Abstract
- Issue: The Financial Alignment Initiative (FAI) aims to improve care for individuals dually eligible for Medicare and Medicaid. Recent interim evaluations of the FAI offer important new information about the successes and limitations of these models.
- Goal: Assess the findings of recent evaluations from five states to identify initial themes and lessons from the Initiative.
- Methods: Review of RTI International evaluations of the FAI in California, Illinois, Massachusetts, Ohio, and Texas.
- Key Findings: Medicare cost saving results under the FAI are mixed. Despite the model’s emphasis on care coordination, few beneficiaries reported receiving care coordination. Implementation challenges suggest the need for adequate resources and timelines. The FAI evaluations were limited by a lack of Medicaid data, particularly in the calculation of cost-savings and in assessing service utilization impacts.
- Conclusion: Efforts to integrate services and financing for dual-eligible enrollees should: ensure there are sufficient resources to support implementation by state governments and other stakeholders; include robust consumer engagement through design, implementation, and oversight stages; watch capacity and performance of plans and providers to meet the needs of a complex population; and have the ability to make midcycle corrections.
The Role of the Financial Alignment Initiative in Improving Care for Dual-Eligible Beneficiaries
Nearly 12 million low-income older adults and people with disabilities qualify for both Medicare and Medicaid. Compared to the Medicare population as a whole, individuals who are dually eligible have more significant health and functional support needs.1 At the same time, dual-eligibles face significant challenges in receiving coordinated care that is aligned with their needs. Medicare and Medicaid operate as separate programs, with Medicare the primary payer for physician visits, hospital stays, postacute skilled care, and prescription drugs. State Medicaid programs augment this coverage by providing financial assistance with Medicare premiums and cost-sharing, as well as covering additional benefits, notably long-term services and supports. Different coverage and payment policies between the two programs create incentives to shift costs between the states and the federal government, resulting in underuse of some services and overuse of others. Lack of coordination and integration leads to fragmented care delivery, high costs, and poor outcomes.
The Affordable Care Act created the Medicare–Medicaid Coordination Office (MMCO) to increase coordination between Medicare and Medicaid. In 2011, MMCO launched the Financial Alignment Initiative (FAI) to test new approaches to aligning incentives between Medicare and Medicaid and to improve care for the dually eligible population. Thirteen states chose to participate in the FAI. Ten participated in a capitated model, in which Medicare and Medicaid benefits are delivered through a single managed care plan. Three opted for a fee-for-service model, which did not fully integrate payments or benefits but allowed states to share in savings that resulted from quality and cost initiatives. In this issue brief, we review and analyze results from evaluation reports of five states that participated in a capitated model: California, Illinois, Massachusetts, Ohio, and Texas.2 (See “How We Conducted This Study.”)
Preliminary Findings from Financial Alignment Initiative Evaluations
Program Implementation: Growing Pains Are Common
Administering integrated benefits involved significant logistical challenges and operational changes for states, the federal government, plans, and providers. Stakeholders from state and federal governments, participating plans, and beneficiary groups interfaced with separate Medicare and Medicaid systems for eligibility, beneficiary notifications, provider payments, and complaint resolution. Solutions often involved manual work-arounds, especially in launch years. In some states, providers of long-term services and supports or behavioral health services had never submitted claims to managed care plans before, which created additional implementation challenges. Managed care plans’ lack of experience in working with certain providers, such as nursing homes in Ohio and California, led to significant difficulties for beneficiaries. State officials also noted the learning curve in understanding Medicare practices and procedures. In Texas, the state agency responsible for implementing the demonstration was restructured during the demonstration period, resulting in shifting staff roles and changes in oversight responsibility.
Beneficiary Experience: Ongoing Focus on Care Coordination Needed
The FAI focused on improving care coordination. However, much of the responsibility for ongoing benefit coordination falls to individual care coordinators, with variation in caseloads and available resources and tools. In focus groups, members reported high rates of turnover among care coordinators. A state official in Ohio said, “[I]t’s tough when we expect care managers to be a central point of integration. . . . It’s difficult to find staff that can adequately convey and understand the different systems across the entire spectrum and understand long-term care, behavioral health, and medical care, and be able to work on behalf of the individual in all of those spaces.”
Despite the focus on care coordination in the demonstrations, only a minority of demonstration participants reported receiving care coordination (Exhibit 1). Across demonstrations, the two plans with the lowest percentages of receiving care coordination in 2015 were in Illinois; they improved their performance in 2016. The two plans with the highest percentages in 2015 were in Massachusetts; in 2016, one of the plans left the program and the other declined slightly.
Across the demonstration, satisfaction surveys suggest that members tended to rate their plans highly, at levels comparable to other Medicare Advantage plans. However, beneficiaries can more readily leave demonstration plans compared with other Medicare Advantage plans if they are unsatisfied with care. This may result in higher satisfaction scores among those members who do choose to remain in the demonstrations.
Service Utilization: Limited Data with Mixed Findings
The evaluations compare changes in Medicare spending and service use in the demonstrations to those occurring in comparison groups using a difference-in-differences approach that assumes the demonstration and comparison groups would have had similar changes in health care use and spending if the demonstration had not taken place. The demonstrations occurred in a context of declining inpatient use across the entire Medicare population, including dual-eligible beneficiaries.3
Evaluations of plans in Illinois, Massachusetts, and Ohio provide data on Medicare service use. In Illinois and Ohio, inpatient use decreased more among dual-eligible beneficiaries who are eligible to participate in the demonstration than in the comparison group. In the first three years of the Massachusetts demonstration, inpatient use declined among the demonstration-eligible population but decreased even more among the comparison group. These findings raise questions about whether long-term analyses of demonstration plans will show reduced inpatient use relative to comparison groups.
Because little Medicaid data is available, the evaluations report only limited information on use of long-term services and supports. However, it will be critical to evaluate this area because of the unique needs and high costs of long-term services users and the opportunity to cut costs by substituting supportive services for more costly acute care. Nursing facilities are also important in influencing enrollment: evaluations described cases in which a nursing home’s entire dual-eligible population opted out on the same day. Based on the data available so far, the Ohio and Massachusetts demonstrations are associated with less use of long-term nursing home care and Illinois with more.
Cost Savings: Impact on Medicare Costs Vary Across States
Cost-savings calculations were only provided for Medicare services. Results should be interpreted in the context of low overall rates of Medicare per-beneficiary spending growth in the past decade. Relative to the comparison groups, the demonstration groups showed increases in Medicare spending in some states and decreases in others (Exhibit 2). Medicare spending in the period before the demonstration also varied across states, with a low of $907 in Massachusetts and a high of $1,494 in Texas. The variation in spending in the period before the demonstration could reflect the fact that the demonstrations included different populations across states: for example, Massachusetts’ demonstration was limited to dual-eligibles under the age of 65, which could contribute to the lower baseline Medicare expenditures.
Strategies for Strengthening Integrated Care
Despite the modest results, the FAI provides valuable lessons for policymakers interested in integration and gives states and plans opportunities to test out strategies for strengthening integrated care programs. States and plans used a range of strategies to address the challenges of care integration (Exhibit 3).
Conclusion
The evaluations to date are hindered by a lack of Medicaid service utilization and expenditure data; policymakers should accelerate efforts to improve the availability of Medicaid data to make timely evaluation of current and future initiatives possible. Policymakers will need to focus on plans’ ability to deliver on care coordination services, which will require sufficient workforce, appropriate training and resources, and close monitoring of performance. A robust quality strategy that encompasses person-centered outcomes, such as quality of life and community integration, should be developed and deployed. Building these models to scale will require policymakers and plans to focus on communicating and engaging with beneficiaries and providers, and a robust enrollment policy. State governments also have operational and capacity issues that may determine whether they can successfully take advantage of integration opportunities. Technical assistance and funding support to state officials as well as state-based stakeholders, provided by the federal government, philanthropy, or other sources will be critical for successful adoption. The most important design feature of all may be a continuous quality improvement approach and mind-set that ensures real-time feedback and the agility to make rapid, midcycle corrections.
Limitations
This issue brief focuses on the FAI evaluation reports from California, Illinois, Massachusetts, Ohio, and Texas, which all utilized a capitated demonstration model. We did not include Colorado, Washington (both managed fee-for-service demonstrations) or Minnesota (administrative alignment demonstration).
The FAI evaluations provide information about program implementation, beneficiaries’ experiences, service utilization, and cost savings calculations. They draw on a range of data sources, including stakeholder interviews, beneficiary focus groups and surveys, demonstration enrollment, complaints and appeals, service utilization, capitation payments, and Medicare claims. Evaluations from Illinois, Massachusetts, and Ohio include service utilization data for Medicare services; the evaluation of Massachusetts included service utilization data for Medicaid. None of the reviewed evaluation reports included information about Medicaid expenditures, which is a limitation for understanding the impact of the FAI.
In addition, there was lower-than-expected enrollment across all demonstrations. When comparing changes in Medicare spending and service use in the demonstrations to changes in comparison groups in other counties or states, the evaluations used an intention-to-treat analysis, meaning that the evaluation looked for changes among all individuals eligible for the demonstration, whether or not they actually participated. Even if demonstration enrollees benefited from participating in the program, the effect is difficult to detect because there were relatively few demonstration enrollees among the eligible population.
In addition, four of the state evaluations assessed demonstrations in their launch year; Massachusetts was evaluated after three years. Given a significant time lag to publication, some of the challenges identified in the evaluations may have been attenuated since they were published.
Finally, quality measures in this set of evaluations were limited. For efforts at integration to be successful, they should achieve more than simply preventing negative outcomes such as avoidable nursing home stays or hospitalizations, and instead drive meaningful improvement in patient outcomes, particularly those related to quality of life and community integration.4 The measures of quality used in the evaluations must be strengthened to detect person-centered improvements in care quality and outcomes.
How We Conducted This Study
The authors reviewed the RTI International evaluations of the Financial Alignment Initiative (FAI) in California, Illinois, Massachusetts, Ohio, and Texas. The Centers for Medicare and Medicaid Services (CMS) contracted with RTI International to evaluate the impact of the FAI on beneficiary experience, quality, service utilization, and cost. In November 2018 and May 2019, CMS published interim state-specific evaluation reports for eight states, of which five participated in the capitated model (California, Illinois, Massachusetts, Ohio, and Texas). These evaluation reports detail important new information about the performance of the FAI, and more broadly, the opportunities and challenges for better integrating care for this population, which are particularly timely given new efforts to expand integrated care for dual-eligible enrollees.5
Acknowledgments
The authors would like to acknowledge Marc Cohen, Ph.D., David Grabowski, Ph.D., and David Stevenson, Ph.D., for their review and input.