This is the accessible text file for GAO report number GAO-14-523 entitled 'Disabled Dual-Eligible Beneficiaries: Integration of Medicare and Medicaid Benefits May Not Lead to Expected Medicare Savings' which was released on September 29, 2014. This text file was formatted by the U.S. Government Accountability Office (GAO) to be accessible to users with visual impairments, as part of a longer term project to improve GAO products' accessibility. Every attempt has been made to maintain the structural and data integrity of the original printed product. Accessibility features, such as text descriptions of tables, consecutively numbered footnotes placed at the end of the file, and the text of agency comment letters, are provided but may not exactly duplicate the presentation or format of the printed version. The portable document format (PDF) file is an exact electronic replica of the printed version. We welcome your feedback. Please E-mail your comments regarding the contents or accessibility features of this document to Webmaster@gao.gov. This is a work of the U.S. government and is not subject to copyright protection in the United States. It may be reproduced and distributed in its entirety without further permission from GAO. Because this work may contain copyrighted images or other material, permission from the copyright holder may be necessary if you wish to reproduce this material separately. United States Government Accountability Office: GAO: Report to Congressional Requesters: August 2014: Disabled Dual-Eligible Beneficiaries: Integration of Medicare and Medicaid Benefits May Not Lead to Expected Medicare Savings: GAO-14-523: GAO Highlights: Highlights of GAO-14-523, a report to congressional requesters. Why GAO Did This Study: In 2009, the Medicare and Medicaid programs spent an estimated $103 billion on disabled dual-eligible beneficiaries—those individuals who are disabled, under age 65, and qualify for both Medicare and Medicaid benefits. Recently, Congress and CMS have emphasized benefit integration for all dual-eligible beneficiaries—both disabled and aged— including beginning a financial alignment demonstration, which CMS expects will improve care and reduce program spending. GAO was asked to provide insights for potentially improving the care provided to disabled dual-eligible beneficiaries while reducing spending. GAO examined (1) spending, utilization, and health status patterns for the portion of this population with the highest spending, (2) the extent to which integrated D-SNPs provided high quality of care for this population while controlling Medicare spending, and (3) D-SNPs’ and traditional MA plans’ performance in serving this population based on quality and resource use measures. To do this work, GAO analyzed Medicare and Medicaid 2009 claims and summary data—the most recent data available. GAO identified D-SNPs that met standards of quality and integration and compared their 2013 costs to expected Medicare FFS spending. GAO used 2011 data—the most recent data available when GAO began its analysis—from the Health Care Effectiveness Data and Information Set to evaluate D-SNPs’ and traditional MA plans’ performance. What GAO Found: Overall spending for high-expenditure disabled dual-eligible beneficiaries—-those in the top 20 percent of spending in their respective states—-was driven largely by Medicaid spending, and the service use and health status often differed widely between those with high Medicare expenditures and high Medicaid expenditures. For these beneficiaries, Medicaid expenditures accounted for nearly two-thirds of overall spending. Also, states with high Medicaid spending often had lower Medicare spending but nearly always had greater overall spending for these beneficiaries. Furthermore, service use and health status often differed widely between high-Medicare-expenditure and high-Medicaid-expenditure disabled dual-eligible beneficiaries. Those with high Medicare expenditures were considerably more likely than those with high Medicaid expenditures to have multiple health conditions and use inpatient services but far less likely to use long- term services and supports. Dual-eligible special needs plans (D-SNP)-—Medicare Advantage (MA) private plans designed to target the needs of dual-eligible beneficiaries—that fully integrated Medicare and Medicaid benefits often met criteria for high quality but had limited experience serving disabled dual-eligible beneficiaries or demonstrating Medicare savings. D-SNPs that the Centers for Medicare & Medicaid Services (CMS)—-the agency that administers Medicare and oversees Medicaid—- designated as Fully Integrated Dual-Eligible (FIDE) SNPs were far more likely to meet high quality criteria compared with other D-SNPs. However, relatively few FIDE-SNPs with high quality served disabled dual-eligible beneficiaries or reported lower costs for Medicare services than expected Medicare fee-for-service (FFS) spending in the same areas. Additionally, FIDE-SNPs that demonstrated the potential for Medicare savings often operated in service areas where D-SNPs with less integration of Medicaid benefits demonstrated more potential for Medicare savings (i.e., lower relative costs for Medicare services). Moderately better health outcomes for disabled dual-eligible beneficiaries in D-SNPs relative to those in traditional MA plans did not translate into lower levels of costly Medicare services (that is, inpatient stays, readmissions, and emergency room visits). These results were also similar whether dual-eligible beneficiaries were at risk for high Medicare spending (those with six or more chronic health conditions), aged (those age 65 and over), or aged and enrolled in FIDE-SNPs. These results suggest that CMS’s expectations regarding the extent to which integration of benefits will produce savings through lower use of costly Medicare services may be optimistic. While operating specialized plans and integrating benefits could lead to improved care, GAO’s results suggest that these conditions may not reduce dual- eligible beneficiaries’ Medicare spending compared with Medicare spending in settings without integrated benefits. CMS reviewed a draft of the report and provided technical comments, which GAO View GAO-14-523. For more information, incorporated as appropriate. View [hyperlink, http://www.gao.gov/products/GAO-14-523]. For more information, contact James Cosgrove at (202) 512-7114 or cosgrovej@gao.gov. [End of section] Contents: Letter: Background: Overall Spending for High-Expenditure Disabled Dual-Eligible Beneficiaries Driven Largely by Medicaid Spending: Service Use and Characteristics Differed Widely between High-Medicare- Expenditure and High-Medicaid-Expenditure Disabled Dual-Eligible Beneficiaries: Fully Integrated D-SNPs Often Provided High Quality Care, but Had Limited Experience Serving Disabled Dual-Eligible Beneficiaries or Demonstrating Medicare Savings: Moderately Better Health Outcomes for Disabled Dual-Eligible Beneficiaries in D-SNPs Relative to Those in Traditional MA Plans Did Not Translate into Lower Levels of Costly Medicare Services: Concluding Observations: Appendix I: Disabled Dual-Eligible Beneficiaries' Service Use and Health Conditions by Expenditure Level, 2009: Appendix II: GAO Contact and Staff Acknowledgments: Tables: Table 1: Medicare and Medicaid Spending for Dual-Eligible Beneficiaries, 2009: Table 2: Performance of Dual-eligible Special Needs Plans (D-SNP) Relative to Traditional Medicare Advantage (MA) Plans on Quality Measures for Disabled Dual-Eligible Beneficiaries, 2011: Figures: Figure 1: Distribution of Combined Medicare and Medicaid Program Spending for Disabled Dual-Eligible Beneficiaries, by Spending Quintile and Program, 2009: Figure 2: Proportion of High-Expenditure Disabled Dual-Eligible Beneficiaries Who Were in the Top Medicare and the Top Medicaid Spending Quintiles, 2009: Figure 3: Per Capita Medicare and Medicaid Spending for High- Expenditure Disabled Dual-Eligible Beneficiaries, by State, 2009: Figure 4: Per Capita Spending for High-Expenditure Disabled Dual- Eligible Beneficiaries Who Used Community-Based Long-term Services and Supports, by State, 2009: Figure 5: Percentage of High-Medicare-Expenditure and High-Medicaid- Expenditure Disabled Dual-Eligible Beneficiaries Who Used Selected Services, 2009: Figure 6: Average Number of Medicare Services Used by High-Medicare- Expenditure Disabled Dual-Eligible Beneficiaries, by Number of Chronic and Mental Health Conditions, 2009: Figure 7: Dual-eligible Special Needs Plans (D-SNP) by High Quality Designation, Fully Integrated Dual-Eligible (FIDE) Designation, and Bid Relative to Medicare Fee-for-Service (FFS) Spending, 2013: Figure 8: Utilization Levels of Selected Costly Medicare Services for Disabled Dual-Eligible Beneficiaries in Dual-eligible Special Needs Plans (D-SNP) and in Traditional Medicare Advantage (MA) Plans, 2011: Abbreviations: CMS: Centers for Medicare & Medicaid Services: D-SNP: dual-eligible special needs plan: ESRD: end-stage renal disease: FFS: fee-for-service: FIDE: Fully Integrated Dual-Eligible: HEDIS: Healthcare Effectiveness Data and Information Set: LTSS: long-term services and supports: MA: Medicare Advantage: MBSF: Master Beneficiary Summary File: MCO: managed care organization: MMLEADS: Medicare-Medicaid Linked Enrollee Analytic Data Source: PPACA: Patient Protection and Affordable Care Act: SNP: special needs plan: SSI: Supplemental Security Income: [End of section] United States Government Accountability Office: GAO: 441 G St. N.W. Washington, DC 20548: August 29, 2014: The Honorable Sander Levin: Ranking Member: Committee on Ways and Means: House of Representatives: The Honorable Henry A. Waxman: Ranking Member: Committee on Energy and Commerce: House of Representatives: Dual-eligible beneficiaries--those qualifying for both Medicare and Medicaid benefits--accounted for under one-fifth of each program's population but over one-third of each program's spending in 2009. [Footnote 1] Dual-eligible beneficiaries are often in poorer health and require more care compared with other Medicare and Medicaid beneficiaries. In 2009, nearly 4 million dual-eligible beneficiaries (40 percent) were under age 65 and qualified for Medicare because they were disabled. Compared with aged (65 and over) dual-eligible beneficiaries, disabled dual-eligible beneficiaries have a higher incidence of mental illness and are less likely to live in an institution. In 2009, the Medicare and Medicaid programs spent an estimated $103 billion combined on all disabled dual-eligible beneficiaries.[Footnote 2] Dual-eligible beneficiaries typically receive their Medicare and Medicaid benefits through each program separately. For Medicare benefits, beneficiaries may opt to enroll in Medicare's traditional fee-for-service (FFS) program or in a private Medicare Advantage (MA) plan (Medicare Part C), which is administered by a Managed Care Organization (MCO), under contract with the Centers for Medicare & Medicaid Services (CMS), the agency in the Department of Health and Human Services that administers the Medicare program and oversees state Medicaid programs. In addition, dual-eligible beneficiaries may choose a type of MA plan called a dual-eligible special needs plan (D- SNP), which is designed to target the needs of this population. [Footnote 3] For Medicaid benefits, beneficiaries generally enroll in their state's Medicaid FFS program or a Medicaid managed care plan [Footnote 4] administered by an MCO under contract with the state. [Footnote 5] Because the Medicare and Medicaid programs are separately responsible for covering certain services for dual-eligible beneficiaries, there may not be an incentive for one program to help control costs in the other program. For example, CMS and others have noted that because state Medicaid programs do not pay for most of the costs of acute hospitalizations for dual-eligible beneficiaries[Footnote 6], there is little incentive for states to ensure that nursing facilities provide a level of care that will avoid unnecessary hospitalizations. In contrast, an incentive may exist for providers to shift beneficiaries from one type of service to another to increase their payments. For example, nursing facilities can benefit financially if their residents are hospitalized for 3 days or more and then returned because nursing facility care, which is typically reimbursed by Medicaid, generally is reimbursed at a higher rate by Medicare for a limited period of time following a hospital stay. These types of program misalignments can result in unnecessary hospitalizations, which can reduce quality of care and increase Medicare costs.[Footnote 7] Recently, Congress and CMS have placed greater emphasis on the coordination and integration of Medicare and Medicaid benefits for dual-eligible beneficiaries. For example, the Medicare Improvements for Patients and Providers Act of 2008 required D-SNPs to contract with state Medicaid agencies to provide Medicaid benefits.[Footnote 8] More recently, the Patient Protection and Affordable Care Act (PPACA) established a type of D-SNP, referred to as a Fully Integrated Dual Eligible (FIDE) SNP, which--unlike other D-SNPs--is designed to integrate program benefits for dual-eligible beneficiaries through a single managed care organization, although payment is generally provided separately by each program.[Footnote 9] In addition, highly integrated D-SNPs--including FIDE-SNPs--that meet certain performance and quality-based standards may seek CMS approval to offer benefits beyond what other MA plans may offer if such benefits would help bridge the gap between Medicare and Medicaid covered services. As required under PPACA, CMS also established the Medicare-Medicaid Coordination Office to more effectively integrate Medicare and Medicaid benefits for dual-eligible beneficiaries and the Center for Medicare and Medicaid Innovation to develop innovative payment and service delivery models for these and other programs. In 2011, CMS, through these two offices, awarded contracts of up to $1 million to 15 states to design new models of care that integrate the two programs' benefits. Later in 2011, CMS announced a financial alignment demonstration that is intended to further integrate the programs' services. CMS expects that the demonstration will decrease incentives for cost shifting and increase care coordination, resulting in improved care for beneficiaries and savings to Medicare and Medicaid. [Footnote 10] You asked us to examine the characteristics, spending, and service use of disabled dual-eligible beneficiaries to help provide insights for potentially increasing quality while reducing Medicare and Medicaid spending. We examined: 1. Medicare and Medicaid spending patterns for high-expenditure disabled dual-eligible beneficiaries; 2. service use and health status of high-Medicare-expenditure and high- Medicaid-expenditure disabled dual-eligible beneficiaries; 3. the extent to which integrated D-SNPs have provided high quality of care for disabled dual-eligible beneficiaries while controlling Medicare spending; and: 4. D-SNPs' and traditional MA plans' performance in serving disabled dual-eligible beneficiaries based on quality and resource use measures. To examine spending patterns for high-expenditure disabled dual- eligible beneficiaries, we analyzed 2009 data, the most recent available, from the Medicare-Medicaid Linked Enrollee Analytic Data Source (MMLEADS) to determine which beneficiaries had the highest Medicare, highest Medicaid, and highest combined program spending. [Footnote 11] We excluded from our analysis disabled dual-eligible beneficiaries who were not eligible for full Medicaid benefits; were under age 21; qualified for Medicare due to end-stage renal disease (ESRD); lived outside the 50 United States and Washington, D.C.; did not have full Medicare coverage (Parts A and B) for all eligible months; or were enrolled in MA for any part of the year. We also excluded individuals who lived in Maine or Arizona because we did not have complete data for these states. We defined high-expenditure beneficiaries as those in the top quintile (top 20 percent) of total program spending in their state.[Footnote 12] We used spending by state because of the variation in states' Medicaid benefits. We analyzed MMLEADS and 2009 Medicare Master Beneficiary Summary File (MBSF) data to examine the distribution of Medicare and Medicaid spending for high-expenditure beneficiaries compared with other beneficiaries.[Footnote 13] We examined the distribution of Medicare spending for all covered services (e.g., inpatient hospital stays, outpatient hospital services, physician and other related services, prescription drugs, post-acute care, and durable medical equipment) and the distribution of Medicaid spending for users of different types of long-term services and supports (LTSS) (e.g., services provided in nursing facilities, intermediate care facilities, or in the community).[Footnote 14] We also analyzed the patterns of program spending across states for high-expenditure beneficiaries. To examine service use and health status of high-Medicare-expenditure and high-Medicaid-expenditure disabled dual-eligible beneficiaries, we used 2009 data from MMLEADS and Medicare carrier claims to determine the proportion of high-expenditure beneficiaries in each program who used certain services.[Footnote 15] We excluded from this analysis the same beneficiaries we excluded from our analysis of spending patterns. For Medicaid services, we focused on the use of different types of LTSS. For Medicare services, we focused on the use of three costly acute services--inpatient stays, readmissions, and emergency room visits--and two other categories of services--primary care and mental health.[Footnote 16] We also determined the proportion of high- expenditure beneficiaries in each program who had chronic and mental health conditions (as defined by CMS's Chronic Conditions Data Warehouse). To examine the extent to which integrated D-SNPs have provided high quality of care for disabled dual-eligible beneficiaries while controlling Medicare spending, we identified D-SNPs that received a FIDE designation from CMS in 2013 and determined whether these plans demonstrated high quality of care, served disabled dual-eligible beneficiaries, and required lower revenue for providing Medicare Part A and B services than comparable spending in Medicare FFS. We classified D-SNPs as high quality if their average performance was in the highest quintile among D-SNPs with sufficient quality data. [Footnote 17] We categorized D-SNPs as serving disabled dual-eligible beneficiaries if those D-SNPs had at least 2 years of experience in serving this population.[Footnote 18] We analyzed the 2013 contract year risk-adjusted bids of D-SNPs relative to the expected Medicare FFS spending for those services in the same service area.[Footnote 19] We used each plan's projected county enrollments and CMS's projected county-level risk-adjusted FFS spending to compute a weighted average of FFS spending in its service area.[Footnote 20] We compared high- quality FIDE-SNPs that bid below or close to Medicare FFS spending with D-SNPs that projected at least 95 percent of their enrollment in the same service area. Because CMS approves benefits flexibility for D- SNPs that meet standards for integration and quality, we also separately examined the quality and relative bids of D-SNPs that were approved for benefits flexibility in 2013. For additional context, we used the quality and bids relative to Medicare FFS spending of highly integrated D-SNPs that CMS approved for flexible benefits to help guide a selection of four FIDE-SNPs with high quality of care and four other highly integrated D-SNPs to interview about their models of care. To examine D-SNPs' and traditional MA plans' performance for disabled dual-eligible beneficiaries on quality and resource use measures, [Footnote 21] we analyzed data from the 2011 MBSF and 2011 Healthcare Effectiveness Data and Information Set (HEDIS) Patient Level Files. [Footnote 22] Specifically, we evaluated performance on 23 process measures and 7 health outcome measures, as well as resource use measures for three costly acute Medicare-covered services: inpatient stays, readmissions, and emergency room visits.[Footnote 23] We excluded beneficiaries who were not eligible for full Medicaid benefits; were under age 18; qualified for Medicare due to ESRD; lived outside the 50 United States and Washington, D.C.; did not have full Medicare coverage (Part A and Part B) for the entire year; were not enrolled in MA for the entire year; or for whom data were missing or inconsistent. As some of D-SNPs' care coordination approaches may be targeted to the sickest beneficiaries, we evaluated performance for all disabled dual-eligible beneficiaries as well as just for those with six or more chronic conditions (as defined by CMS's Hierarchical Condition Categories).[Footnote 24] To provide additional context, we also evaluated performance for aged dual-eligible beneficiaries, including how D-SNPs' performance varied by FIDE designation.[Footnote 25] To assess the reliability of the data we used in our analyses, we reviewed related documentation, interviewed knowledgeable officials from CMS and its contractors, and performed appropriate electronic data checks. This allowed us to determine that the data were suitable for our purposes. We conducted this performance audit from April 2013 to July 2014 in accordance with generally accepted government auditing standards. Those standards require that we plan and perform the audit to obtain sufficient, appropriate evidence to provide a reasonable basis for our findings and conclusions based on our audit objectives. We believe that the evidence obtained provides a reasonable basis for our findings and conclusions based on our audit objectives. Background: Medicare and Medicaid Coverage for Disabled Dual-Eligible Beneficiaries: Beneficiaries under age 65 may qualify for Medicare coverage on the basis of disability (such as a physical disability, developmental disability, or disabling mental health condition).[Footnote 26] Disabled individuals typically enroll in the federal Social Security Disability Insurance program and then have a 24-month waiting period before Medicare benefits begin.[Footnote 27] During the waiting period, low-income individuals who qualify for the Supplemental Security Income program (SSI) in their state can also qualify for Medicaid coverage. SSI is a means-tested income assistance program that provides cash benefits to individuals who meet certain disability criteria and have low levels of income and assets. After the Medicare waiting period ends, beneficiaries become dually enrolled in both programs. Medicare becomes the primary payer for most services, but Medicaid continues to cover benefits not offered by Medicare.[Footnote 28] Medicare coverage for dual-eligible beneficiaries includes hospitalizations, physician services, prescription drugs, skilled nursing facility care, home health visits, and hospice care. Under Medicaid, states are required to cover certain items and services for dual-eligible beneficiaries, including nursing facility services and home health services.[Footnote 29] Although states are required to cover certain populations and services, they have the option to expand coverage beyond these mandatory levels, and accordingly state Medicaid programs vary in scope. Characteristics of Disabled Dual-Eligible Beneficiaries Compared with Aged Dual-Eligible Beneficiaries: Compared with aged dual-eligible beneficiaries, disabled dual-eligible beneficiaries in 2009 were more likely to be male and African- American, and they tended to have a much higher incidence of mental illness. However, they had a far lower incidence of three or more chronic conditions and were less likely to be institutionalized.[Footnote 30] In terms of relative spending, disabled dual-eligible beneficiaries had lower per capita Medicare spending but higher per capita Medicaid spending in 2009. However, among beneficiaries who did not use LTSS, per capita Medicare and Medicaid spending were both slightly higher for disabled dual-eligible beneficiaries. Disabled dual-eligible beneficiaries were less likely to live in an institution than aged dual eligible beneficiaries; however, among those that did, Medicaid spending was significantly higher than for aged dual-eligible beneficiaries in institutions.[Footnote 31] See table 1. Table 1: Medicare and Medicaid Spending for Dual-Eligible Beneficiaries, 2009: Type of beneficiary: All; Per capita Medicare spending: Aged: $16,878; Per capita Medicare spending: Disabled: $14,183; Per capita Medicaid spending: Aged: $13,501; Per capita Medicaid spending: Disabled: $13,651. Type of beneficiary: Nonuser of long-term services and supports (LTSS); Per capita Medicare spending: Aged: $11,704; Per capita Medicare spending: Disabled: $11,806; Per capita Medicaid spending: Aged: $2,078; Per capita Medicaid spending: Disabled: $2,704. Type of beneficiary: LTSS user; Per capita Medicare spending: Aged: $24,585; Per capita Medicare spending: Disabled: $20,893; Per capita Medicaid spending: Aged: $30,513; Per capita Medicaid spending: Disabled: $44,560. Type of beneficiary: Institutional LTSS[A]; Per capita Medicare spending: Aged: [B]; Per capita Medicare spending: Disabled: [B]; Per capita Medicaid spending: Aged: $38,196; Per capita Medicaid spending: Disabled: $67,299. Type of beneficiary: Community-based LTSS[C]; Per capita Medicare spending: Aged: [B]; Per capita Medicare spending: Disabled: [B]; Per capita Medicaid spending: Aged: $13,582; Per capita Medicaid spending: Disabled: $28,672. Source: MedPAC. GAO-14-523. Notes: Analysis excludes end-stage renal disease beneficiaries. Medicare spending includes only Medicare fee-for-service spending. Medicaid spending includes both Medicaid fee-for-service and managed care spending, but excludes Medicaid payments of Medicare premiums. Spending amounts include spending for all services covered by Medicare and Medicaid and vary according to whether the beneficiary used LTSS, and if so, which type. [A] Institutional LTSS include services provided in nursing facilities, intermediate care facilities for persons with intellectual disabilities, inpatient psychiatric facilities for individuals under age 21, and institutions for individuals age 65 or older with a mental illness. [B] Data on Medicare spending for institutional and community-based LTSS users were not provided. [C] Community-based LTSS include home health services, personal care services, and home and community-based services. [End of table] Special Needs Plans: In addition to requiring SNPs to meet all the requirements of other MA plans, CMS requires SNPs to provide specialized services targeted to the needs of their unique beneficiaries, including beneficiaries with certain severe and disabling chronic conditions, beneficiaries who live in institutions, or beneficiaries dually enrolled in both the Medicare and Medicaid programs. SNPs that provide specialized services, such as case management, for dual-eligible beneficiaries are referred to as D-SNPs.[Footnote 32] CMS pays D-SNPs the same way that it pays other MA plans, that is, a monthly amount determined by the plan bid--the plan's estimated cost of providing Medicare Part A and Part B benefits--in relation to a benchmark, which is the maximum amount the Medicare program will pay MA plans in a given locality. CMS then adjusts the monthly payments to MA plans on the basis of beneficiaries' risk scores. For MA plans that bid below the benchmark, CMS provides a rebate that is modified by an overall assessment of quality at the contract level[Footnote 33] using a 5-star rating scale based on measures of clinical quality and patients' reported care experience for Medicare Part C and Part D. [Footnote 34] CMS designates certain D-SNPs as FIDE-SNPs, which is a designation for plans that integrate Medicare and Medicaid program benefits for dual- eligible beneficiaries through a single MCO. All FIDE-SNPs are financially at risk for enrollees' nursing facility services for at least 6 months of the year. In 2013, CMS designated 35 FIDE-SNPs that enrolled about 98,000 beneficiaries across seven states.[Footnote 35] Beginning in contract year 2013, CMS may give certain D-SNPs that meet a high standard of integration and specified performance and quality- based standards the flexibility to offer supplemental benefits beyond those that CMS currently allows for other MA plans if the agency determines these benefits would better integrate care.[Footnote 356] This benefits flexibility is designed to assist dual-eligible beneficiaries who are at risk of institutionalization to remain in the community and may prevent health status decline and reduce the quantity and cost of health care services. As part of the qualifying criteria that CMS currently applies to benefits flexibility, the agency requires D-SNPs either to be in a contract with a 3 star (or higher) overall rating for the previous contract year[Footnote 37] or if the D-SNP is part of a contract that does not have sufficient enrollment to generate a star rating, to score 75 percent or above on five of seven specific SNP plan-level HEDIS measures. For 2013, CMS approved 21 highly integrated D-SNPs for benefits flexibility that enrolled about 75,000 beneficiaries. CMS Financial Alignment Demonstration: In 2011, CMS announced a financial alignment demonstration that is intended to align Medicare and Medicaid services and funding to reduce costs and improve the quality of care for dual-eligible beneficiaries. As of June 2014, 12 states have been approved to participate in the demonstration and 4 states have active proposals pending.[Footnote 38] Each demonstration is authorized for at least 3 years. Most states participating in the financial alignment demonstration plan to use a capitated model.[Footnote 39] Under the capitated model, CMS and states provide a single capitated payment to health plans to provide all Medicare and Medicaid benefits to enrolled dual-eligible beneficiaries. Payment rates to health plans will be reduced up front each year based on a predetermined combined Medicare and Medicaid savings estimate by CMS. In general, CMS and states expect the savings percentages to increase during the second and third years of the demonstration. Furthermore, to encourage quality of care improvements, CMS and states will withhold a portion of the payments--starting at 1 percent in the first year and up to 3 percent in the third year--that participating health plans can earn back by meeting a certain threshold of performance on quality measures. CMS expects that the capitated model will result in savings (1) to Medicare by reducing hospital admissions, emergency room visits, and skilled nursing care, and (2) to Medicaid by avoiding costly long-term nursing home care.[Footnote 40] At the same time, CMS expects that there may be increased use of primary care services, outpatient services, behavioral health services, and community-based LTSS, due to a greater emphasis on care coordination and maintaining beneficiaries in the community.[Footnote 41] Although CMS projects that approximately 61 to 75 percent of savings will come from reductions in costly Medicare-covered services, the agency requires that--as part of a more integrated approach--both the Medicare and Medicaid programs adjust their payment rates to plans based on aggregate savings percentages.[Footnote 42] In Massachusett's demonstration, only disabled dual-eligible beneficiaries between the ages of 21 and 64 are eligible for enrollment. Voluntary enrollment for the program began on October 1, 2013, and passive enrollment--whereby individuals are automatically enrolled in the program but can opt out--began January 1, 2014. Since 2004, Massachusetts has separately participated in a financial alignment program for dual-eligible beneficiaries age 65 and older known as Senior Care Options. This program provides all of the services covered by Medicare and MassHealth--the Massachusetts Medicaid program--and is funded by a combined capitated payment from both programs. Overall Spending for High-Expenditure Disabled Dual-Eligible Beneficiaries Driven Largely by Medicaid Spending: Medicaid Spending--Particularly for Users of Community-based LTSS-- Accounted for Nearly Two-Thirds of Overall Spending for High- Expenditure Beneficiaries: High Medicaid spending for disabled dual-eligible beneficiaries drove high combined (Medicare and Medicaid) program spending for these beneficiaries. Beneficiaries ranked within the top 20 percent--or top quintile--of spending in their respective states accounted for more than 60 percent of national combined program spending for disabled dual-eligible beneficiaries. Furthermore, for these high-expenditure beneficiaries, nearly two-thirds (63 percent) of combined program spending was Medicaid spending and slightly over one-third (37 percent) was Medicare spending. (See figure 1.) Figure 1: Distribution of Combined Medicare and Medicaid Program Spending for Disabled Dual-Eligible Beneficiaries, by Spending Quintile and Program, 2009: [Refer to PDF for image: vertical bar graph with subchart] Population: Bottom quintile: 20%; Middle quintiles: 60%; Top quintile: 20%. Spending: Bottom quintile: 1%; Middle quintiles: 36%; Top quintile: 63%: of that: - Medicaid: 63%; - Medicare: 37%. Source: GAO analysis of CMS data. GAO-14-523. Notes: The figure is based on analysis of 2009 Medicare-Medicaid Linked Enrollee Analytic Data Source data. The analysis excluded beneficiaries who were not eligible for full Medicaid benefits; were under age 21; qualified for Medicare due to end-stage renal disease; lived outside the 50 United States and Washington, D.C.; did not have full Medicare coverage (Part A and Part B) for all eligible months; or were enrolled in Medicare Advantage for any part of the year. The analysis also excluded beneficiaries who lived in Maine or Arizona because we did not have complete data for these states. To determine beneficiaries' spending quintile, beneficiaries were ranked within their state of residence by their total combined (Medicare and Medicaid) program spending and divided into quintiles. [End of figure] High-expenditure beneficiaries in the top combined spending quintile were also more likely to be in the top Medicaid spending quintile in their states than in the top Medicare spending quintile. Specifically, 72 percent of high-expenditure beneficiaries were in the top Medicaid spending quintile, while 54 percent of these beneficiaries were in the top Medicare spending quintile. Only 26 percent of high-expenditure beneficiaries were in both the top Medicare and the top Medicaid spending quintiles in their states. (See figure 2.) Figure 2: Proportion of High-Expenditure Disabled Dual-Eligible Beneficiaries Who Were in the Top Medicare and the Top Medicaid Spending Quintiles, 2009: [Refer to PDF for image: horizontal bar graph] Top Medicaid spending quintile only: 46%; Top Medicaid and Medicare spending quintile: 26%; Top Medicare spending quintile only: 28%. Source: GAO analysis of CMS data. GAO-14-523. Notes: The figure is based on analysis of 2009 Medicare-Medicaid Linked Enrollee Analytic Data Source data. The analysis excluded beneficiaries who were not eligible for full Medicaid benefits; were under age 21; qualified for Medicare due to end-stage renal disease; lived outside the 50 United States and Washington, D.C.; did not have full Medicare coverage (Part A and Part B) for all eligible months; or were enrolled in Medicare Advantage for any part of the year. The analysis also excluded beneficiaries who lived in Maine or Arizona because we did not have complete data for these states. High- expenditure beneficiaries are those in the top combined (Medicare and Medicaid) spending quintile in their states. To determine beneficiaries' spending quintiles, beneficiaries were ranked within their state of residence by their total combined program spending, their total Medicare spending, and their total Medicaid spending, and divided into quintiles for each type of spending. Less than 1 percent of beneficiaries in the top combined spending quintile were in neither the top Medicare nor the top Medicaid spending quintiles. [End of figure] Just over half (52 percent) of Medicaid spending for high-expenditure disabled dual-eligible beneficiaries was for those who used community- based LTSS, while for low-expenditure beneficiaries (those in the bottom combined spending quintile), community-based LTSS users accounted for only 4 percent of Medicaid spending.[Footnote 43] Inpatient stays accounted for the largest share (39 percent) of Medicare spending for high-expenditure beneficiaries but accounted for less than 1 percent of Medicare spending for low-expenditure beneficiaries. States with High Medicaid Spending Often Had Lower Medicare Spending but Greater Overall Spending for High-Expenditure Beneficiaries: Because Medicaid benefits differ across states, per capita Medicaid spending for high-expenditure disabled dual-eligible beneficiaries varied more across states than per capita Medicare spending did. Per capita Medicaid spending for these beneficiaries ranged from about $27,000 in Michigan to about $188,000 in New York, while per capita Medicare spending ranged from about $18,000 in North Dakota to about $49,000 in Florida. Largely because of the variation in Medicaid spending, combined program spending ranged from about $68,000 in Alabama to about $220,000 in New York. Although states with greater per capita Medicaid spending for high- expenditure beneficiaries often had less per capita Medicare spending, they were usually among the states with the greatest per capita combined spending.[Footnote 44] (See figure 3.) For example, 5 of the 10 states with the greatest per capita Medicaid spending had per capita Medicare spending that was less than the average across states. Nevertheless, all 10 states were among those with the greatest per capita combined spending. In fact, 17 of the 20 states with the greatest per capita Medicaid spending were also among those with the greatest per capita combined spending. Figure 3: Per Capita Medicare and Medicaid Spending for High- Expenditure Disabled Dual-Eligible Beneficiaries, by State, 2009: [Refer to PDF for image: vertical bar graph] Per capita spending: State: Michigan; Medicare: $45,144; Medicaid: $26,840; Combined: $71,984. State: Alabama; Medicare: $31,645; Medicaid: $36,691; Combined: $68,336. State: South Carolina; Medicare: $35,723; Medicaid: $38,309; Combined: $74,032. State: Oregon; Medicare: $36,808; Medicaid: $39,557; Combined: $76,365. State: California; Medicare: $46,100; Medicaid: $39,576; Combined: $85,676. State: Montana; Medicare: $32,711; Medicaid: $40,439; Combined: $73,151. State: Georgia; Medicare: $32,628; Medicaid: $40,710; Combined: $73,338. State: Wisconsin; Medicare: $33,194; Medicaid: $41,936; Combined: $75,130. State: Kentucky; Medicare: $34,155; Medicaid: $43,485; Combined: $77,641. State: North Carolina; Medicare: $37,568; Medicaid: $43,908; Combined: $81,476. State: Mississippi; Medicare: $36,118; Medicaid: $45,412; Combined: $81,531. State: Illinois; Medicare: $42,549; Medicaid: $46,259; Combined: $88,808. State: West Virginia; Medicare: $31,147; Medicaid: $47,659; Combined: $78,806. State: Florida; Medicare: $48,983; Medicaid: $48,923; Combined: $97,906. State: Missouri; Medicare: $41,290; Medicaid: $49,654; Combined: $90,944. State: Oklahoma; Medicare: $41,720; Medicaid: $50,393; Combined: $92,113. State: Washington; Medicare: $36,659; Medicaid: $51,087; Combined: $87,746. State: Hawaii; Medicare: $29,261; Medicaid: $52,600; Combined: $81,861. State: Texas; Medicare: $45,072; Medicaid: $53,437; Combined: $98,509. State: Tennessee; Medicare: $38,054; Medicaid: $53,490; Combined: $91,544. State: Idaho; Medicare: $33,305; Medicaid: $53,915; Combined: $87,220. State: South Dakota; Medicare: $27,088; Medicaid: $57,214; Combined: $84,302. State: Utah; Medicare: $29,870; Medicaid: $57,419; Combined: $87,289. State: Nevada; Medicare: $47,258; Medicaid: $57,468; Combined: $104,727. State: Arkansas; Medicare: $29,889; Medicaid: $57,545; Combined: $87,434. State: Vermont; Medicare: $24,107; Medicaid: $58,094; Combined: $82,201. State: Nebraska; Medicare: $35,262; Medicaid: $59,796; Combined: $95,057. State: Pennsylvania; Medicare: $37,591; Medicaid: $59,906; Combined: $97,496. State: Massachusetts; Medicare: $40,596; Medicaid: $61,803; Combined: $102,399. State: Kansas; Medicare: $33,743; Medicaid: $63,621; Combined: $97,364. State: Louisiana; Medicare: $38,163; Medicaid: $64,019; Combined: $102,181. State: Colorado; Medicare: $35,563; Medicaid: $67,490; Combined: $103,053. State: Iowa; Medicare: $28,786; Medicaid: $70,590; Combined: $99,376. State: Virginia; Medicare: $31,586; Medicaid: $73,446; Combined: $105,032. State: New Mexico; Medicare: $22,834; Medicaid: $74,728; Combined: $97,562. State: Indiana; Medicare: $35,726; Medicaid: $76,777; Combined: $112,503. State: Maryland; Medicare: $38,884; Medicaid: $79,056; Combined: $117,941. State: Ohio; Medicare: $37,809; Medicaid: $79,691; Combined: $117,500. State: New Hampshire; Medicare: $31,269; Medicaid: $79,707; Combined: $110,976. State: Rhode Island; Medicare: $37,345; Medicaid: $82,998; Combined: $120,343. State: Alaska; Medicare: $36,642; Medicaid: $83,876; Combined: $120,518. State: Minnesota; Medicare: $29,450; Medicaid: $92,099; Combined: $121,548. State: District of Columbia; Medicare: $44,059; Medicaid: $93,671; Combined: $137,730. State: Wyoming; Medicare: $31,806; Medicaid: $96,399; Combined: $128,205. State: New Jersey; Medicare: $37,555; Medicaid: $104,160; Combined: $141,714. State: North Dakota; Medicare: $18,035; Medicaid: $109,525; Combined: $127,560. State: Delaware; Medicare: $36,952; Medicaid: $112,649; Combined: $149,601. State: Connecticut; Medicare: $32,519; Medicaid: $144,463; Combined: $176,982. State: New York; Medicare: $32,635; Medicaid: $187,821; Combined: $220,456. Source: GAO analysis of CMS data. GAO-14-523. Notes: The figure is based on analysis of 2009 Medicare-Medicaid Linked Enrollee Analytic Data Source data. The analysis excluded beneficiaries who were not eligible for full Medicaid benefits; were under age 21; qualified for Medicare due to end-stage renal disease; lived outside the 50 United States and Washington, D.C.; did not have full Medicare coverage (Part A and Part B) for all eligible months; or were enrolled in Medicare Advantage for any part of the year. The analysis also excluded beneficiaries who lived in Maine or Arizona because we did not have complete data for these states. High expenditure beneficiaries represent those in the top combined (Medicare and Medicaid) spending quintile in their state. To determine beneficiaries' spending quintile, beneficiaries were ranked within their state of residence by their total combined program spending and divided into quintiles. Within the figure, states are ordered by increasing Medicaid spending. [End of figure] Because the majority of spending for high-expenditure disabled dual- eligible beneficiaries was for beneficiaries who used community-based LTSS, and because CMS expects to see an increase in the use of these services under the financial alignment demonstration, we repeated this analysis for only beneficiaries who used community-based LTSS and found similar results.[Footnote 45] States that had greater per capita Medicaid spending for high-expenditure beneficiaries who used community-based LTSS tended to have less per capita Medicare spending but greater per capita combined spending for these beneficiaries. [Footnote 46] (See figure 4.) In particular, 8 of the 10 states with the greatest per capita Medicaid spending for community-based LTSS users had per capita Medicare spending that was less than the average across states. Nevertheless, 9 of the 10 states with the greatest per capita Medicaid spending were among the 10 states with the greatest per capita combined spending. Figure 4: Per Capita Spending for High-Expenditure Disabled Dual- Eligible Beneficiaries Who Used Community-Based Long-term Services and Supports, by State, 2009: [Refer to PDF for image: vertical bar graph] Per capita spending: State: Mississippi; Medicare: $41,292; Medicaid: $23,557; Combined: $64,849. State: Oregon; Medicare: $36,233; Medicaid: $31,753; Combined: $67,986. State: Michigan; Medicare: $37,128; Medicaid: $36,061; Combined: $73,188. State: North Carolina; Medicare: $37,149; Medicaid: $37,472; Combined: $74,621. State: Rhode Island; Medicare: $53,826; Medicaid: $38,996; Combined: $92,822. State: California; Medicare: $38,993; Medicaid: $39,696; Combined: $78,689. State: Montana; Medicare: $35,073; Medicaid: $39,894; Combined: $74,967. State: South Carolina; Medicare: $33,433; Medicaid: $40,416; Combined: $73,849. State: Illinois; Medicare: $38,671; Medicaid: $40,578; Combined: $79,249. State: Idaho; Medicare: $31,882; Medicaid: $43,645; Combined: $75,526. State: Texas; Medicare: $38,314; Medicaid: $44,320; Combined: $82,634. State: Florida; Medicare: $41,504; Medicaid: $47,662; Combined: $89,166. State: Georgia; Medicare: $21,266; Medicaid: $50,425; Combined: $71,691. State: Alabama; Medicare: $23,150; Medicaid: $50,710; Combined: $73,860. State: South Dakota; Medicare: $21,842; Medicaid: $53,356; Combined: $75,198. State: Pennsylvania; Medicare: $25,253; Medicaid: $53,824; Combined: $79,077. State: Missouri; Medicare: $33,672; Medicaid: $54,390; Combined: $88,062. State: West Virginia; Medicare: $23,902; Medicaid: $56,384; Combined: $80,286. State: Kentucky; Medicare: $22,728; Medicaid: $56,834; Combined: $79,562. State: Iowa; Medicare: $24,453; Medicaid: $57,520; Combined: $81,974. State: Wisconsin; Medicare: $19,675; Medicaid: $58,795; Combined: $78,471. State: Arkansas; Medicare: $21,855; Medicaid: $59,048; Combined: $80,903. State: Washington; Medicare: $30,456; Medicaid: $59,180; Combined: $89,636. State: Oklahoma; Medicare: $37,236; Medicaid: $59,326; Combined: $96,562. State: Kansas; Medicare: $31,184; Medicaid: $61,814; Combined: $92,998. State: Nevada; Medicare: $38,753; Medicaid: $62,745; Combined: $101,498. State: Nebraska; Medicare: $25,550; Medicaid: $63,645; Combined: $89,195. State: Hawaii; Medicare: $11,564; Medicaid: $66,462; Combined: $78,026. State: Vermont; Medicare: $16,065; Medicaid: $68,162; Combined: $84,226. State: Utah; Medicare: $17,137; Medicaid: $70,991; Combined: $88,128. State: Colorado; Medicare: $30,558; Medicaid: $71,736; Combined: $102,294. State: Louisiana; Medicare: $28,763; Medicaid: $73,481; Combined: $102,244. State: Virginia; Medicare: $23,728; Medicaid: $74,770; Combined: $98,498. State: New Jersey; Medicare: $27,417; Medicaid: $82,114; Combined: $109,530 State: Ohio; Medicare: $34,639; Medicaid: $82,278; Combined: $116,917. State: Massachusetts; Medicare: $23,424; Medicaid: $82,373; Combined: $105,797. State: New Hampshire; Medicare: $24,717; Medicaid: $87,262; Combined: $111,979. State: New Mexico; Medicare: $12,525; Medicaid: $87,757; Combined: $100,282. State: Tennessee; Medicare: $19,680; Medicaid: $88,988; Combined: $108,667. State: Alaska; Medicare: $28,209; Medicaid: $90,985; Combined: $119,194. State: District of Columbia; Medicare: $37,090; Medicaid: $91,075; Combined: $128,166. State: Maryland; Medicare: $20,267; Medicaid: $91,601; Combined: $111,868. State: Minnesota; Medicare: $26,966; Medicaid: $93,884; Combined: $120,850. State: Indiana; Medicare: $26,368; Medicaid: $94,051; Combined: $120,420. State: Wyoming; Medicare: $29,538; Medicaid: $94,426; Combined: $123,964. State: North Dakota; Medicare: $16,559; Medicaid: $98,272; Combined: $114,831. State: Delaware; Medicare: $26,502; Medicaid: $114,178; Combined: $140,681. State: Connecticut; Medicare: $22,664; Medicaid: $139,123; Combined: $161,787. State: New York; Medicare: $22,161; Medicaid: $167,174; Combined: $189,335. Source: GAO analysis of CMS data. GAO-14-523. Notes: The figure is based on analysis of 2009 Medicare-Medicaid Linked Enrollee Analytic Data Source data. The analysis excluded beneficiaries who were not eligible for full Medicaid benefits; were under age 21; qualified for Medicare due to end-stage renal disease; lived outside the 50 United States and Washington, D.C.; did not have full Medicare coverage (Part A and Part B) for all eligible months; or were enrolled in Medicare Advantage for any part of the year. The analysis also excluded beneficiaries who lived in Maine or Arizona because we did not have complete data for these states. High expenditure beneficiaries represent those in the top combined (Medicare and Medicaid) spending quintile in their state. To determine beneficiaries' spending quintile, beneficiaries were ranked within their state of residence by their total combined program spending and divided into quintiles. Within the figure, states are ordered by increasing Medicaid spending. [End of figure] Service Use and Characteristics Differed Widely between High-Medicare- Expenditure and High-Medicaid-Expenditure Disabled Dual-Eligible Beneficiaries: Beneficiaries with High Medicare Expenditures Were More Likely to Use Inpatient Services; Beneficiaries with High Medicaid Expenditures Were More Likely to Use LTSS: The services most commonly used by disabled dual-eligible beneficiaries in the top Medicare-spending quintile and those in the top Medicaid-spending quintile often differed widely. As expected-- because Medicare is the primary payer for acute hospital stays-- beneficiaries with high Medicare expenditures were highly likely to use inpatient services. However, a relatively low percentage of beneficiaries with high Medicaid expenditures used these services. In contrast, beneficiaries with high Medicaid expenditures were much more likely than beneficiaries with high Medicare expenditures to use LTSS-- particularly community-based LTSS.[Footnote 47] (See figure 5.) Beneficiaries with high Medicaid expenditures were also more likely to use community-based LTSS and less likely to use inpatient services than most beneficiaries with lower Medicaid expenditures. Figure 5: Percentage of High-Medicare-Expenditure and High-Medicaid- Expenditure Disabled Dual-Eligible Beneficiaries Who Used Selected Services, 2009: [Refer to PDF for image: horizontal bar graph] Acute Medicare services: 1+ inpatient stays: High-Medicare-expenditure beneficiaries: 63%; High-Medicaid-expenditure beneficiaries: 22%. 1+ readmissions[A]: High-Medicare-expenditure beneficiaries: 31%; High-Medicaid-expenditure beneficiaries: 25%. 1+ emergency room visits: High-Medicare-expenditure beneficiaries: 76%; High-Medicaid-expenditure beneficiaries: 44%. Medicaid LTSS: Any setting: High-Medicare-expenditure beneficiaries: 44%; High-Medicaid-expenditure beneficiaries: 90%. Institutional nursing facility: High-Medicare-expenditure beneficiaries: 13%; High-Medicaid-expenditure beneficiaries: 18%. Intermediate care facility[B]: High-Medicare-expenditure beneficiaries: 2%; High-Medicaid-expenditure beneficiaries: 14%. Community[C]: High-Medicare-expenditure beneficiaries: 29%; High-Medicaid-expenditure beneficiaries: 59%. Source: GAO analysis of CMS data. GAO-14-523. Notes: This analysis used 2009 Medicare-Medicaid Linked Enrollee Analytic Data Source data. The analysis excluded beneficiaries who were not eligible for full Medicaid benefits; were under age 21; qualified for Medicare due to end-stage renal disease; lived outside the 50 United States and Washington, D.C.; did not have full Medicare coverage (Part A and Part B) for all eligible months; or were enrolled in Medicare Advantage for any part of the year. The analysis also excluded beneficiaries who lived in Maine or Arizona because we did not have complete data for these states. High-Medicare-expenditure beneficiaries represent those in the top Medicare spending quintile in their states. High-Medicaid-expenditure beneficiaries represent those in the top Medicaid spending quintile in their states. To determine beneficiaries' spending quintiles, beneficiaries were ranked within their state of residence by their total Medicare spending and their total Medicaid spending and divided into quintiles for each type of spending. We categorized as LTSS users those beneficiaries who used LTSS for at least 3 months during the year, then categorized them according to the type of LTSS that was used for the greatest number of months. Percentages for LTSS settings--institutional nursing facility, intermediate care facility, or community--may not sum to percentage for "any setting" due to rounding. [A] Readmissions are a subset of beneficiaries who had at least 1 inpatient stay and include admissions for any reason within 30 days of a hospital discharge. [B] Intermediate care facility for persons with intellectual disabilities. [C] Community-based LTSS include home health services, personal care services, and home and community-based services. [End of figure] Despite differences in service use, high-Medicare-expenditure and high- Medicaid-expenditure disabled dual-eligible beneficiaries had similar utilization levels of Medicare-covered primary care and mental health services. Nearly 80 percent of beneficiaries in each group received at least one primary care service, and approximately 40 percent of beneficiaries in each group received at least one mental health service. Both groups also were more likely to receive primary care or mental health services than beneficiaries with lower expenditures in their respective programs. (See appendix I. for more information on the use of selected Medicare and Medicaid services by disabled dual-eligible beneficiary expenditure levels.) Beneficiaries with High Medicare Expenditures Were Far More Likely than Those with High Medicaid Expenditures to Have Multiple Health Conditions: Disabled dual-eligible beneficiaries with high Medicare expenditures were considerably more likely than beneficiaries with high Medicaid expenditures to have multiple chronic or mental health conditions. About 35 percent of beneficiaries with high Medicare expenditures had six or more chronic conditions, compared with 14 percent of beneficiaries with high Medicaid expenditures. In addition, 25 percent of beneficiaries with high Medicare expenditures had three or more mental health conditions, compared with 13 percent of beneficiaries with high Medicaid expenditures. The presence of multiple health conditions may drive the use of costly Medicare services among beneficiaries with high Medicare expenditures. As the number of chronic and mental health conditions increased among these beneficiaries, the average number of emergency room visits, inpatient stays, and readmissions also increased. The increase in the average number of emergency room visits was particularly dramatic as the number of mental health conditions increased. (See figure 6.) Furthermore, as the number of chronic conditions increased, the percentage of beneficiaries with high Medicare expenditures who had at least one inpatient admission increased substantially, although as the number of mental health conditions increased, the percentage was relatively stable. Figure 6: Average Number of Medicare Services Used by High-Medicare- Expenditure Disabled Dual-Eligible Beneficiaries, by Number of Chronic and Mental Health Conditions, 2009: [Refer to PDF for image: 2 multiple line graphs] Number of services: Number of chronic conditions: 0; Inpatient stays: 0.4; Readmissions[A]: 0.3; Emergency room visits: 1.7. Number of chronic conditions: 1; Inpatient stays: 0.7; Readmissions[A]: 0.4; Emergency room visits: 2.5. Number of chronic conditions: 2; Inpatient stays: 1; Readmissions[A]: 0.4; Emergency room visits: 2.9. Number of chronic conditions: 3; Inpatient stays: 1.1; Readmissions[A]: 0.5; Emergency room visits: 3.2. Number of chronic conditions: 4; Inpatient stays: 1.3; Readmissions[A]: 0.5; Emergency room visits: 3.3. Number of chronic conditions: 5; Inpatient stays: 1.4; Readmissions[A]: 0.5; Emergency room visits: 3.5. Number of chronic conditions: 6+; Inpatient stays: 2.1; Readmissions[A]: 0.8; Emergency room visits: 4.3. Number of services: Number of mental health conditions: 0; Inpatient stays: 1.3; Readmissions[A]: 0.5; Emergency room visits: 2.3. Number of mental health conditions: 1; Inpatient stays: 1.3; Readmissions[A]: 0.6; Emergency room visits: 2.7. Number of mental health conditions: 2; Inpatient stays: 1.5; Readmissions[A]: 0.6; Emergency room visits: 3.5. Number of mental health conditions: 3; Inpatient stays: 1.5; Readmissions[A]: 0.7; Emergency room visits: 4.3. Number of mental health conditions: 4; Inpatient stays: 1.7; Readmissions[A]: 0.8; Emergency room visits: 5.5. Number of mental health conditions: 5; Inpatient stays: 1.8; Readmissions[A]: 1; Emergency room visits: 7.2. Number of mental health conditions: 6+; Inpatient stays: 2.1; Readmissions[A]: 1.2; Emergency room visits: 10. Source: GAO analysis of CMS data. GAO-14-523. Notes: This analysis used 2009 Medicare-Medicaid Linked Enrollee Analytic Data Source data. The analysis excluded beneficiaries who were not eligible for full Medicaid benefits; were under age 21; qualified for Medicare due to end-stage renal disease; lived outside the 50 United States and Washington, D.C.; did not have full Medicare coverage (Part A and Part B) for all eligible months; or were enrolled in Medicare Advantage for any part of the year. The analysis also excluded beneficiaries who lived in Maine or Arizona because we did not have complete data for these states. High-Medicare-expenditure beneficiaries represent those in the top Medicare spending quintile in their state. To determine beneficiaries' spending quintile, beneficiaries were ranked within their state of residence by their total Medicare spending and divided into quintiles. [A] Readmissions include admissions for any reason within 30 days of a hospital discharge. [End of figure] Fully Integrated D-SNPs Often Provided High Quality Care, but Had Limited Experience Serving Disabled Dual-Eligible Beneficiaries or Demonstrating Medicare Savings: Fully Integrated D-SNPs Often Met Criteria for High Quality, but Relatively Few of Those Plans Served Disabled Dual-Eligible Beneficiaries: FIDE-SNPs in 2013 were far more likely than other D-SNPs to meet criteria for high quality. Among those that reported sufficient data to receive a quality score, 14 (56 percent) of the 25 FIDE-SNPs met criteria for high quality but only 24 (14 percent) of all other 169 D- SNPs met these criteria.[Footnote 48] While FIDE-SNPs often received an overall quality score within the top two quintiles, 7 FIDE-SNPs (28 percent) scored below this mark, including 6 FIDE-SNPs that scored below the 40th percentile and 3 FIDE-SNPs that scored below the 25th percentile.[Footnote 49] The 14 high quality FIDE-SNPs operated across four states under programs through which all D-SNPs fully integrated Medicare and Medicaid benefits.[Footnote 50] In contrast, all 6 FIDE- SNPs that operated outside of these four states received a quality score below the 60th percentile.[Footnote 51] While 207 (64 percent) of the 323 D-SNPs in 2013 served disabled dual- eligible beneficiaries,[Footnote 52] only 13 (37 percent) of the 35 FIDE-SNPs[Footnote 53] served this population (see figure 7).[Footnote 54] The proportion of disabled dual-eligible beneficiaries in each of these 13 FIDE-SNPs ranged from 12 percent to nearly all of plan enrollment. These 13 D-SNPs operated across four states. Among the 22 FIDE-SNPs that did not serve disabled beneficiaries, most operated in state fully integrated programs that excluded this population from enrollment in those plans. Furthermore, only 2 FIDE-SNPs met criteria for high quality and served disabled beneficiaries.[Footnote 55] Figure 7: Dual-eligible Special Needs Plans (D-SNP) by High Quality Designation, Fully Integrated Dual-Eligible (FIDE) Designation, and Bid Relative to Medicare Fee-for-Service (FFS) Spending, 2013: [Refer to PDF for image: stacked horizontal bar graph] Number of D-SNPs: D-SNP characteristics: High quality D-SNP: Served disabled dual-eligible beneficiaries: 25; Other: 13; Total: 38. FIDE-SNP: Served disabled dual-eligible beneficiaries: 13; Other: 22; Total: 35. FIDE-SNP (by quality designation): High quality: Served disabled dual-eligible beneficiaries: 2; Other: 12; Total: 14. Nonhigh quality: Served disabled dual-eligible beneficiaries: 5; Other: 6; Total: 11. Unavailable quality score: Served disabled dual-eligible beneficiaries: 6; Other: 4; Total: 10. 2013 bid below Medicare FFS spending: FIDE-SNP: Served disabled dual-eligible beneficiaries: 3; Other: 5; Total: 8. FIDE-SNP and high quality: Served disabled dual-eligible beneficiaries: 0; Other: 3; Total: 3. Source: GAO analysis of CMS data. GAO-14-523. Notes: To designate D-SNPs as high quality, the 2012 performance year data were normalized across the 13 most common measures in the SNP Healthcare Effectiveness Data and Information Set Public Use Files. All D-SNPs with a high quality designation had an average performance (weighted higher for outcome measures used in CMS's star ratings) within the highest quintile among D-SNPs that reported data for all 13 measures. Those without an available quality score did not report data for all 13 measures. FIDE-SNPs were determined by CMS for 2013. Using CMS July 2011 enrollment data, we designated D-SNPs as having served disabled dual-eligible beneficiaries (under the age of 65) if (1) they enrolled at least 10 of those beneficiaries in July 2011 (the month most representative of annual enrollment according to CMS) and (2) those beneficiaries encompassed at least 5 percent of the D-SNP's enrollment at that time. D-SNPs' risk-adjusted bids--the capitated spending the plan requires to provide Medicare Part A and B services in its service area--were compared with CMS's risk-adjusted estimate of FFS spending at the county level and adjusted to plans' projected service area enrollment. FFS spending was also adjusted using a Sustainable Growth Rate correction factor provided by CMS. [End of figure] While 11 of the 21 highly integrated D-SNPs that CMS approved for 2013 benefits flexibility were high quality FIDE-SNPs, CMS's quality requirements did not prevent some D-SNPs with relatively low quality from being approved for benefits flexibility. Although CMS approved 18 of the 21 D-SNPs for benefits flexibility through their contract star rating, only D-SNPs without a star rating are assessed for quality at the plan level. Just 3 of the 18 D-SNPs approved through their contract star rating would have met CMS's plan-level quality requirements.[Footnote 56] Furthermore, 3 of the 18 D-SNPs approved for benefits flexibility through their star rating performed below the median quality score, including the only two D-SNPs approved with an overall star rating of 3.0. Relatively Few High Quality FIDE-SNPs Showed Potential for Medicare Savings, Regardless of Whether They Served Disabled Dual-Eligible Beneficiaries: Only 8 of the 35 FIDE-SNPs--and 3 of the 14 with high quality--bid below Medicare FFS spending in 2013, an indication that these plans can provide standard Medicare Part A and B benefits at a lower cost than what Medicare would have likely spent for these beneficiaries in FFS.[Footnote 57] Also, only 3 FIDE-SNPs that served disabled dual- eligible beneficiaries bid below Medicare FFS spending, and none of these met criteria for high quality.[Footnote 58] Among the 11 high quality FIDE-SNPs that bid at or above Medicare FFS spending, only 2 bid within 3 percentage points of FFS spending. On average (weighted by July 2013 enrollment), the 35 FIDE-SNPs bid 6 percent above FFS spending, and the 14 high quality FIDE-SNPs bid 3 percent above FFS spending.[Footnote 59] In contrast, D-SNPs without a FIDE designation bid 4 percent below FFS spending. Prior GAO research found that MA plans in service areas with high Medicare FFS spending were more likely to bid below Medicare FFS spending than MA plans in service areas with low FFS spending; FIDE- SNPs in 2013 generally followed that pattern.[Footnote 60] Of the eight FIDE-SNPs that bid below FFS spending, five operated in the 70th percentile or higher of MA service area FFS spending, two operated between the 60th and 70th percentiles, and only one operated below the median. In addition, among the 10 FIDE-SNPs that bid below FFS spending or had high quality and bid within 3 percentage points above FFS spending, 5 operated in the same service area as D-SNPs that were not fully integrated. In each of these 5 cases, total bids as a percentage of FFS spending were lower for D-SNPs with less integration of Medicare and Medicaid benefits. Furthermore, FIDE-SNPs were not notably more likely to bid below FFS spending based on CMS approval for benefits flexibility[Footnote 61], not-for-profit status, relatively smaller projected profit margin, or plan enrollment size above 1,000.[Footnote 62] Moderately Better Health Outcomes for Disabled Dual-Eligible Beneficiaries in D-SNPs Relative to Those in Traditional MA Plans Did Not Translate into Lower Levels of Costly Medicare Services: D-SNPs' performance for disabled dual-eligible beneficiaries relative to traditional MA plans' was similar on average for process measures, but was moderately better on health outcome measures.[Footnote 63] On average, D-SNPs' relative performance on 23 process measures (including screening for certain diseases and annual monitoring for patients on certain prescriptions) was 1 percentage point higher both for all beneficiaries and for those with six or more chronic conditions. While D-SNPs performed better on approximately two-thirds of the process measures,[Footnote 64] D-SNPs' relative performance varied substantially--ranging from 6 percentage points lower to 9 percentage points higher for all beneficiaries, and by an even wider range for those with six or more chronic conditions. In contrast, on average, D-SNPs' performance on seven health outcome measures (including maintaining healthy cholesterol, blood pressure, and blood sugar levels) was 5 percentage points higher for all beneficiaries and 7 percentage points higher for those with six or more chronic conditions. D-SNPs' relative performance was consistently better on each health outcome measure--ranging from 3 to 6 percentage points higher for all beneficiaries and 2 to 16 percentage points higher for those with six or more chronic conditions. (See table 2.) Table 2: Performance of Dual-eligible Special Needs Plans (D-SNP) Relative to Traditional Medicare Advantage (MA) Plans on Quality Measures for Disabled Dual-Eligible Beneficiaries, 2011: Process measures (N=23): Minimum difference; Percentage point difference between D-SNPs' and traditional MA plans' performance: All disabled: -6; Disabled with six or more chronic conditions: -12. Maximum difference; Percentage point difference between D-SNPs' and traditional MA plans' performance: All disabled: +9; Disabled with six or more chronic conditions: +13. Average difference; Percentage point difference between D-SNPs' and traditional MA plans' performance: All disabled: +1; Disabled with six or more chronic conditions: +1. Health outcome measures (N=7): Minimum difference; Percentage point difference between D-SNPs' and traditional MA plans' performance: All disabled: +3; Disabled with six or more chronic conditions: +2. Maximum difference; Percentage point difference between D-SNPs' and traditional MA plans' performance: All disabled: +6; Disabled with six or more chronic conditions: +16. Average difference; Percentage point difference between D-SNPs' and traditional MA plans' performance: All disabled: +5; Disabled with six or more chronic conditions: +7. Source: GAO analysis of CMS data. GAO-14-523. Notes: Positive numbers in the table indicate D-SNPs' performance was better by the indicated number of percentage points; performance on each process and outcome measure could range from 0 to 100 percent. The analysis excluded beneficiaries who were not eligible for full Medicaid benefits; were under age 18; qualified for Medicare due to end-stage renal disease; lived outside the 50 United States and Washington, D.C.; did not have full Medicare coverage (Part A and Part B) for the entire year; were not enrolled in MA for the entire year; or for whom data were missing or inconsistent. Chronic conditions are measured using the Hierarchical Conditions Categories CMS uses for risk-adjustment. Six percent of disabled dual-eligible beneficiaries in D-SNPs had six or more chronic conditions compared to 7 percent of those in traditional MA plans. We analyzed all 23 process measures and 7 intermediate health outcome measures in the Healthcare Effectiveness Data and Information Set (HEDIS) that are relevant to beneficiaries under age 65. We were not able to compare D-SNPs' and traditional MA plans' performance for disabled dual-eligible beneficiaries with six or more chronic conditions for 1 of the 23 process measures because there were fewer than 30 relevant beneficiaries. [End of table] However, D-SNPs' moderately better performance on health outcome measures did not translate into lower utilization levels of costly Medicare services for either all disabled dual-eligible beneficiaries or those with six or more chronic conditions. Among all disabled dual- eligible beneficiaries, those in D-SNPs had levels of inpatient stays, readmissions, and emergency room visits similar to those for beneficiaries in traditional MA plans. Among beneficiaries with six or more chronic conditions, those in D-SNPs had levels of inpatient stays and readmissions similar to those in traditional MA plans, but had somewhat higher levels of emergency room use--the percentage of beneficiaries in D-SNPs with zero emergency room visits was 4 percentage points lower. (See figure 8.) Figure 8: Utilization Levels of Selected Costly Medicare Services for Disabled Dual-Eligible Beneficiaries in Dual-eligible Special Needs Plans (D-SNP) and in Traditional Medicare Advantage (MA) Plans, 2011: [Refer to PDF for image: 4 horizontal bar graphs] All disabled: Inpatient stays[A]: 0; D-SNPs: 85%; Traditional MA plans: 84%. Inpatient stays[A]: 1; D-SNPs: 10%; Traditional MA plans: 10%. Inpatient stays[A]: 2; D-SNPs: 3%; Traditional MA plans: 3%. Inpatient stays[A]: 3; D-SNPs: 1%; Traditional MA plans: 1%. Inpatient stays[A]: 4; D-SNPs: 1%; Traditional MA plans: 1%. Inpatient stays[A]: 5+; D-SNPs: 1%; Traditional MA plans: 1%. Risk-adjusted readmission rate[B]: Rate: D-SNPs: 11%; Traditional MA plans: 11%. ER visits: Inpatient stays[A]: 0; D-SNPs: 60%; Traditional MA plans: 62%. Inpatient stays[A]: 1; D-SNPs: 19%; Traditional MA plans: 18%. Inpatient stays[A]: 2-4; D-SNPs: 16%; Traditional MA plans: 15%. Inpatient stays[A]: 5; D-SNPs: 5%; Traditional MA plans: 5%. Disabled with six or more chronic conditions: Inpatient stays[A]: 0; D-SNPs: 50%; Traditional MA plans: 50%. Inpatient stays[A]: 1; D-SNPs: 22%; Traditional MA plans: 21%. Inpatient stays[A]: 2; D-SNPs: 12%; Traditional MA plans: 12%. Inpatient stays[A]: 3; D-SNPs: 6%; Traditional MA plans: 7%. Inpatient stays[A]: 4; D-SNPs: 4%; Traditional MA plans: 4%. Inpatient stays[A]: 5+; D-SNPs: 7%; Traditional MA plans: 7%. Risk-adjusted readmission rate[B]: Rate: D-SNPs: 11%; Traditional MA plans: 12%. ER visits: Inpatient stays[A]: 0; D-SNPs: 36%; Traditional MA plans: 40%. Inpatient stays[A]: 1; D-SNPs: 21%; Traditional MA plans: 21%. Inpatient stays[A]: 2-4; D-SNPs: 27%; Traditional MA plans: 25%. Inpatient stays[A]: 5; D-SNPs: 17%; Traditional MA plans: 14%. Source: GAO analysis of CMS data. GAO-14-523. Note: Numbers in the figure show (a) the percentage of beneficiaries with the specified number of inpatient stays, (b) the risk-adjusted percentage of beneficiaries' admissions that resulted in readmissions, and (c) the percentage of beneficiaries with the specified number of emergency room (ER) visits. Percentages may not sum to 100 due to rounding. The analysis excluded beneficiaries who were not eligible for full Medicaid benefits; were under age 18; qualified for Medicare due to end-stage renal disease; lived outside the 50 United States and Washington, D.C.; did not have full Medicare coverage (Part A and Part B) for the entire year; were not enrolled in MA for the entire year; or for whom data were missing or inconsistent. Chronic conditions are measured using the Hierarchical Conditions Categories CMS uses for risk-adjustment. Six percent of disabled dual-eligible beneficiaries in D-SNPs had six or more chronic conditions compared to 7 percent of those in traditional MA plans. [A] Count of inpatient stays excludes those associated with principal diagnoses related to mental health or chemical dependency and those that did not result in a discharge from the hospital. [B] Denotes risk-adjusted percentage of beneficiaries' inpatient stays that resulted in a readmission--for any cause--within 30 days. Readmissions exclude those following maternity stays, non-acute stays (such as long-term care hospitalizations and rehabilitation stays), and same-day discharges. Readmission rate was risk-adjusted based on presence of surgeries, discharge condition, comorbidity, age, and gender. [End of figure] D-SNPs' performance relative to traditional MA plans on both process and health outcome measures was better for aged dual-eligible beneficiaries than it was for disabled dual-eligible beneficiaries; however, consistent with the results for disabled dual-eligible beneficiaries, D-SNPs' consistently better performance on health outcome measures for aged dual-eligible beneficiaries did not translate into lower levels of costly Medicare services. Among aged dual-eligible beneficiaries, D-SNPs' relative performance on process measures was 3 percentage points higher, on average (compared to 1 percentage point for disabled dual-eligible beneficiaries), and on health outcomes was 6 percentage points higher, on average (compared to 5 percentage points for disabled dual-eligible beneficiaries). However, aged dual-eligible beneficiaries in D-SNPs still had levels of inpatient stays, readmissions, and emergency room visits similar to such beneficiaries in traditional MA plans; for each measure we examined, D-SNPs' performance was within 2 percentage points of traditional MA plans' performance. Furthermore, this similarity was present even though there was a smaller percentage of aged dual- eligible beneficiaries in D-SNPs who had six or more chronic conditions (9 percent in D-SNPs vs. 13 percent in traditional MA plans). Among aged dual-eligible beneficiaries with six or more chronic conditions, D-SNPs' performance relative to traditional MA plans was 2 percentage points higher, on average, on process measures and 9 percentage points higher on outcome measures, yet this better performance also did not translate into lower utilization of costly Medicare-covered services. On the contrary, consistent with the results for disabled dual-eligible beneficiaries, D-SNPs had somewhat higher levels of ER use than traditional MA plans for aged dual- eligible beneficiaries with six or more chronic conditions. Looking within D-SNPs, aged dual-eligible beneficiaries in FIDE-SNPs performed 5 percentage points better on health outcomes than other D-SNPs, but this better performance did not translate into lower utilization of costly Medicare services.[Footnote 65] Concluding Observations: These results suggest that CMS's expectations regarding the extent to which integration of benefits will produce savings through lower use of costly Medicare services may be optimistic. Whether CMS and participating states will be able to improve quality without increasing overall program spending for disabled dual-eligible beneficiaries is uncertain. * While increasing the use of community-based LTSS may improve the quality of care for disabled dual-eligible beneficiaries who utilize those services, community-based LTSS users drove high combined program spending for disabled dual-eligible beneficiaries. Likewise, states with the highest per capita Medicaid spending for disabled dual- eligible beneficiaries were usually among the states with the highest overall program spending. In addition, the wide differences in health characteristics between disabled dual-eligible beneficiaries with the highest Medicare spending and those with the highest Medicaid spending may indicate the potential challenge of providing additional services without disproportionately impacting the costs of each program. Although most disabled dual-eligible beneficiaries with the highest Medicaid spending used community-based LTSS to assist them with activities of daily living, these beneficiaries generally did not have the numerous chronic conditions associated with those who had the highest Medicare spending. * Furthermore, if the models of care in the financial alignment demonstrations or other integrated models build on fully integrated D- SNP models, these efforts may improve the care provided to dual- eligible beneficiaries but may not produce significant Medicare savings for dual-eligible beneficiaries. D-SNPs that fully integrated Medicare and Medicaid benefits were far more likely than other D-SNPs to meet criteria for high quality but usually operated under fully integrated state programs that excluded disabled dual-eligible beneficiaries from enrollment. Regardless of whether they served disabled beneficiaries, high quality fully integrated D-SNPs did not usually demonstrate the potential for Medicare savings. In addition, many fully integrated D-SNPs that demonstrated the potential for Medicare savings operated in service areas where D-SNPs with less integration of benefits demonstrated more potential for Medicare savings. Our findings also suggest that even if there is moderate improvement in the performance of health outcome measures, and if dual-eligible beneficiaries are enrolled in plans specifically designed for them, instead of enrolled in traditional MA plans, these conditions are not necessarily sufficient to reduce disabled dual-eligible beneficiaries' use of costly Medicare services. Despite moderately better performance on health outcome measures for both disabled and aged dual-eligible beneficiaries, the fact that D-SNPs had similar levels of costly Medicare-covered services (i.e., inpatient admissions, readmissions, and emergency room visits) as traditional MA plans for this population has significant implications for program costs. Furthermore, for dual- eligible beneficiaries with six or more chronic conditions--a group that is at risk for high Medicare spending--although D-SNPs had better relative performance on health outcome measures, they still had similar, if not higher, levels of costly Medicare-covered services. Agency Comments: We provided a draft of this report to CMS for comment. CMS did not have any general comments. The agency provided technical comments, which we incorporated as appropriate. As agreed with your office, unless you publicly announce the contents of this report earlier, we plan no further distribution until 30 days from the report date. At that time, we will send copies to the Secretary of Health and Human Services, interested congressional committees, and others. The report also is available at no charge on GAO's website at [hyperlink, http://www.gao.gov]. If you or your staffs have any questions regarding this report, please contact me at (202) 512-7114 or cosgrovej@gao.gov. Contact points for our Offices of Congressional Relations and Public Affairs may be found on the last page of this report. GAO staff who made major contributions to this report are listed in appendix II. Signed by: James Cosgrove: Director, Health Care: [End of section] Appendix I: Disabled Dual-Eligible Beneficiaries' Service Use and Health Conditions by Expenditure Level, 2009: Medicare services: Service: 1+ inpatient stays; Medicare spending quintiles: Top: 63%; Middle: 11%; Bottom: 0%; Medicaid spending quintiles: Top: 22%; Middle: 23%; Bottom: 5%. Service: 1+ readmissions[C]; Medicare spending quintiles: Top: 31%; Middle: 4%; Bottom: 10%; Medicaid spending quintiles: Top: 25%; Middle: 21%; Bottom: 14%. Service: 1+ emergency room visits; Medicare spending quintiles: Top: 76%; Middle: 45%; Bottom: 20%; Medicaid spending quintiles: Top: 44%; Middle: 52%; Bottom: 29%. Service: 1+ primary care services; Medicare spending quintiles: Top: 79%; Middle: 71%; Bottom: 42%; Medicaid spending quintiles: Top: 76%; Middle: 68%; Bottom: 52%. Service: 1+ mental health services; Medicare spending quintiles: Top: 42%; Middle: 32%; Bottom: 9%; Medicaid spending quintiles: Top: 37%; Middle: 32%; Bottom: 14%. Medicaid services: Service: Any long-term services and supports (LTSS); Medicare spending quintiles: Top: 44%; Middle: 31%; Bottom: 22%; Medicaid spending quintiles: Top: 90%; Middle: 22%; Bottom: 2%. Service: Institutional nursing facility services; Medicare spending quintiles: Top: 13%; Middle: 4%; Bottom: 1%; Medicaid spending quintiles: Top: 18%; Middle: 2%; Bottom: 2%. Service: Intermediate care facility services[A]; Medicare spending quintiles: Top: 2%; Middle: 4%; Bottom: 2%; Medicaid spending quintiles: Top: 14%; Middle: 0; Bottom: 0. Service: Community-based services[B]; Medicare spending quintiles: Top: 29%; Middle: 23%; Bottom: 20%; Medicaid spending quintiles: Top: 59%; Middle: 20%; Bottom: 0. Health conditions: Service: 4+ chronic conditions; Medicare spending quintiles: Top: 59%; Middle: 24%; Bottom: 2%; Medicaid spending quintiles: Top: 29%; Middle: 31%; Bottom: 11%. Service: 6+ chronic conditions; Medicare spending quintiles: Top: 35%; Middle: 7%; Bottom: 0; Medicaid spending quintiles: Top: 14%; Middle: 13%; Bottom: 3%. Service: 3+ mental health conditions; Medicare spending quintiles: Top: 25%; Middle: 10%; Bottom: 2%; Medicaid spending quintiles: Top: 13%; Middle: 14%; Bottom: 4%. Source: GAO analysis of CMS data. GAO-14-523. Notes: This analysis used 2009 Medicare-Medicaid Linked Enrollee Analytic Data Source data and 2009 Medicare carrier claims. The analysis excluded beneficiaries who were not eligible for full Medicaid benefits; were under age 21; qualified for Medicare due to end-stage renal disease; lived outside the 50 United States and Washington, D.C.; did not have full Medicare coverage (Part A and Part B) for all eligible months; or were enrolled in Medicare Advantage for any part of the year. The analysis also excluded beneficiaries who lived in Maine or Arizona because we did not have complete data for these states. To determine beneficiaries' spending quintiles, beneficiaries were ranked within their state of residence by their total Medicare spending and their total Medicaid spending and divided into quintiles for each type of spending. The middle quintiles included beneficiaries in the 2nd, 3rd, and 4th quintiles. We categorized as LTSS users beneficiaries who used LTSS for at least 3 months during the year, then categorized them according to the type of LTSS that was used for the greatest number of months. Percentages for LTSS types--institutional nursing facility, intermediate care facility, or community-based--may not sum to the percentage for "any LTSS" due to rounding. [A] Readmissions are a subset of beneficiaries who had at least 1 inpatient stay and include admissions for any reason within 30 days of a hospital discharge. [B] Intermediate care facility for persons with intellectual disabilities. [C] Community-based LTSS include home health services, personal care services, and home and community-based services. [End of table] [End of section] Appendix II: GAO Contact and Staff Acknowledgments: GAO Contact: James Cosgrove, (202) 512-7114 or cosgrovej@gao.gov: Staff Acknowledgments: In addition to the contact named above, individuals making key contributions to this report include Catina Bradley, Assistant Director; Phyllis Thorburn, Assistant Director; Alison Binkowski; Aubrey Naffis; and Luis Serna III. Todd Anderson, Emily Johnston, Elizabeth T. Morrison, and Hemi Tewarson also provided valuable assistance. [End of section] Footnotes: [1] Medicare Payment Advisory Commission and Medicaid and CHIP Payment Access Commission, Data Book: Beneficiaries Dually Eligible for Medicare and Medicaid (Washington, D.C.: December 2013), 26. Medicare is the federal health insurance program for seniors, certain individuals with disabilities, and individuals with end-stage renal disease (ESRD). Medicaid is a joint federal-state program providing coverage of medical and health-related services for certain low-income individuals, such as children and individuals who are disabled or elderly. [2] Congressional Budget Office, Dual-Eligible Beneficiaries of Medicare and Medicaid: Characteristics, Health Care Spending, and Evolving Policies (Washington, D.C.: June 2013), 7. [3] In 2011, about 12 percent of disabled dual-eligible beneficiaries who qualified for full Medicaid benefits were enrolled in MA, and approximately 74 percent of these beneficiaries were enrolled in D- SNPs. [4] Dual-eligible beneficiaries may be enrolled in more than one Medicaid managed care plan and states may approve a range of plan types, such as plans that provide only mental health services or dental services. [5] States have allowed Medicaid beneficiaries to voluntarily enroll in a managed care plan, but more frequently, states are requiring these beneficiaries to enroll in managed care. [6] Medicaid does pay forsome of the beneficiary cost sharing associated with Medicare coverage of hospital services. [7] See Centers for Medicare & Medicaid Services, Cost Drivers for Dually Eligible Beneficiaries: Potentially Avoidable Hospitalizations from Nursing Facility, Skilled Nursing Facility, and Home and Community-Based Services Waiver Program (Washington, D.C.: August 2010); Mathematica Policy Research, Inc., Coordinating and Improving Care for Dual-Eligibles in Nursing Facilities: Current Obstacles and Pathways to Improvement (Princeton, N.J.: March 2010); and Congressional Budget Office, Dual-Eligible Beneficiaries of Medicare and Medicaid. [8] Only about one-third of the 2012 contracts contained any provisions for benefit integration, and only about one-fifth provided for active care coordination between D-SNPs and Medicaid agencies, which indicates that most care coordination was done exclusively by D- SNPs, without any involvement of state Medicaid agencies. See GAO, Medicare Special Needs Plans: CMS Should Improve Information Available about Dual-Eligible Plans' Performance, [hyperlink, http://www.gao.gov/products/GAO-12-864] (Washington, D.C.: Sept. 13, 2012). [9] FIDE SNPs must meet certain criteria, which include (i) operating under a CMS approved capitated contract with the State Medicaid agency that includes coverage of specified primary, acute, long term care benefits and services; and (ii) coordinating the delivery of Medicare and Medicaid LTSS and other services using aligned case management and specialty care network methods for high risk beneficiaries. [10] See Centers for Medicare & Medicaid Services, State Medicaid Director Memorandum, Memo Re: Financing Models to Support State Efforts to Integrate Care for Medicare-Medicaid Enrollees, SMDL # 11- 008 (Baltimore, Md.: July 8, 2011). [11] MMLEADS includes Medicare and Medicaid enrollment and claims data for beneficiaries who were enrolled in both programs. MMLEADS does not include Medicaid supplemental payments to providers that are made in addition to base payments for services provided to individual beneficiaries or Medicaid payments for Medicare premiums. [12] We ranked beneficiaries within their states by their total Medicare, Medicaid, and combined program spending and divided them into quintiles for each type of spending. We grouped beneficiaries in the second, third, and fourth quintiles into one middle group of beneficiaries. [13] MBSF includes data on enrollment, spending, and use of services for all Medicare beneficiaries. [14] LTSS are designed for beneficiaries who have limited ability to care for themselves; these services can be provided in institutions or within the community. Institutional LTSS include services provided in nursing facilities, intermediate care facilities for persons with intellectual disabilities, inpatient psychiatric facilities for individuals under age 21, and institutions for individuals age 65 or older with a mental illness. Community-based LTSS include home health services, personal care services, and other home and community-based services. We defined LTSS users as beneficiaries who used LTSS for at least 3 months during the year, and we categorized beneficiaries according to the type of LTSS--institutional or community-based--used for the greatest number of months. [15] Medicare carrier claims contain FFS claims from mostly noninstitutional providers (e.g., physicians) and free-standing facilities (e.g., independent clinical laboratories). [16] A readmission is generally defined as an admission to a hospital within 30 days of a discharge from the same or another hospital. We defined primary care services using guidance in CMS's Primary Care Incentive Payment Program and defined mental health services using guidance in CMS's Mental Health Services Billing Guide. [17] To calculate D-SNPs' performance on quality, we used quantile normalization to compare plan performance against 13 measures related to either effectiveness of care or readmissions in the SNP Healthcare Effectiveness Data and Information Set (HEDIS) Public Use Files for performance in 2012--the most recent data available. Each of the measures contained data for at least 75 percent of the D-SNPs in the Public Use Files. We calculated a quality score by using the weighted average performance of D-SNPs that reported data for all 13 measures (74 percent of D-SNPs with any data for these measures). We weighted measures according to CMS's star rating methodology, which recognizes 3 of the 13 measures as outcome-based and assigns those measures a weight of 3 to signify relative importance. All D-SNPs categorized as high quality performed better than the national average on at least 2 of the 3 outcome measures and the majority of any measures for which the D-SNP reported data. We ensured these D-SNPs were not in a contract under sanction by CMS. Our high-quality threshold aligned similarly with performance of 4 stars and above in CMS's star ratings. The lowest performing D-SNP in a one-plan contract that received at least 4 stars on CMS's 2014 Part C and overall ratings (which are mainly based on performance during 2012) received a quality score at the 82nd percentile. Furthermore, among D-SNPs in contracts with a star rating, 31 of the 35 D-SNPs that met our criteria for high quality also received an overall rating of 4 stars or higher. None of the D-SNPs that met our criteria for high quality received a rating below 3.5 stars. [18] For this study, we designated plans as having served disabled dual-eligible beneficiaries if (1) they enrolled at least 10 of those beneficiaries in July 2011 and (2) those beneficiaries encompassed at least 5 percent of the D-SNP's enrollment at that time. Twenty-nine of the 35 FIDE-SNPs operated in July 2011. [19] While risk scores for beneficiaries in MA plans tend to be higher relative to the risk scores of beneficiaries in Medicare FFS with the same health conditions and demographic characteristics, we did not reduce MA risk scores to account for this difference in diagnostic coding. Our estimate of MA bids relative to FFS spending is conservative. If we had made an adjustment for the difference in diagnostic coding, the standardized bid of MA plans and their relative costs to comparable spending in Medicare FFS would have increased. [20] In doing so, we assumed that Medicare physician fees would remain at 2012 levels. CMS provided an adjustment factor. [21] We defined traditional MA plans as those that were not a type of special needs plan (SNP) and had a plan type of local preferred provider organization, regional preferred provider organization, health maintenance organization (with or without point of service options), private FFS, or provider sponsored organization. [22] HEDIS measures are used by more than 90 percent of America's health plans, including MA plans, to measure performance on important dimensions of care and service. [23] The process, health outcome, and resource use measures were from the 2012 (2011 measurement year) HEDIS Patient Level Files--the most recently available data at the time we started our analysis. We analyzed all 23 process measures and 7 intermediate health outcome measures in HEDIS that are relevant to beneficiaries under age 65. Process measures assess whether health professionals deliver services according to evidence-based guidelines designed to prevent future adverse health outcomes. While process measures address whether or not a specific action was performed, the intermediate health outcome measures we analyzed assess the physiological result of such actions. [24] In 2010, Medicare beneficiaries with six or more chronic conditions had substantially higher utilization of costly Medicare services--and substantially higher per capita Medicare spending-- relative to those with fewer chronic conditions. See Centers for Medicare & Medicaid Services, Chronic Conditions among Medicare Beneficiaries, Chartbook, 2012 Edition (Baltimore, Md.: 2012). [25] In examining plans' performance among aged dual-eligible beneficiaries, we included an additional 7 process measures that are relevant to those aged 65 and older but not to beneficiaries under 65. We were unable to examine how D-SNPs' performance varied by FIDE designation for disabled dual-eligible beneficiaries due to the small number of FIDE-SNPs that served disabled dual-eligible beneficiaries and the concentrated enrollment within these plans. [26] The Social Security Administration defines disability as the inability to engage in any substantial gainful activity by reason of any medically determinable physical or mental impairment that can be expected to result in death or that has lasted or can be expected to last for a continuous period of at least 12 months. Individuals under age 65 may also qualify for Medicare on the basis of ESRD. ESRD is a medical condition in which individuals have permanent kidney failure that requires long-term dialysis or a kidney transplant to maintain life. [27] Individuals who are diagnosed with ESRD or who become disabled as a result of amyotrophic lateral sclerosis may become eligible for Medicare without a 24-month waiting period. [28] Dual-eligible beneficiaries may be categorized as full-benefit or partial-benefit. Those with full benefits may receive the entire range of Medicaid benefits; those with partial-benefits do not receive Medicaid-covered services, but Medicaid covers their Medicare premiums or cost-sharing, or both. Partial benefit dual-eligible beneficiaries have limited income and assets, but their income and assets are not low enough to qualify them for full Medicaid benefits in their state. In 2011, over 70 percent of disabled dual-eligible beneficiaries qualified as full benefit. [29] States may also cover personal care services and other types of home and community-based services. [30] Congressional Budget Office, Dual-Eligible Beneficiaries, 7. [31] Medicare Payment Advisory Commission, Report to the Congress: Medicare and the Health Care Delivery System (Washington, D.C.: June 13, 2013), 154-156. [32] D-SNPs, and other SNPs, have been reauthorized by Congress several times since they were first established in 2003. Current authorization of SNP programs expires in January 2017. [33] MCOs are allowed flexibility in designing various plan benefit packages. Although multiple MA plans may be included under a single contract between a managed care organization and CMS, the agency assesses quality at the contract level. As a result, every plan covered under the same contract receives the same star rating. [34] The Medicare Part D program provides voluntary, outpatient prescription drug coverage for eligible individuals. [35] Because the National Committee for Quality Assurance--the entity responsible for managing many of the performance measures for MA plans-- requires a minimum of 30 observations to validate performance measures, the 35 FIDE SNPs only include those with July 2013 enrollment of at least 30 beneficiaries. [36] D-SNPs may offer a range of benefits that include personal care in the home, nonskilled nursing care in the home, respite care for caregivers, in-home food delivery, and adult daycare services. D-SNPs must describe these benefits as part of their plan benefit packages at the time of bid submission and must offer the benefits without any cost-sharing or additional premium charges. [37] In 2014, CMS reported that only about 4 percent of MA contracts with prescription drug coverage (enrolling 1 percent of beneficiaries in these contracts) received an overall rating of less than 3 stars. In addition, only about 25 percent of these contracts (enrolling 17 percent of beneficiaries in these contracts) received an overall rating of 3 stars. In contrast, about 33 percent of these contracts (enrolling 30 percent of beneficiaries in these contracts) received an overall rating of 3.5 stars. [38] One of the 12 states with an approved demonstration uses an alternative financial alignment model. [39] States that participate in the demonstration typically use either a capitated model or a managed FFS model. Under the capitated model, health plans are responsible for delivering an integrated set of services for dual-eligible enrollees. Under the managed FFS model, states enter into an agreement with CMS to be eligible for savings resulting from initiatives that improve quality and reduce costs for Medicare. [40] The evidence of Medicaid savings due to rebalancing--increasing the proportion of LTSS provided in the home or community while reducing the proportion furnished in institutions--is limited, and study findings are mixed. See Medicare Payment Advisory Commission, Report to the Congress: Medicare and the Health Care Delivery System, 152. [41] Research has shown that while a small proportion of Medicare beneficiaries account for most of the program's costs and the biggest sources of spending for these high-cost beneficiaries are those related to emergency department visits and inpatient hospitalizations, only 10 percent of these costs were potentially preventable. Karen E. Joynt, Atul A. Gawande, E. John Orav, and Ashish K. Jha, "Contribution of Preventable Acute Care Spending to Total Spending for High-Cost Medicare Patients", Journal of the American Medical Association 309, vol. 309, no. 24 (2013): 2572-2578. [42] See Centers for Medicare & Medicaid Services, Joint Rate-Setting Process for the Capitated Financial Alignment Model (Baltimore, Md,: August 2013). CMS projects that approximately 60 to 70 percent of savings in the demonstration will come from fewer hospital admissions (including readmissions), and approximately 1 to 5 percent of savings will come from fewer ER visits. [43] Because we did not have data on Medicaid spending for LTSS alone, we examined total Medicaid spending for beneficiaries who used different types of LTSS. As a result, total Medicaid spending for community-based LTSS users may include Medicaid spending for services beyond this category of services. [44] We calculated correlation coefficients to measure the strength of the relationships between the various types of spending. We found that per capita Medicaid spending was much more strongly correlated with per capita combined spending than with per capita Medicare spending. [45] See Centers for Medicare & Medicaid Services, Joint Rate-Setting Process. [46] We calculated correlation coefficients to measure the strength of the relationships between the various types of spending. We found that per capita Medicaid spending for community-based LTSS users was much more strongly correlated with per capita combined spending than with per capita Medicare spending. Because CMS expects that integration of Medicare and Medicaid services will primarily result in a reduction in the use of high-cost Medicare services (see Centers for Medicare & Medicaid Services, Joint Rate-Setting Process), we also looked at the relationship between different types of spending for community-based LTSS users in the top Medicare spending quintile. We found that states with greater per capita Medicaid spending for these beneficiaries did not necessarily have less per capita Medicare spending but usually had greater per capita combined spending. [47] Previously, we found that hospital services and LTSS made up nearly 65 percent of total expenditures for high-expenditure Medicaid- only beneficiaries (those not dually eligible for Medicare). See GAO, Medicaid: Demographics and Service Usage of Certain High-Expenditure Beneficiaries, [hyperlink, http://www.gao.gov/products/GAO-14-176] (Washington, D.C.: Feb. 19, 2014). [48] In interviews with officials of selected D-SNPs, those operating high quality FIDE-SNPs were more likely than officials of other D-SNPs to report assigning care transition staff to hospitals and integrating financial incentives with providers. [49] Twenty-six of the 35 FIDE-SNPs operated under a not-for-profit MCO, including all 14 high quality FIDE-SNPs. [50] The four states administering these programs were Minnesota, Wisconsin, Massachusetts, and California. California's program is limited to D-SNPs operated by one MCO. Among FIDE-SNPs with a quality score, only 1 of those 19 that operated in the four states received a quality score below the 60TH percentile. [51] In interviews, officials of one high quality FIDE-SNP reported that the flexibility in benefits afforded by its state's fully integrated program was key to improving quality and reducing costs. [52] The 323 D-SNPs do not include those with operations exclusively outside of the 50 states and the District of Columbia and exclude plans with fewer than 30 beneficiaries. The total enrollment of the 207 D-SNPs that served disabled dual-eligible beneficiaries represented 82 percent of all 323 D-SNPs' enrollment in July 2013. In July 2011, disabled dual-eligible beneficiaries represented 39 percent of these plans' enrollment. [53] Among the 35 D-SNPs, the 13 that served disabled beneficiaries represented 21 percent of total enrollment. [54] We categorized six FIDE-SNPs as not serving disabled dual- eligible beneficiaries because they did not operate in July 2011. We confirmed that three of these six FIDE-SNPs did not serve disabled dual-eligible beneficiaries in 2013. Because the other three FIDE-SNPs operated in states where other FIDE-SNPS served disabled dual-eligible beneficiaries, these 3 D-SNPs may have enrolled disabled dual-eligible beneficiaries in 2013. [55] In interviews with officials of selected D-SNPs, some reported challenges with finding providers that could address some of disabled dual-eligible beneficiaries' needs. The challenges they cited included finding psychiatrists who are willing to make home visits, arranging inpatient beds for those with severe mental health issues, knowing which medications a beneficiary was taking prior to hospitalization, and finding physician specialist offices with beds that are equipped to handle physical limitations. [56] Fourteen D-SNPs approved for benefits flexibility through their contract star rating had sufficient data to assess whether they met CMS's plan-level quality requirement. [57] Because payment benchmarks have historically exceeded Medicare FFS spending, most plans that bid below FFS spending receive payments above that spending. Only 1 of the 323 D-SNPs--and 0 FIDE-SNPs--in our study bid at or above their payment benchmark. While only 24 percent of D-SNPs bid within 5 percent of their plan benchmark, nearly half (46 percent) of FIDE-SNPs bid within this range. [58] In interviews with officials of selected D-SNPs, some stated that current reimbursement and risk-adjustment may not properly account for disabled dual-eligible beneficiaries' needs. Officials of one D-SNP reported that these beneficiaries have a greater prevalence of transience, which can make care coordination more difficult and increase the resources used by the health plan without receiving additional reimbursement. Suggestions by officials to account for the additional risk associated with disabled dual-eligible beneficiaries included accounting for a wider range of mental health diagnoses, extreme levels of obesity, and social instability factors such as unemployment. Some research has shown that the prevalence of obesity and extreme obesity is significantly higher among individuals with disabilities. For those with and without disabilities, increased weight was generally associated with higher levels of chronic disease risk factors. Katherine Froehlich-Grobe, Jaehoon Lee, and Richard A. Washburn, "Disparities in Obesity and Related Conditions among Americans with Disabilities," American Journal of Preventive Medicine vol. 45, no. 1 (2013): 83-90. [59] While modest growth in MA payment benchmarks in recent years may have encouraged some MA plans to control their costs, 2014 MA bids relative to FFS spending were about the same--and may have increased for SNPs--compared with 2013 bids relative to FFS spending. Medicare Payment Advisory Commission, Report to the Congress: Medicare Payment Policy (Washington, D.C.: Mar. 14, 2014), 331-333. [60] See GAO, Medicare Advantage: Comparison of Plan Bids to Fee-for- Service Spending by Plan and Market Characteristics, [hyperlink, http://www.gao.gov/products/GAO-11-247R] (Washington, D.C.: Feb. 4, 2011). [61] Among the 21 highly integrated D-SNPs approved for benefits flexibility, only 3 bid below FFS spending. Weighting by July 2013 enrollment, the 21 highly integrated D-SNPs bid 4 percent above FFS spending. [62] Among all D-SNPs in our study, for-profit D-SNPs had higher profit margins but lower bids relative to FFS spending compared with not-for-profit D-SNPs, despite having similar Medicare FFS spending in their service areas. A prior GAO study found that D-SNPs reported higher profits, on average, compared with MA plans available to all beneficiaries in 2011. See GAO, Medicare Advantage: Special Needs Plans Were More Profitable, on Average, than Plans Available to All Beneficiaries in 2011, [hyperlink, http://www.gao.gov/products/GAO-14-210R] (Washington, D.C.: Dec. 19, 2013). However, the enrollment-weighted average of 2013 projected profit margins for FIDE-SNPs and other D-SNPs was below 5 percent. In addition, 26 of the 35 FIDE-SNPs projected profit margins at or below 4 percent. [63] The results in this finding do not control for potential differences in health status; however, (1) as we reported in [hyperlink, http://www.gao.gov/products/GAO-12-864], disabled dual- eligible beneficiaries in D-SNPs had risk scores similar to beneficiaries in traditional MA plans; (2) for the current report, we confirmed that, for the beneficiaries in our analysis, the distribution of the number of chronic conditions was similar for beneficiaries in D-SNPs and in traditional MA plans; and (3) we report comparisons for all beneficiaries and for just those with six or more chronic conditions. We analyzed all 23 process measures and 7 intermediate health outcome measures in HEDIS that are relevant to beneficiaries under age 65. [64] D-SNPs' performance for all beneficiaries was better on 15 of the 23 process measures: by less than 3 percentage points on 11 process measures, by 3 to 5 percentage points on 3 measures, and by more than 5 percentage points on 1 measure. D-SNPs' performance for beneficiaries with 6 or more chronic conditions was better on 14 of 22 process measures; we were not able to accurately compare the performance for one measure because there were fewer than 30 relevant beneficiaries. [65] Aged dual-eligible beneficiaries in FIDE-SNPs performed 9 percentage points better than those in traditional MA plans on health outcomes but did not have lower utilization of costly Medicare services. [End of section] GAO's Mission: The Government Accountability Office, the audit, evaluation, and investigative arm of Congress, exists to support Congress in meeting its constitutional responsibilities and to help improve the performance and accountability of the federal government for the American people. GAO examines the use of public funds; evaluates federal programs and policies; and provides analyses, recommendations, and other assistance to help Congress make informed oversight, policy, and funding decisions. GAO's commitment to good government is reflected in its core values of accountability, integrity, and reliability. Obtaining Copies of GAO Reports and Testimony: The fastest and easiest way to obtain copies of GAO documents at no cost is through GAO's website [hyperlink, http://www.gao.gov]. Each weekday afternoon, GAO posts on its website newly released reports, testimony, and correspondence. To have GAO e-mail you a list of newly posted products, go to [hyperlink, http://www.gao.gov] and select "E-mail Updates." Order by Phone: The price of each GAO publication reflects GAO's actual cost of production and distribution and depends on the number of pages in the publication and whether the publication is printed in color or black and white. Pricing and ordering information is posted on GAO's website, [hyperlink, http://www.gao.gov/ordering.htm]. Place orders by calling (202) 512-6000, toll free (866) 801-7077, or TDD (202) 512-2537. Orders may be paid for using American Express, Discover Card, MasterCard, Visa, check, or money order. Call for additional information. Connect with GAO: Connect with GAO on facebook, flickr, twitter, and YouTube. Subscribe to our RSS Feeds or E mail Updates. Listen to our Podcasts. Visit GAO on the web at [hyperlink, http://www.gao.gov]. To Report Fraud, Waste, and Abuse in Federal Programs: Contact: Website: [hyperlink, http://www.gao.gov/fraudnet/fraudnet.htm]; E-mail: fraudnet@gao.gov; Automated answering system: (800) 424-5454 or (202) 512-7470. Congressional Relations: Katherine Siggerud, Managing Director, siggerudk@gao.gov: (202) 512-4400: U.S. Government Accountability Office: 441 G Street NW, Room 7125: Washington, DC 20548. Public Affairs: Chuck Young, Managing Director, youngc1@gao.gov: (202) 512-4800: U.S. Government Accountability Office: 441 G Street NW, Room 7149: Washington, DC 20548. [End of document]