent method for decomposing the cost differences across plans into component parts due to demographic characteristics of plan participants, the mix of diagnoses for which participants are treated, and the cost of treatment given particular diagnoses. The goal is to quantify the contribution of each of these components to the difference between average cost and the cost in a given firm. The demographic mix of plan enrollees accounts for wide differences in cost ($649). Perhaps the most noticeable feature of the results is that, after adjusting for demographic mix, the difference in expenditures accounted for by the treatment costs given diagnosis ($807) is almost as wide as the unadjusted range in expenditures ($838). Differences in cost due to the different mixes of illnesses that are treated, after adjusting for demographic mix, also accounts for large differences in cost ($626). These components of cost do not move together; for example, demographic mix may decrease expenditure under a particular plan while the diagnosis mix may increase costs. We have also provided an approximate decomposition of the 'variation in expenditures across firms. Although outpatient care accounts for almost 50% of expenditures on average, it accounts for only about 20% of the variation in expenditure across firms. Inpatient care accounts for about 34% of expenditures on average, but almost 59% of the variation in expenditures. Thus one can conclude that variations in high-cost inpatient treatments are a principal cause of the substantial cost variation across firms. (A residual group accounts for about 16% of expenditure and about 20% of variation in expenditure across firms.) The most important component of variation is the diagnosis rate, which accounts for about 52% of variation across firms. Treatment cost differences, given treatment, account for about 40%, with the remainder accounted for by the interaction between the two. Our results suggest that efforts to understand the substantial differences in private insurance expenditures should focus on the relatively intensive, inpatient care, and that further analysis should address the sources of variation in both admissions for intensive treatment (the diagnosis rate) and the cost given treatment.

JEL Classification: I11, I12