The FAI Is Dead

The Financial Alignment Initiative ended on December 31, 2025, after more than a decade of testing whether integrated Medicare-Medicaid financing could improve care and reduce costs for dual eligible beneficiaries. Authorized under Section 3021 of the Affordable Care Act and administered through CMMI, the FAI launched in 2013 as the federal government’s most ambitious attempt to solve the structural problem at the center of dual eligible care: two separate programs, two separate payment streams, two separate regulatory frameworks, and no single entity accountable for the whole person.

Thirteen states participated at the demonstration’s peak. Ten tested a capitated model through Medicare-Medicaid Plans, which operated under three-way contracts among CMS, the state Medicaid agency, and the health plan. The MMP received a blended Medicare-Medicaid capitation payment and was responsible for all medical, behavioral health, and long-term services and supports for its enrolled dual eligible population. Two states tested a managed fee-for-service model, and Minnesota operated an alternative model that aligned administrative processes within its existing Senior Health Options program. At enrollment’s height, approximately 470,000 beneficiaries participated. By the time the demonstration ended, eight states remained in the capitated model, having extended through multiple rounds of federal authorization. Colorado and Virginia had exited in 2017. New York’s FIDA demonstration ended in 2019. California transitioned out in 2023.

The FAI is gone. The population it served remains. The question is whether the replacement vehicles, primarily FIDE SNPs and to a lesser extent PACE, can deliver what the demonstration achieved in its strongest markets while avoiding what failed in its weakest ones.

What the FAI Tested

#

The capitated model’s defining feature was the three-way contract. Unlike D-SNPs, which hold separate contracts with CMS for Medicare and with the state for Medicaid coordination, MMPs operated under a single agreement that specified obligations across both programs. The three-way contract included provisions for integrated member materials, unified grievance and appeals processes, integrated health risk assessments, and state-specific care coordination requirements. A Contract Management Team, with representatives from the state Medicaid agency, CMS regional offices, and the Medicare-Medicaid Coordination Office, provided ongoing joint oversight.

The blended capitation was the financial mechanism that made integration operational. CMS and the state each contributed their respective per-member payments into a single capitated rate paid to the MMP. The plan bore risk for the full spectrum of services: primary and specialty medical care, inpatient and outpatient hospital services, behavioral health, pharmacy, and crucially, long-term services and supports including home and community-based services, personal care, and nursing facility care. That full-risk arrangement created incentives the fragmented system lacks. An MMP that invested in care coordination to prevent hospitalizations captured the Medicare savings. An MMP that invested in home-based services to prevent nursing home admissions captured the Medicaid savings. In the fragmented system, neither program captures the return on the other’s investment.

The demonstrations operated under waiver authority that gave them flexibility unavailable to D-SNPs operating within the standard MA regulatory framework. States could design custom benefit packages, implement passive enrollment with opt-out provisions, require care coordination staffing ratios, and mandate specific LTSS delivery models. The regulatory freedom was significant, but it was also temporary. Demonstration authority expires. The structures built under it do not automatically convert to permanent programs.

The political trajectory of the FAI tracked the broader arc of dual eligible policy across three administrations. The Obama administration launched the initiative and approved the initial state demonstrations. The first Trump administration continued the demonstrations without expansion, allowing extensions but signaling no interest in new enrollments. The Biden administration invested in the transition framework, finalizing the CY 2023 rule that established the pathway for MMP-to-D-SNP conversion and tightening D-SNP integration requirements. The second Trump administration let the demonstrations end on schedule, betting that the permanent D-SNP regulatory framework is sufficient to carry integration forward without ongoing demonstration authority.

What the FAI Achieved

#

RTI International, CMS’s contracted evaluator, conducted multi-year assessments across participating states. The findings were uneven, which is both the honest summary and the reason the FAI was never certified for expansion under CMMI’s statutory standard.

In Massachusetts, the One Care program demonstrated measurable improvements in care coordination and beneficiary experience. Enrollees reported higher satisfaction with care management services and greater ease navigating the system compared to dual eligibles in non-demonstration areas. Massachusetts invested heavily in care coordination staffing, required MMPs to employ community health workers, and built integration infrastructure that connected medical, behavioral health, and LTSS providers within a shared care planning framework.

Ohio’s MyCare program showed similar strengths in care coordination processes, though the cost and utilization outcomes were more mixed. The demonstration achieved operational integration in a state with a complex Medicaid managed care landscape, and the evaluation identified improvements in beneficiary-reported measures of care coordination and access to services.

Other states struggled with enrollment, provider participation, or both. Passive enrollment with opt-out provisions, while necessary to achieve scale, generated opposition from beneficiaries and advocates who viewed the default enrollment mechanism as coercive. Several states experienced high opt-out rates that limited the demonstration’s enrolled population below the scale needed for meaningful evaluation. Provider participation varied by market; in areas where dominant health systems declined to contract with MMPs, beneficiaries faced network adequacy problems that undermined the integration promise.

The cost findings were the most consequential for the FAI’s legacy. CMMI’s statutory standard requires that a model reduce Medicare spending without reducing quality, or improve quality without increasing spending. The FAI evaluations did not produce consistent evidence of net savings across the capitated model states. Individual states showed savings in specific service categories, particularly reductions in inpatient utilization and skilled nursing facility admissions in markets with strong care coordination, but the aggregate picture was not sufficiently compelling to meet the certification threshold.

The lesson from the FAI is not that integration failed. It is that integration succeeded in markets where state Medicaid agencies invested in infrastructure, where MMPs hired and retained care coordination staff, where providers participated in integrated networks, and where the enrolled population was large enough for the care management model to operate at scale. Integration underperformed in markets where any of those conditions was absent. The question for the post-FAI era is whether the permanent regulatory framework can replicate the conditions that produced the strongest results.

What Fills the Gap

#

FIDE SNPs are now the primary vehicle for dual eligible integration. The transition was deliberate. CMS used the CY 2023, 2025, and 2026 rulemaking cycles to progressively tighten D-SNP integration requirements, pushing the regulatory floor for D-SNPs upward toward the integration level the FAI demonstrated.

The key regulatory milestones include exclusively aligned enrollment for all FIDE SNPs starting in 2025, meaning the D-SNP and the Medicaid managed care organization must share a parent company and the member must receive both Medicare and Medicaid through that single organization. The CY 2026 rule added requirements for integrated identification cards and integrated health risk assessments by 2027. The CY 2027 proposed rule, at §422.514(h), would require same-parent MCO alignment for all D-SNPs, with existing unaligned enrollees grandfathered through 2030.

Eight FAI states transitioned their MMPs directly into D-SNPs at the demonstration’s end. Illinois, Massachusetts, Ohio, and Rhode Island moved into FIDE SNPs. Michigan, South Carolina, and Texas transitioned into HIDE SNPs. The transition preserved enrollment relationships and, in most cases, the same health plans that operated the MMPs now operate the successor D-SNPs.

The transition from MMP to D-SNP is not a lateral move. Several features of the FAI that beneficiaries, advocates, and states valued are not automatically carried over into the D-SNP framework. The three-way contract’s joint oversight mechanism, with its Contract Management Team and regular CMS-state-plan meetings, has no direct analogue in D-SNP operations. The unified grievance and appeals process, which applied both Medicare and Medicaid standards to every coverage dispute, is not standard in D-SNPs unless the state negotiates it into the State Medicaid Agency Contract. The supplemental benefits that states authorized under demonstration flexibility may not survive unless the plan can fund them through MA bid economics or the state authorizes them through a Section 1115 waiver, as Massachusetts has done.

Massachusetts filed a Section 1115 waiver amendment to allow its successor FIDE SNPs to continue providing community-based services that were available under One Care but fall outside standard Medicaid state plan authority. That waiver represents the kind of state-level innovation needed to preserve FAI-era integration gains, and it also illustrates the administrative complexity involved. Not every state will pursue a waiver. Not every state has the policy infrastructure to design one.

AHEAD, the geographic total-cost-of-care model operating in eight states with hospital participation, intersects with dual eligible populations in participating markets. AHEAD holds participating hospitals accountable for total cost of care across their attributed Medicare population, which includes dual eligibles. In states where AHEAD operates alongside FIDE SNPs, the model creates a complementary accountability structure: the FIDE SNP coordinates care for the dual eligible, and the AHEAD hospital bears cost risk for the same beneficiary’s acute care utilization. Whether this layered accountability produces better outcomes or creates coordination friction between the plan and the hospital is an empirical question the first performance years will begin to answer.

PACE, the Program of All-Inclusive Care for the Elderly, remains the only model that provides fully integrated Medicare-Medicaid capitation under a single entity for community-dwelling, nursing-home-eligible adults. With approximately 68,000 participants in roughly 170 programs across 32 states, PACE operates at a scale far smaller than the D-SNP market. The FAI’s end does not directly expand PACE’s role, but it eliminates the only other model that offered comparable integration depth. In markets where FIDE SNPs are not yet available and the FAI demonstration has ended, PACE is the sole remaining option for a dual eligible seeking fully integrated care under a single organization.

The State Decision Framework

#

The post-FAI landscape puts state Medicaid agencies in the position of primary decision-maker for dual eligible integration strategy. CMS sets the regulatory floor through D-SNP requirements, but the state determines which integration tier its D-SNPs will operate at, which plans receive State Medicaid Agency Contracts, whether to pursue FIDE or HIDE as the target integration level, and whether to invest in the infrastructure that makes integration operational rather than nominal.

States that participated in the FAI have a head start. They have existing plan relationships, established care coordination protocols, trained state staff who understand Medicare-Medicaid interaction, and beneficiaries who have experienced integrated care and can articulate what they need from the successor model. These states are best positioned to lead the next phase of integration.

States that did not participate in the FAI and do not currently offer FIDE SNPs face a longer path. Building FIDE SNP infrastructure requires Medicaid managed care contracting changes, plan readiness assessments, state staff training on Medicare regulations, and beneficiary education. It requires the state Medicaid agency to engage with CMS regional offices on a regular basis, a relationship that the FAI’s Contract Management Teams facilitated but that has no automatic successor mechanism.

The simultaneous implementation of work requirements creates a resource competition that will slow integration progress in many states. The same state Medicaid agency that must build work requirement verification systems by January 2027 is also expected to strengthen D-SNP integration, negotiate State Medicaid Agency Contracts with higher integration standards, and prepare for the CY 2027 same-parent alignment requirement. For states with limited Medicaid agency capacity, the binding constraint is not policy ambition. It is implementation bandwidth.

Related Reading

#

MCR-01_07 LEAD and ASM: New Pathways for ACOs and Specialists MCR-11_07 New York and Illinois: High-Cost, High-Regulation, High-Unmet-Need