The CMMI Track Record, Honestly

Now apply the whole course to the biggest question in the field: after more than a decade and dozens of models, has the Center for Medicare and Medicaid Innovation actually worked? The honest answer is sobering, and getting it right is exactly what separates rigorous analysis from advocacy.

What the numbers show

The Congressional Budget Office, which scores CMMI’s budget effects, reached a striking conclusion. Over CMMI’s first decade:

• CMMI increased direct federal spending by about $5.4 billion from 2011 to 2020.
• It spent roughly $7.9 billion operating models, which reduced spending on benefits by about $2.6 billion. The operating costs and payouts exceeded the medical savings.
• Early projections were far rosier: CBO had once estimated $2.8 billion in savings over that decade, and once projected $77.5 billion in savings for 2021 to 2030, a figure it later revised to a small spending increase.

And on expansion: of roughly 70 models tested since CMMI’s creation, only four have been certified for nationwide expansion.

Worth remembering: this is the gross-versus-net lesson at national scale. Individual models often produced gross medical savings. Net of what CMMI spent to run them and paid providers, the portfolio increased federal spending. Both statements are true, and only the net one answers “did it pay off for the taxpayer.”

How to hold this honestly

A nonpartisan reading resists two temptations. The first is to use these numbers to declare value-based care a failure. They do not show that. They show that a portfolio of experiments, many of them voluntary, underpowered, and confounded in exactly the ways this course described, has not yet produced net savings at scale, and that a few models clearly worked.

The second temptation is to explain the numbers away. The honest analyst states them plainly: after a decade, the aggregate financial case is unproven, and the burden is on proponents to show which specific designs work, not to assert that the idea works in general.

Why the record looks this way

Everything in this course predicts it. Voluntary participation invited favorable selection. High-cost targeting invited regression to the mean. Constructed benchmarks made savings assumption-dependent. Gross figures got the headlines while net figures told the fiscal story. Early-adopter pilots did not generalize. The disappointing aggregate is not a mystery; it is what happens when hard evaluation meets hopeful expectations.

Key takeaways

• CBO found CMMI increased net federal spending by about $5.4 billion in its first decade, against early projections of savings.
• Only four of roughly 70 models have been certified for national expansion.
• The honest reading is neither dismissal nor denial: the aggregate financial case is unproven, and specific designs, not the idea in general, are what the evidence can support.

Sources

• Congressional Budget Office, Federal Budgetary Effects of the Activities of the Center for Medicare and Medicaid Innovation (September 2023) (opens in a new tab)