Learning from Program Differences between Medicaid and Medicare

摘要

ABSTRACT Medicare and Medicaid are the largest social health insurance programs in the United States, together providing health insurance for over 130 million Americans. While the programs share some key attributes, they also differ in important ways, with the key institutional difference being that states have significant control over the design of their Medicaid programs while Medicare is managed at the national level. Despite their importance to the U.S. health care system, there is little rigorous quasi-experimental work studying how the two programs compare to each other in terms of their effects on health care access, quality, and outcomes for enrollees. In this project, we will examine the relative effects of these two programs by following adults with disabilities who are enrolled in Medicaid before age 65 into Medicare at age 65 and beyond—that is, as disabled adults with only Medicaid switch to become dual-eligibles enrolled in both Medicaid and Medicare. Adults with disabilities who are dually enrolled in both Medicaid and Medicare before age 65 will serve as the control group, as they do not experience a transition between programs at age 65. Studying these changes to dual eligibility will allow us to isolate the average effects of state Medicaid programs on patient care and outcomes relative to a structurally different Medicare program (e.g., with a wider set of participating providers than Medicaid), as well as to explore variation in these effects across features of state Medicaid programs. In Aim 1, we will estimate how enrollees' care and outcomes differ in these two programs. In Aim 2, we will explore the mechanisms behind the program differences. To do this, we will leverage the fact that there is not one Medicaid program but 51, with varying levels of provider access and payment. Specifically, we will analyze how the Medicaid-Medicare transition differs in states with high provider payments (relative to Medicare) vs. states with low provider payments. We will do the same to compare states that contract out the provision of Medicaid benefits to private managed care plans vs. states that do not. This will provide new evidence on the extent to which these two important factors explain overall Medicaid-Medicare differences and will have implications for program design. In Aim 3, we focus on one specific component of the Medicaid and Medicare benefit packages: prescription drugs. We will use the introduction of Medicare Part D (which shifted dual- eligibles from Medicaid drug coverage to Medicare drug coverage) to assess how the drug benefit component of these two programs differentially affects enrollees' use of drugs and related health outcomes. Here, we will also leverage variation across state Medicaid programs to explore mechanisms, focusing on variation in state use of caps on the number of prescriptions beneficiaries can fill and prior authorization requirements. These analyses will provide critical lessons for the effects of important program design decisions (payment rates, managed care, drug caps, prior authorization) on health care delivery and health outcomes for low-income and disabled Americans.