The Assessment of Medical Expenditure Risk and the Impact of an Insurance Benefit: Implications for the Development of a Measure of Medical Care Economic Risk

The Affordable Care Act (ACA) extends access to coverage for millions of individuals through the Medicaid expansion, the new Insurance Marketplaces and additional coverage provisions.  The Congressional Budget Office (May 2013) projects that the reduction in the number of uninsured as a result of the ACA will reach 25 million in 2016, with this decrease holding steady through 2023.  Meanwhile, the CBO estimates that the net cost of coverage components of the ACA from 2014 to 2023 will amount to $1,363 billion.  The development of measures that adequately quantify the value of this coverage, in terms of reduction in an individual’s or a family’s medical care economic risk, is an area where further research is needed.  Specifically, developing a measure designed to capture the extent to which enrollement in coverage reduces risk of poverty and economic hardship is an important component of evaluating the impact of the most recent and previous coverage expansions.

Drawing from the prior work of Manning, Basu and Mullahy (2005), Handel (2010) and the conceptual framework in Meier and Wolfe (2012), the analysis used data from panels 2-8 of the Medical Expenditure Panel Survey (MEPS) and the 1996-2004 Risk Adjustment Scores Public Use file of the MEPS to begin exploring components of the operational development of a framework for measuring medical care economic risk.  The analysis focused on individual risk assessment by category of risk for the Medicare population at or above the age of 65.  Assessment of medical expenditure risk entailed baseline risk classification using year one information followed by risk class specific probit estimation and distribution fitting to model year two expenditures.  The use of risk class specific distribution-based assessment of medical expenditure risk paired with the assessment of insurance effects was adopted from the work of Handel (2010).  Drawing from prior work of Manning, Basu and Mullahy (2005), the approach to distribution fitting for positive expenditures focused on examining the fit of special cases of the generalized gamma distribution.   For select total expenditure amounts, insurance characteristics were combined with expenditures for each risk class to map total medical care expenditure amounts to out-of-pocket expenditures.  The practice of mapping to out-of-pocket expenditures is employed by Handel in previous work.

Using modeled expenditures, the analysis demonstrated the risk class specific value of an insurance benefit by comparing the probability of exceeding expenditure thresholds with insurance to the comparable probability without insurance.  By examining two approaches to baseline risk classification, the analysis illustrated the sensitivity of the assessment to baseline risk classification.