|Title||Unintended Consequences of Medicaid Policy for High-Need Beneficiaries|
|Year of Publication||2020|
|Academic Department||Public Policy|
|Degree||Doctor of Philosophy|
|University||University of Chicago|
|Keywords||Long-term Care, Managed care, Medicaid|
This dissertation examines how two healthcare policies impact older adults and individuals with disabilities: long-term care and Medicaid managed care. The first chapter provides the first national examination of long-term care (LTC) setting, hospitalization and spending among the elderly, Medicare-Medicaid dually enrolled. Using national Medicaid claims data, we found that overall hospitalization rates were similar for HCBS and nursing facility users, despite nursing facility users generally being sicker as reflected in their claims history. Among HCBS users, blacks were more likely to be hospitalized than whites, and the gap widened among blacks and whites with dementia. Also, conditional on receiving HCBS, Medicaid HCBS spending was higher for whites than non-whites; higher Medicare and Medicaid spending on hospitalizations for blacks and Hispanics did not offset this difference. In the second chapter, I analyze the Health and Retirement Study (HRS) to estimate the effect of home care versus nursing home care on several measures of physical and mental health. While the analysis in Chapter 1 used claims data to study the effects of Medicaid-paid LTC, the use of the nationally representative HRS allows me to estimate effects of care setting across all payers and across a wider variety of health outcomes than can be studied in the claims data alone. Instrumental variables methods were used to estimate plausibly causal effects of care setting for a specific group of ``marginal'' individuals: those using home care (versus nursing home care) because of state-level Medicaid policies. I find that hospitalization rates were higher but functional decline was slower among home care users. While I find differences in observable characteristics by LTC setting by race and ethnicity, the state-level instrument lacked sufficient power in the black and Hispanic subsamples to explore differences in causal effects of care setting by race/ethnicity. Finally, in the third chapter I examine effects of another Medicaid policy change: inclusion of the Medicare-Medicaid dually enrolled (duals) in Medicaid managed care (MMC). In order to control program costs, many states have recently expanded their MMC programs to include duals. This chapter provides the first national estimates of the effects of MMC expansions from 2005 to 2012 for duals. I use difference-in-differences (DID) and instrumental variables (IV) methods to estimate plausibly causal impacts of three different types of MMC: comprehensive managed care (CMC), managed long-term service and supports (MLTSS), and primary care case management (PCCM). MLTSS plans are associated with increases in hospitalization and potentially avoidable hospitalization and these increases are concentrated among beneficiaries with many chronic conditions. I find mixed effects of CMC expansion: in mandatory enrollment settings, I find modest increased hospitalization while in voluntary enrollment settings, I find decreased hospitalization. Finally, PCCM plans are not associated with changes in hospital use. This study provides the first national estimates of how a major financing change, the inclusion of duals in MMC, impacts hospital use, providing policymakers with much needed evidence as they face the challenge of financing public health insurance programs as health care costs rise and the population ages.