The CMS Innovation Center: A Mixed Bag of Savings, Quality, and Lessons 

The Center for Medicare and Medicaid Innovation (CMMI), established under the Affordable Care Act in 2010, was designed with a bold mission: to test innovative payment and care delivery models that could reduce costs while maintaining or improving the quality of care in Medicare, Medicaid, and the Children’s Health Insurance Program (CHIP). Focused primarily on Medicare, the federal health insurance program for older Americans, CMMI was envisioned as a laboratory for value-based care, aiming to shift the healthcare system away from fee-for-service models toward approaches that prioritize outcomes over volume. Yet, more than a decade later, a recent analysis by Avalere Health reveals a complex picture. While some CMMI models have delivered cost savings and quality improvements, many have fallen short, raising questions about the agency’s effectiveness and its path forward under increasing scrutiny. 

A Sobering Financial Reality 

Avalere Health’s analysis of 18 CMMI payment models, each active for at least two years and impacting over 25,000 beneficiaries, paints a sobering picture. Collectively, these models cost the federal government $7.7 billion, including $6.4 billion in net model expenses and $1.3 billion in implementation and evaluation costs. Only one-third of the models produced substantial savings, while another third generated significant losses, and the remaining third had a nominal financial impact. This uneven performance aligns with a 2023 Congressional Budget Office (CBO) report, which found that CMMI’s activities increased federal spending by $5.4 billion between 2011 and 2020—roughly 0.1% of net Medicare spending during that period. 

These numbers challenge the optimistic projections made when CMMI was created. In 2010, the CBO estimated the agency would save $1.3 billion over its first decade, and by 2016, it projected savings of $34 billion from 2017 to 2026. However, Avalere’s earlier analyses, including a 2022 estimate of $9.4 billion in net losses over the same period, suggested those forecasts were overly ambitious. The reality is that not every model was designed to save money outright—some were intended to test concepts or improve care delivery—but the aggregate financial burden has drawn criticism, particularly under the Trump administration, which has targeted CMMI for spending reductions. 

Bright Spots Amid the Losses 

Despite the overall financial shortfall, Avalere’s report highlights that not all CMMI models have underperformed. One-third of the 18 models analyzed—six in total—delivered substantial savings. For example, the Maryland All-Payer Model saved approximately $975 million, and the Maryland Total Cost of Care Model contributed $689 million in savings. These models, which align payments across all payers to incentivize cost-effective care, demonstrate that CMMI’s experiments can yield meaningful results when designed thoughtfully. Other successes, like the Home Health Value-Based Purchasing Model and the Pioneer Accountable Care Organization (ACO) Model, have shown promise in reducing costs without compromising quality, with some elements even integrated into broader Medicare programs. 

These bright spots suggest that CMMI’s work, while imperfect, is not without value. The agency’s iterative approach—testing, refining, and sometimes scaling models—has produced lessons that could inform future efforts. For instance, models like Bundled Payments for Care Improvement (BPCI) and its successor, BPCI Advanced, have reduced Medicare payments for certain episodes of care by encouraging providers to discharge patients to home rather than costly post-acute facilities, all while maintaining quality. Such findings underscore the potential for targeted interventions to drive value-based care, even if the broader portfolio has yet to deliver consistent savings. 

Quality Outcomes: A Mixed Bag 

Cost savings are only half of CMMI’s mandate; the other is improving care quality. Here, too, the results are mixed. Avalere’s analysis found that four of the 18 models demonstrated clear improvements in quality metrics, such as reducing emergency department visits or enhancing chronic disease management. Three showed minimal progress, four had no statistically significant effect, and seven produced mixed outcomes. Patient experience surveys, conducted for most models, revealed similarly varied results: four models showed small improvements, eight had no significant changes, and two had mixed impacts. 

This inconsistency highlights a key challenge: measuring quality is complex, and not all models are designed to produce immediate, measurable improvements. For example, reducing hospitalizations—a common quality metric—may take years to achieve, especially in models targeting chronic conditions. Moreover, quality metrics often rely on patient-reported outcomes, which can be subjective and difficult to standardize. The Maryland Total Cost of Care Model, praised for its savings, faced criticism in a 2022 analysis for not significantly improving health equity or detecting disparities in underserved populations, illustrating the difficulty of balancing cost and quality goals. 

Transparency and Stakeholder Input: Room for Improvement 

Beyond finances and quality, Avalere’s report raises concerns about CMMI’s transparency and engagement with stakeholders. Only half of the 18 models solicited public input through rulemaking or requests for information during their design phase. None were endorsed by the Physician-Focused Payment Model Technical Advisory Committee (PTAC), a group established to propose Medicare payment reforms. This lack of external validation has fueled perceptions that CMMI operates in a bubble, potentially missing opportunities to refine models based on real-world insights. 

The Healthcare Leadership Council, which commissioned Avalere’s analysis, emphasized the need for greater collaboration with the private sector. Its president, Maria Ghazal, noted that while some models have advanced value-based care, others highlight the importance of incorporating feedback from providers and patients to ensure success. As CMMI navigates a new era under the Trump administration, which has already canceled four models and halted two planned demonstrations, opening channels for stakeholder input could help align future experiments with practical realities. 

The Political Spotlight 

CMMI’s financial track record has placed it under intense scrutiny, particularly from Republican lawmakers. In a 2024 House Energy and Commerce Committee hearing, members criticized the agency for failing to deliver on its projected $80 billion in savings over two decades, pointing to the CBO’s $5.4 billion loss estimate as evidence of inefficiency. The Trump administration’s recent decision to terminate models like Primary Care First, Making Care Primary, End-Stage Renal Disease Treatment Choices, and Maryland Total Cost of Care by the end of 2025 reflects this cost-cutting mindset. CMS estimates these cancellations will save taxpayers $750 million, though the long-term impact on care delivery remains unclear. 

Yet, defenders of CMMI argue that innovation inherently involves risk. Not every model is expected to succeed, and failures can yield valuable lessons. Elizabeth Fowler, CMMI’s deputy administrator, highlighted in 2024 that some models have reduced Medicare fee-for-service costs by 2.1% and hospitalizations by 16%, particularly for underserved populations. Democratic lawmakers, like Rep. Frank Pallone, have also praised CMMI for benefiting 41 million beneficiaries and strengthening primary care access. These perspectives underscore the tension between short-term fiscal accountability and the long-term goal of transforming healthcare delivery. 

Lessons for the Future 

Avalere’s report doesn’t just critique CMMI—it offers a roadmap for improvement. The one-third of models that saved money, like the Maryland models, suggest that aligning incentives across payers and focusing on high-cost care episodes can work. Meanwhile, the mixed-quality outcomes highlight the need for better metrics, perhaps incorporating more patient-centered measures like functional status or quality of life. The report also calls for greater transparency, urging CMMI to engage stakeholders early and often to refine model designs. 

CMMI’s iterative process—testing models, learning from failures, and scaling successes—remains a cornerstone of its approach. Only four of its 50 models have been expanded nationwide, including the Pioneer ACO Model and the Medicare Diabetes Prevention Program. However, these successes prove that scaling is possible when savings and quality align. As CMMI moves forward, it could prioritize models with proven elements, like bundled payments or accountable care organizations, while refining evaluation methods to better capture long-term impacts. 

A Balancing Act 

The CMMI story is one of ambition tempered by reality. Its mandate to lower costs and improve care is undeniably complex, requiring experimentation in a healthcare system resistant to change. Avalere’s analysis shows that while CMMI has not yet achieved consistent savings, its successes offer hope. The agency’s ability to learn from wins and losses will determine whether it can fulfill its promise of value-based care. 

As political pressures mount, CMMI must balance innovation with accountability. Canceling underperforming models is a start, but the agency’s future hinges on designing experiments that deliver measurable results without alienating providers or patients. By embracing transparency, refining quality metrics, and building on proven strategies, CMMI can move closer to a healthcare system that spends smarter and cares better. For now, it remains a work in progress—one with high stakes and higher potential.