Will your insurance type influence clinical quality outcomes? An investigation of contributing factors, underlying mechanism, and consequences

Abstract

Although past studies have found differences in clinical quality outcomes based on patients' insurance type, the extent of these discrepancies, the underlying mechanisms supporting this relationship, and factors that can moderate this effect have not yet been addressed. We seek to fill this gap by studying the relationship between insurance type and clinical quality outcomes using the level of care received, that is, treatment intensity, as a mediator. Using agency theory, we argue that underinsured patients (patients insured by Medicaid) will experience more adverse clinical quality outcomes due to lower treatment intensity compared to privately insured patients. We then explore whether two contextual factors—patient complexity and resource utilization—influence the relationship between insurance type and treatment intensity, and in turn clinical quality outcomes. We use 2010–2016 data from California to test these relationships. The insurance types used for comparison are private and Medi‐Cal (California's equivalent of Medicaid, representing underinsured patients). A patient's 30‐day readmission rate is used as a measure of clinical quality and the log of charges normalized by wage index and inflation (called treatment intensity) is used as a proxy for the level of care they received. We use within‐hospital coarsened exact score matching to create a comparable group of patients using private and Medi‐Cal insurance followed by a hospital fixed effects regression with errors clustered by hospital to test these relationships. Results indicate that treatment intensity mediates the relationship between insurance type and clinical quality outcomes. Further, the difference in treatment intensity based on insurance type is reduced as patient complexity increases, while higher utilization at a hospital lowers its incentive to provide equitable care to underinsured patients.