Health Care Spending and Service Use among High-Cost Medicaid Beneficiaries, 2002-2004

Article excerpt

Using program administrative data, this paper examines spending and service use patterns for the national Medicaid population between 2002 and 2004, with a focus on high-cost beneficiaries. We observed a high degree of spending persistence: 57.9% of those who were among the top 10% of Medicaid spenders in 2002 remained in the top 10% of spenders in the two subsequent years. We identified two distinct subgroups of high spenders those with persistently high costs and those with episodically high costs each with different services driving their costs.

**********

Medicaid, the nation's health insurance program for low-income families, elderly individuals, and people with disabilities, cares for some of our most medically complex citizens. Many Medicaid beneficiaries have multiple chronic conditions, and, as a result, many incur high health care costs, often for prolonged periods. Past research has shown that the concentration of spending among Medicaid beneficiaries is high: in 2004, for example, 5% of Medicaid beneficiaries accounted for 57% of Medicaid spending (Cohen and Holahan 2006).

Despite the implications of high-cost beneficiaries to Medicaid spending, scant information is available on the subgroup, especially for those who incur persistently high costs. One recent study (Schneider, Lambrew, and Shenouda 2005) examined a single year of spending for community-based Medicaid beneficiaries but excluded institutionalized beneficiaries who account for nearly a third of Medicaid expenditures (Sommers, Cohen, and O'Malley 2006). Another study (Sommers and Cohen 2006) looked at spending for both community-based and institutionalized high-cost Medicaid beneficiaries, but for just one year.

While studies have been conducted on the persistence of high health care expenditures among the U.S. population overall and the Medicare population (Monheit 2003; CBO 2005; Cohen and Yu 2006), to our knowledge, except for a study on spending among New York City Medicaid beneficiaries (Billings and Mijanovich 2007), there has been no broad examination of the persistence of high Medicaid spending over time for both acute and long-term care.

With Medicaid costs nearly $330 billion in 2007 (CMS 2009) and projected to grow rapidly in the future (CBO 2007), state and federal policymakers have stepped up efforts to find ways to better manage care for beneficiaries and improve program efficiency while potentially reaping some cost savings.

One group that has been a focus of efficiency measures is high-cost beneficiaries (Gillespie and Rossiter 2003; CHCS 2008). Having more insight into these beneficiaries, including those who have persistently high Medicaid costs, would be helpful to state and federal decision makers in crafting sound policies targeted at this population.

Central to developing such strategies is a solid understanding of who these beneficiaries are. For example, how many beneficiaries with high costs in one year have high costs the following year? Or, alternatively, how many experience episodic high costs because of an acute illness or injury'? What personal characteristics or medical conditions are associated with beneficiaries who have persistent or episodic high costs? What types of services are associated with high costs? Are there inefficiencies in delivery of care to high-cost users?

Our study begins to fill this information gap. Building on the limited existing research literature, we examine program spending patterns and service use for the national Medicaid population (both those living in the community and in institutions) over a three-year period from 2002 to 2004.

Data Source

The study's main data set is the 2002 to 2004 Medicaid Statistical Information System (MSIS) Summary File. Under federal Medicaid law, all 50 states and the District of Columbia are required to submit Medicaid eligibility and claims data to the Centers for Medicare and Medicaid Services (CMS) on a quarterly basis. …

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.