The Impact of Patient Health Insurance Coverage and Latent Health Status on Hospital Readmissions

Hospital readmission rate is adopted as a key hospital quality metric by the Centers for Medicare and Medicaid Services (CMS). In 2012, CMS began penalizing 2200 hospitals with excess readmission rates for a total about $280 million in Medicare payments. However, readmission rate may not be an appropriate quality metric because it may be driven by factors other than the care given at hospitals. In this research, rather than treating readmission as a consequence of a mere inpatient care quality, we investigate several other important factors that may lead to readmission, such as insurance coverage and patient's unobserved health status outside hospitals. We posit that generous insurance coverage may result in higher readmission rates; and deteriorated health status, which is unobservable most of the time, may result in higher readmission rates. We utilize a comprehensive inpatient panel dataset of Congestive Heart Failure patient visits across 68 hospitals in North Texas from 2005 to 2011. We apply a quasi-experimental approach to investigate the impact of switching to Medicare plan on increasing readmission rate and develop a Hidden Markov Model (HMM) to capture the unobservable health status and its impact on the readmission rate. Our quasi-experimental results demonstrate that privately insured or self-pay patients face higher readmission propensity after they enroll in Medicare. HMM estimation results indicate that there is a substantial difference in the readmission rates among different health states (e.g., less healthy had significantly higher readmission rates compared to healthier), all of which reveals the association between level of unobserved health status and readmission rate. This research sheds light onto debate over readmission rate as a sole quality metric and tries to explain that unobservable health status and financial risk borne by patients are likely causes of readmission.