Mandatory or Irrelevant: CMMI’s Last Chance to Fix Value-Based Care

By Elena Pak, Credentialing Department, WCH

The CMS Innovation Center — created in 2010 with a $10 billion mandate to prove that paying for value could save Medicare money — has a credibility problem. Of 49 models launched in its first decade, only six generated statistically significant savings, and just four were certified for expansion. Now, under CMS Administrator Dr. Mehmet Oz and CMMI Director Abe Sutton, the center is doubling down on the one tool that theory and evidence suggest could actually work: mandatory participation. The message delivered at the March 2026 Accountable for Health summit was clear and deliberate — voluntary models alone cannot fix a fee-for-service system that has proven remarkably resistant to change. Whether mandatory models can is the central unresolved question in American healthcare payment policy.

Key Takeaways

1. The voluntary model problem is structural, not incidental. When providers can choose whether to participate, those most likely to succeed do. Those least likely to succeed — the ones whose patients most need better care coordination — stay on the sidelines. This selection bias doesn’t just skew the results. It makes the model’s savings ungeneralizable to the broader system. Mandatory models eliminate the problem by design.

2. CMMI proposed more mandatory models in 2025 than in any prior year. Three mandatory models in a single year — one-third of all proposals — represents a sharp directional signal. The TEAM model for hospitals launched January 1, 2026. The Ambulatory Specialty Model for physician specialists launches in 2027. The administration is moving from rhetoric to implementation.

3. Only 4 out of ~50 models have been certified for expansion. That is not a rounding error — it is a structural indictment of how CMMI has operated. The center’s own director acknowledged it while defending the high certification bar. Providers and health systems that have invested heavily in model participation only to see models expire or thresholds shift have legitimate grievances. Trust in CMMI’s staying power is a real participation barrier.

4. Fee-for-service entrenchment is the enemy — and the data confirms it. Despite fifteen years of effort, most Medicare payment still flows through fee-for-service. Voluntary models can succeed if benchmarks and incentives are crafted correctly — but this has proven very difficult, particularly with prospective benchmarking. Mandatory models force the experiment to run on the full population, not just early adopters.

5. The new model portfolio reflects political priorities as much as clinical ones. MAHA ELEVATE (lifestyle medicine), GLOBE and GUARD (drug pricing), ACCESS (tech-enabled chronic care) — these are not value-neutral technical designs. They embed specific theories of what drives cost and quality. Providers applying to new models should read the political subtext, not just the payment methodology.

6. The question for providers is no longer whether to engage with value-based models — it’s which ones, and how fast. CMS is signaling that mandatory participation is coming to more specialties and settings. Waiting for the landscape to stabilize is itself a strategic choice — and increasingly a risky one.

Why CMMI Exists — and Why It Has Struggled

The Center for Medicare and Medicaid Innovation was born from a specific frustration: Congress knew the fee-for-service payment system was driving overutilization and escalating costs, but didn’t know precisely how to fix it. The Innovation Center received $10 billion in funding for its first decade, with another $10 billion for each decade thereafter, specifically to test new payment models before potentially scaling them system-wide. The logic was sound — run controlled experiments, find what works, expand what saves money.

The center was initially expected to reduce net spending by $2.8 billion between 2011 and 2020. The Congressional Budget Office had further estimated in 2016 that CMMI would save $34 billion between 2017 and 2026. Neither projection materialized. Of 49 models with published evaluations from the first decade, six generated statistically significant savings, and only four were certified for expansion by CMS’s chief actuary.

That track record is the context for everything the current administration is doing. Oz called CMMI a “nuclear reactor of good ideas” and a “remarkable treasure trove for our government” — but the praise was accompanied by an implicit diagnosis: the reactor has been running below capacity, and the fuel mix needs to change.

The Voluntary Model Trap

To understand why CMMI is pivoting toward mandatory participation, it helps to understand exactly why voluntary models have underperformed.

Mandatory models eliminate the favorable selection problem whereby participants choose only models they are likely to profit from. The true potential of a model to achieve program-wide savings is unclear if the participants are functionally limited to only those who are best equipped to achieve those savings.

The mechanism is straightforward. A voluntary ACO model attracts providers who are already efficient, already coordinating care well, and already positioned to generate savings against a benchmark. Providers who are wasteful, poorly coordinated, or serving complex populations — the providers whose transformation would generate the largest system-wide savings — have little incentive to join a model that exposes them to financial downside risk. So they don’t.

The result is a self-fulfilling prophecy: voluntary models produce modest, hard-to-generalize savings among a self-selected group of high performers, while the bulk of Medicare spending remains untouched in fee-for-service. Despite CMS’s stated goal of having all Medicare beneficiaries in an accountable care relationship by 2030, approximately 50% of providers remain on the value-based care sidelines. After fifteen years of voluntary model design, that number should give everyone pause.

Mandatory models do require the Innovation Center to go through rulemaking — which significantly delays implementation, and comments from entrenched interests can impact design. They also generate substantially more industry resistance. But as CMMI Director Sutton acknowledged directly: “Mandatory models are going to have to be part of the equation.”

What’s Actually Launching — and What It Means

The current CMMI pipeline is the most active in the center’s history, and the design choices reveal clear priorities.

The TEAM (Transforming Episode Accountability Model) began January 1, 2026 — a mandatory episode-based payment model for hospitals in selected markets. The Ambulatory Specialty Model (ASM), mandatory for physician specialists, launches in 2027. These two models together represent the most aggressive mandatory push since the Obama-era Comprehensive Joint Replacement model, which itself generated meaningful savings and became a template for subsequent episode-based designs.

LEAD (Long-term Enhanced ACO Design), launching in 2027 after ACO REACH concludes, is designed specifically to expand participation among rural and independent physicians — the population historically most absent from value-based arrangements — through improved benchmarking, prospective payments, and specialist integration.

The ACCESS model, launching July 2026, tests outcome-aligned payment for technology-enabled chronic disease management across cardiometabolic and kidney conditions — a 10-year voluntary model with rolling applications through 2033. For practices already in ACO arrangements, ACCESS is designed to complement existing models rather than compete with them, though its integration with MSSP benchmarks becomes operative in 2028.

MAHA ELEVATE tests up to 30 evidence-based proposals for whole-person and lifestyle medicine interventions not currently covered by traditional Medicare — an explicit nod to RFK Jr.’s public health priorities, and a signal that CMMI’s model portfolio will increasingly reflect the administration’s broader “Make America Healthy Again” agenda.

The Credibility Problem CMMI Must Solve

The strongest barrier to provider participation in value-based models is not financial risk tolerance. It is institutional trust.

An ACO could generate savings as measured against a benchmark but generate no savings — or even net costs — relative to what would have happened without the model at all, because benchmarks are not designed to approximate counterfactual spending. Providers who have operated inside CMMI models know this. They have watched benchmarks shift mid-model, seen models expire before participants could recoup their infrastructure investments, and observed the center’s certification bar function more as a ceiling on success than a standard of accountability.

Avalere Health analysis found that community-based CMMI models generated $2.1 billion in net savings, while health plan models resulted in $4.8 billion in net losses — with the MA VBID model alone accounting for the largest share of losses before being terminated five years early. The record is genuinely mixed, which makes Sutton’s defense of the high certification standard simultaneously understandable and insufficient. A high bar is appropriate when the models are well-designed. It is a rationalization when models expire before the evidence matures.

For providers evaluating participation in 2026 and 2027, the relevant question is not whether CMMI’s models are theoretically sound — most of them are. The question is whether the institution will honor its commitments long enough to make the investment worthwhile. The administration’s enthusiasm is not in doubt. Its attention span, across a regulatory and political environment that changes rapidly, is.

What Providers Should Do Now

Three concrete steps for organizations navigating the current CMMI landscape:

Assess mandatory exposure before it arrives. TEAM is already running. ASM launches in 2027. If your organization operates in one of the selected markets for either model, the time to build operational capacity — care coordination infrastructure, episode cost tracking, specialist alignment — is now, not after the model begins. Organizations that entered CJR without preparation underperformed. Those that prepared generated meaningful shared savings.

Evaluate LEAD and ACCESS applications seriously. LEAD’s application window opens in March 2026 for a January 2027 start. For physician-led organizations and rural practices that have found prior ACO models inaccessible, LEAD’s design accommodations — flexible benchmarking, prospective payments, specialist integration pathways — address the most commonly cited barriers. ACCESS applications for the July 2026 start closed March 20, 2026, but rolling applications continue through 2033.

Build internal data infrastructure regardless of model participation. The providers who perform best across all CMMI models — mandatory and voluntary — are those who can track cost and utilization at the patient and episode level in near real-time. That capability is not model-specific. It is the foundational asset that makes any value-based arrangement manageable. Investing in it now hedges against whatever the regulatory environment produces next.

The direction of travel in Medicare payment is not in question. Fee-for-service will continue to shrink as a share of total payment; mandatory risk arrangements will expand; and the providers who have built the operational infrastructure to succeed under value-based models will have a durable competitive advantage over those who have not. CMMI’s execution record is imperfect. Its trajectory is not.

If you enjoyed this piece, subscribe to our monthly newsletter to receive even more in-depth insights.

Sources: Healthcare Dive, “CMS innovation center remains focused on mandatory models, officials say” (March 4, 2026); Congressional Budget Office, “Federal Budgetary Effects of the Activities of the Center for Medicare & Medicaid Innovation” (September 2023); Avalere Health / Healthcare Leadership Council, “Analysis of CMMI Model Costs, Quality Performance, and Transparency” (April 2025); AAFP, “CMMI releases nine new payment models” (January 2026); HFMA, “CMMI payment models for 2026 explained” (January 2026); Holland & Knight, “CMMI Launches Voluntary Payment Model for Qualifying Chronic Conditions with Tech-Enabled Care” (December 2025); Paragon Institute, “How to Reform the CMS Innovation Center with a Choice and Competition Approach” (October 2025); CareJourney, “CMMI Flexing New Primary Care Model” (2024).