Evaluations of four state-run dually eligible demonstrations


Emily Boerger


Federal and state policymakers have been working to improve the cost-effectiveness of care provided to the “dually eligible” — those entitled to services under both Medicare and Medicaid — for more than two decades, but the results to date have been disappointing. The two giant public insurance programs are run by separate bureaucracies, have different cultures, and operate with different eligibility systems and financial incentives, all of which make coordination challenging. With so much at stake, it has proven difficult to find a replicable formula that all parties find acceptable.


Stay one step ahead. Join our email list for the latest news.



That’s regrettable because the payoff from a better system would be substantial. The  Medicare Payment Advisory Commission (MedPAC) and the Medicaid and CHIP Payment and Access Commission (MACPAC) estimate that the dually eligible account for 20 percent of Medicare’s enrollment but 34 percent of its spending. For Medicaid, the figures are 15 percent and 32 percent, respectively. In theory, the federal and state governments have an incentive to coordinate with each other to eliminate avoidable expenses. In practice, business-as-usual is the norm, which means the beneficiaries often get care in institutional settings that might have been avoided with more effective prevention and social services.

Since 2011, the Centers for Medicare and Medicaid Services (CMS) has sponsored tests of two models of federal-state coordination under its financial alignment initiative (FAI), with the federal government sharing savings from Medicare with states that implement effective reforms.

The first FAI model relies on a three-way contract between the federal government, the states, and managed care plans. The federal and state governments make capitated payments to the insurers to cover the full cost of care under both programs, with the expectation that the combined payment will be less than that the amount that would be spent in an unmanaged environment.

The other model retains fee-for-service as the basis for payment and allows states to contract with third parties, which may or may not be insurance plans, to steer beneficiaries to services that are cost-effective and may prevent more expensive care later. The beneficiaries retain the right to seek care from providers of their choosing.

Currently, eleven states have active FAI demonstrations underway (Colorado and Virginia terminated their programs at the end of 2017). In recent months, CMS has received, and released, the independent annual evaluations for four of them. The findings in these reports are illuminating.


  • California: The state is testing the full capitation model in selected regions. As of June 2019, the demonstration was serving nearly 107,000 dually eligible beneficiaries. Unfortunately, statistical assessments of the cost results show the program has increased spending for Medicare parts A and B relative to what would have occurred without the test. The evaluation does not cover the cost effects for Medicaid, or Medicare part D.


  • Illinois: The state began its full capitation program in March 2014 and contracts with seven plans. As of the end of 2019, the demonstration was serving more than 58,000 dually eligible beneficiaries. The most recent evaluation found that the test did not produce statistically significant savings in Medicare over its first three years.


  • Massachusetts: The state began its “One Care” demonstration program in October 2013 and contracts with three plans providing services in ten of the state’s fourteen counties. Massachusetts is the only state restricting its dually eligible test to beneficiaries age 21 to 64, which means the participants are entitled to Medicare because of a significant disability and not age. The most recent annual evaluation of One Care showed no statistically significant savings or costs in Medicare over the first four years of its implementation.


  • Washington: The state stands out because it has chosen to implement a health home model using fee-for-service rather than capitated payments. The dually eligible enrolled in the demonstration are assigned to one of six contractors. These organizations work with their assigned patients to steer them through the various medical and social services available to them. Unlike the results from the other states discussed above, Washington’s program has delivered consistent savings. The most recent evaluation estimates cumulative savings of $293 million in Medicare over six years.


While Washington’s results are encouraging, the overall picture is one of significant remaining challenges across the country. The demonstrations underway were never intended to be comprehensive reforms. Even so, it is striking how small they are relative to the size of the eligible population. After many years of discussion and debate, the reality is that the vast majority of the dually eligible remain in unmanaged settings and receive services from providers who have weak incentives for cost avoidance. The result is that the federal government and the states are spending far more than is necessary, or affordable, and the beneficiaries often receive inferior care too.

After years of trial and error, the states and the federal government have a better handle today than in the past on what will work, and what won’t. Even so, a full solution would seem to be some distance away. Faster progress may require another intervention from Congress, with a mandate for a fully integrated benefit administered by the states or the federal government. Another push in national legislation may be the only way to get all of the relevant players on the same page.

James C. Capretta is a columnist for State of Reform and a senior fellow at the American Enterprise Institute.