Three Years of ACO REACH:
What the Data Actually Shows

ACO REACH is winding down. The program launched as GPDC in PY2021, transitioned to ACO REACH in PY2023, and runs through PY2026 — but with the LEAD application window now closed, the successor program is set. Public data is available through PY2023, giving us three full years to examine. What did those years actually show about what drives performance in a total cost of care model with capitation?

The honest answer is that the data is more clarifying than most program-level summaries suggest. Configuration mattered enormously. Experience compounded in ways the headline numbers obscure. And a small subset of ACOs captured a disproportionate share of the savings — not randomly, but in ways that are largely predictable from the structural choices each organization made before the performance year began.

We built the ACO REACH → LEAD Explorer to make this data accessible. What follows is what we found when we looked at all 150 ACOs across all three years.

Finding 1: The Program Got Dramatically Better in Three Years

The trajectory from PY2021 to PY2023 is striking. In the first performance year, the program was still finding its footing — many ACOs were transitioning from GPDC, capitation infrastructure was new, and the public health emergency was creating real disruption in utilization patterns. Total earned savings across the program were $508 million in PY2022. By PY2023, that number had nearly doubled to $923 million.

The share of ACOs earning savings also climbed steadily: from 57% in PY2021 to 72.7% in PY2023 — the highest rate in the program's history and significantly above what MSSP achieves in a typical year. This isn't just a function of ACOs getting better at the program. It reflects a selection effect: organizations that couldn't make the model work left, and the remaining cohort increasingly represents organizations with the infrastructure, patient panel, and clinical workflows to succeed in total cost of care.

Finding 2: Configuration Was the Single Biggest Performance Driver

If you want to understand why two ACOs with similar patient populations and similar clinical programs can generate dramatically different savings rates, start with configuration. The data is unambiguous on this point.

In PY2023, ACOs operating under the Global risk arrangement with Total Care Capitation and no stop loss averaged a 6.4% savings rate, with 93% of those ACOs earning positive savings. ACOs on Professional risk averaged less than half that. The gap isn't marginal — it's structural, and it reflects something important about how capitation shapes organizational behavior.

Total Care Capitation — where the ACO receives a risk-adjusted monthly payment covering all services for aligned beneficiaries — creates a fundamentally different set of incentives than Primary Care Capitation. When your financial exposure extends to all spend, not just primary care, the organizational focus shifts accordingly. Specialist relationships, care coordination protocols, and post-acute utilization management become revenue drivers rather than cost centers. ACOs that built those capabilities and then elected TCC captured the benefit.

Finding 3: CI/SEP Status Is the Strongest Predictor of Performance

The Continuous Improvement / Sustained Exceptional Performance designation was one of the most consequential program features, and the data shows why. In PY2023, CI/SEP ACOs averaged a 5.80% savings rate. Non-CI/SEP ACOs averaged 1.07%. That's a 5.4× difference — not a modest advantage, but a categorically different level of performance.

CI/SEP qualification requires demonstrating either sustained exceptional savings or continuous improvement over multiple years. It's not a lottery — it reflects durable organizational capability. CI/SEP ACOs also have access to the High Performers Pool, which in PY2023 directed 44% of all shared savings to just 24 organizations (18% of participants). The concentration of earnings at the top of the distribution is a feature of program design, not an anomaly.

Finding 4: Experience Compounds — and the Numbers Prove It

Forty-four ACOs participated in all three performance years of ACO REACH — PY2021, PY2022, and PY2023. Their savings rate trajectory tells a clear story about what happens when organizations stay in the model and learn it.

This cohort averaged a 3.95% savings rate in PY2021. By PY2022, that had risen to 6.95%. By PY2023, they averaged 7.41% — an 87% improvement over three years. Seventy percent of all-3-year ACOs improved their savings rate year over year. This is not what random variation looks like. It's what organizational learning looks like.

The mechanism is straightforward: capitation payments create predictable revenue, which funds infrastructure investment, which improves care coordination and cost management, which generates savings, which funds more investment. ACOs that stayed long enough to complete this cycle demonstrated compounding returns that first-year participants simply couldn't match.

Finding 5: Quality and Savings Are Related — But Not the Way You'd Expect

PY2021 and PY2022 operated under a pay-for-reporting quality framework — ACOs earned full quality credit for submitting data, which meant quality scores were uniformly high and carried limited signal. PY2023 shifted to pay-for-performance, and the picture became more interesting. The average quality score dropped to 81.0 out of 100, with meaningful variance emerging for the first time.

What the data shows is that CI/SEP ACOs — which by definition have demonstrated sustained or improving performance — also tended to score higher on quality measures. The CI/SEP designation functions as a proxy for organizational maturity: organizations that have built the infrastructure to generate compounding savings tend to have also built the clinical workflows that drive quality performance. The two aren't perfectly correlated, but they share a common root cause: operational discipline applied consistently over multiple years.

This matters for LEAD, which makes quality an explicit gating mechanism for sharing rate determination. Organizations entering LEAD with strong CI/SEP track records from REACH are better positioned not just financially but on the quality dimension that will increasingly determine their sharing rate.

What This Means for LEAD

The LEAD Model — launching January 1, 2027, for organizations that applied before May 17 — is explicitly designed around what ACO REACH taught. The 10-year agreement period with no benchmark rebasing is a direct response to Finding 4: organizations that demonstrate compounding performance deserve a structure that lets them realize the return on that investment without being penalized for succeeding. The expanded concurrent risk adjustment is a direct response to what REACH showed about high-needs populations. The CARA episode arrangements for Global track ACOs extend the TCC logic to specialist relationships in ways PCC never could.

For ACO REACH participants, the transition decision is not primarily about whether LEAD is a good program. The evidence strongly suggests it is. The decision is about configuration, partnership structure, and whether the organizational infrastructure built over three years of REACH is positioned to generate the compounding returns that the longitudinal data predicts are available to experienced operators in a stable-benchmark environment.

For MSSP ACOs considering LEAD, the REACH data is useful precisely because it shows what total cost of care performance looks like when capitation is the payment mechanism and the agreement term is long enough to let organizational learning compound. The question is whether your savings rate trajectory, quality infrastructure, and post-acute relationships put you on the right side of the performance distribution that the REACH data reveals.