Daniel Shenfeld is an adjunct Assistant Professor in the Department of Medical Ethics and Health Policy at the University of Pennsylvania. He is also the Founder and Principal of Manganese Health Data Solutions.
In 2010, Congress launched the Center for Medicare and Medicaid Innovation (CMMI) to explore new payment models and help control healthcare costs. However, CMMI’s impact has been limited, with most initiatives failing to achieve significant savings, largely due to lengthy development and evaluation timelines. These delays result primarily from outdated processes and technology. By modernizing its approach, CMMI can streamline innovation, reduce costs, improve patient outcomes, and strengthen the federal government’s capacity to shape and respond to a rapidly evolving healthcare market.
Traditionally, healthcare payments rewarded providers based on service volume (“fee-for-service”), often leading to unnecessary treatments. “Value based care” (VBC) offers an alternative approach, tying provider compensation to the quality of care and patient outcomes. One of the objectives of the Affordable Care Act was to promote the transition to VBC. However, Congress recognized the challenges of shifting from the deeply-entrenched fee-for-service model, and established CMMI to develop, test, and scale innovative VBC models.
In the 14 years since its inception, CMMI has launched over 50 pilot programs, under both Democratic and Republican administrations. However, few of the programs have achieved any meaningful success, and CMMI has failed to deliver significant taxpayer savings. A 2023 CBO report found that only six programs produced statistically significant savings, all of which were modest–and even then, none surpassed $220M in annual savings. By comparison, the Medicare Shared Savings Program, which CMMI does not operate but regards as a foundation for testing new payment models, generates about $2 billion in annual savings, while total Medicare expenditures in 2023 were just over $1 trillion. When factoring in CMMI’s $10 billion-per-decade operating budget and the cost of implementing its programs, the agency’s total budgetary impact has been marginally negative.
Unsurprisingly, the CBO findings triggered heated political backlash against CMMI. What the backlash failed to account for, however, is the reality that developing successful programs requires multiple iterations, and meaningful savings emerge only after a program scales. While CMMI’s effectiveness in its first decade should not be judged solely by total savings, the CBO report does point towards a pressing question: how can CMMI improve and accelerate its success rate? This is an issue that matters not only for improving healthcare quality and reducing costs but also for strengthening state capacity by facilitating the government’s ability to drive innovation.
There are indicators that CMMI recognizes these challenges. For example, in 2022,it launched a “strategy refresh,” outlining strategic objectives for the remainder of the decade and proposing next steps–including a critical pledge to reduce program complexity. Nonetheless, the “strategy refresh” failed to answer these critical questions: How will these strategic objectives translate into a higher success rate? And how exactly will CMMI implement its proposed next steps?
The new strategic guidelines also conspicuously overlook the issue of speed—arguably the biggest obstacle to CMMI’s success. Most of its pilot programs follow lengthy evaluation periods (five, seven, or 10 years), and are considered for expansion only after evaluation is complete… Meanwhile, the healthcare market continues to quickly evolve, so that by the time a program is deemed ready for expansion, it may already be obsolete. This slow-moving approach makes it difficult to achieve meaningful savings within a reasonable timeframe.
A key reason for CMMI’s slow progress is its inefficient program performance analysis. Outdated tools and inefficient processes hinder evaluations, while participating providers–who must track their own performance–often lack the technical skills to do so effectively. Market solutions have proved to be extremely costly, eroding providers’ already thin financial incentives to participate in CMMI’s programs and further reducing success rates.
What’s more, despite the critical role of technology in addressing these challenges, the “strategy refresh” mentions it only in passing. A modern technology and analytics platform could streamline CMMI’s operations, accelerate evaluations, and boost program adoption. It would also empower providers to assess and improve their performance, ultimately enhancing CMMI’s overall impact. Achieving this requires CMMI to embrace a modern approach to software development—an ambitious but attainable goal. Investing in this effort is not only crucial for fulfilling CMMI’s mission but also for strengthening the federal government’s broader digital capabilities.
Success in CMMI’s programs requires an investment in technology and analytics
To understand how CMMI can accelerate its process, it’s essential to view program implementation from the perspective of participating providers–hospitals, clinics, and the organizations that manage them. VBC programs are designed to incentivize improvement to care quality and outcomes while reducing unnecessary or wasteful care. However, for providers to succeed under VBC, they must adopt significant clinical and operational workflow changes. These may include proactive engagement with at-risk patients to prevent hospital admissions or directing patients away from emergency rooms toward urgent care centers.
VBC programs also present a significant financial challenge to providers, as they require a fundamentally different financial model than the traditional fee-for-service system.To achieve savings, providers must be able to forecast the financial impact of workflow changes with a reasonable degree of confidence. However, gaining this level of insight demands substantial investment in data analytics and financial modeling. Many providers, especially smaller ones, rely on costly consultants to generate financial forecasts—often spending six-figure sums just to determine whether participation in a program is financially viable.
This is a significant barrier to entry, which contributes to low uptake of some CMMI programs. For example, the “Primary Care Flex” model only attracted 24 participants out of several hundred eligible organizations. Similarly, the “Making Care Primary” program, which targeted small primary care providers, only had 128 participants, which–presumably–played a role in its cancelation earlier this year.
Similar challenges persist throughout program implementation, as financial projections must be continuously re-assessed– again requiring robust analytics and financial modeling capabilities. Many participants, particularly smaller providers, lack the resources to invest in these capabilities, leading to poorer performance and ultimately lowering the program’s overall success rate.
At its core, the key obstacle for providers is a lack of financial visibility—timely, reliable, and detailed financial performance data at an accessible cost. Addressing this issue would not only increase provider participation in CMMI programs, but also accelerate the implementation process and improve success rates.
A modern analytics platform can accelerate CMMI’s success
To improve financial visibility, providers need several specific analytic capabilities. First, claims data–that is, detailing the healthcare services provided to patients–must be collected and stored in a modern database. This raw data then needs to be processed so that providers can extract actionable insights about their performance, such as identifying which services are being delivered at higher costs or frequencies compared to peers and which patient populations are most affected. Providers must also have the ability to model the financial impact of workflow changes, such as estimating the savings from reducing the use of high-cost services.
The same analytic capabilities are also needed in CMMI’s own work to design and evaluate programs. But currently, this work is fragmented. Programs are typically evaluated separately by different contractors, often using outdated tools such as SAS, a statistical analysis software rarely used in modern production systems, in a process that takes several months.
To address this, both CMMI and participating providers need a modern analytics platform capable of processing data automatically and at scale. This platform would enable participating providers to access detailed and actionable financial and operational reports in near-real time with benchmark performance against regional peers. Over time, the platform could also enable providers to model the financial impact of potential workflow changes. Additionally, this platform would streamline CMMI’s program design and evaluation cycle, as data processing and much of the analysis would be standardized and automated.
Some may question whether the government rather than the private sector. However, in this case, the private sector has proven inefficient. Many providers assess their performance using benchmarks–comparisons of metrics across regional or peer groups—but accessing the necessary historical CMS data costs hundreds of thousands of dollars annually, a price most providers cannot afford. Instead, benchmarks are produced by a small number of vendors, who markup the price to providers, essentially leveraging monopoly pricing power.
As a consultant who helps providers navigate these challenges, I’ve learned through conversations that even small providers pay vendors hundreds of thousands of dollars each year for these services, despite the fact that most see little financial benefit from participating in VBC programs. Meanwhile, larger providers build capabilities internally, leading to duplicative work and increasing administrative costs, which ultimately undermines CMMI’s objectives.
A new CMMI analytics platform would reduce barriers to entry for its programs, lower administrative costs, increase the probability of program success thanks to improved financial visibility, and accelerate program design, launch, and evaluation timelines. This would be a significant undertaking that would require CMMI to adopt a modern approach to software and analytics development. However, with all the necessary data already available in CMMI’s systems, this undertaking is both achievable and crucial.
To build the platform, it is critical that CMMI follow agile development processes, emphasizing incremental progress and fast delivery cycles. While software development would be contracted out, CMMI should remain closely involved to effectively guide the work to meet real user needs. Various software development shops, many of them quite small, have helped private healthcare companies build similar systems and would be well-suited for the task. The development can be accelerated by leveraging production-grade open-source tools for claims processing and analytics. Based on my experience helping participating providers build internal analytics systems, a six-month development timeline would be an ambitious but achievable goal for a competent development team with focused guidance.
By leading the way in building capacity and new technology practices within the federal government to drive innovation, CMMI can realize its potential and bring about much-needed transformative change to the U.S. healthcare system.
Contact: daniel.shenfeld@pennmedicine.upenn.edu