On August 14, 2023, the Center for Medicare and Medicaid Innovation (CMMI) released guidance on several methodology changes to the Accountable Care Organization Realizing Equity, Access, and Community Health (ACO REACH) Model effective for performance year 2024 (PY2024).¹ The stated goals of these changes include increasing the predictability of the model, limiting inappropriate risk score growth and maintaining consistency across Centers for Medicare and Medicaid Services (CMS) programs and CMMI models, refining policies to advance health equity, and maintaining consistency across ACO models.

We do not expect the changes CMMI has introduced to significantly improve the predictability of ACO financial benchmarks under most scenarios. While these changes dampen the potential effects of the Coding Intensity Factor (CIF) and Retrospective Trend Adjustment (RTA), ACO financial benchmarks may still receive material retrospective adjustments, which are difficult to predict. In addition, the introduction of the V28 CMS-hierarchical condition category (HCC) risk score to the ACO REACH model is similar to the recent changes to Medicare Advantage (MA) and the Medicare Shared Savings Program (MSSP) and helps maintain risk score consistency across programs. The changes introduced to the health equity component of the benchmark calculation assist in refining the existing calculation and measurement to more appropriately capture populations in need of additional resources. However, the scale of the health equity benchmark adjustment has not changed meaningfully and is expected to have a relatively small impact on the financial benchmark for most ACOs.

We summarize notable changes related to risk scores, increased benchmark predictability, health equity, High Needs Population ACOs, and beneficiary alignment in the sections below. We do not cover all of the rule changes outlined in the document but instead focus on those that we expect to have the most material impact on REACH ACOs currently participating in the model. REACH ACOs should review the rule changes closely to assess the implications for their specific circumstances.

Risk score changes

Introduction of 2024 CMS-HCC Risk Model (V28)

Rule Change

Risk scores for PY2024 and all future PYs will be calculated using the same blend of V24 and V28 that is being used in the MA model (33% V28 in 2024, 66% V28 in 2025, 100% V28 in 2026). The risk scores for all of the benchmark years (including the risk score reference year and the rate book) will be calculated using the same blend of risk score models that is used for the PY.

Potential Implications

Historical shifts (during the benchmark years) in normalized risk scores using older risk score models (V22 and V24) may be significantly different from shifts observed using the new blending of the V28 risk model. However, we anticipate less disruption under this approach than for ACOs subject to differing risk score models across benchmark and performance years (such as those proposed to be used in the MSSP model in certain situations). CMMI estimates that the impact of this risk score blending on 2024 benchmarks across all REACH ACOs will be a reduction of about 0.4%. However, some ACOs will be affected more severely than others.

1% Coding Intensity Factor (CIF) Cap for PY2024

Rule Change

The CIF accounts for the aggregate shift in normalized risk score among all ACO REACH assigned beneficiaries (versus the normalization factor that accounts for the aggregate shift in the reference population). CMS will now cap the impact of the CIF at 1% (i.e., no higher than a 1% reduction in normalized risk scores or benchmarks) for PY2024.

Potential Implications

As REACH ACOs continue to mature and improve coding accuracy among their assigned populations, the possibility of a material CIF (representing a material reduction in all REACH ACO benchmarks) has increased. By introducing this 1% CIF cap for PY2024, CMS is limiting the negative impact the CIF can have on participating ACOs in 2024. The CIF is one key source of benchmark uncertainty, so this change will make ACO REACH benchmarks more predictable.

If the 1% cap is reached or exceeded, the CIF would not be a revenue-neutral coding adjustment.

Increased benchmark predictability changes

Retrospective Trend Adjustment (RTA) Dampening in Excess of +/-4%

Rule Change

The RTA currently accounts for any observed differences (in excess of +/-1%) between the prospective nationwide trend factors included in the rate book and the actual nationwide observed trend for the REACH reference population. Starting in PY2024, the benchmark for REACH ACOs will only be adjusted for a portion of the RTA in excess of 4% (100% of the RTA up to +/-4%, 50% of the RTA between +/-4% and +/-8%, and 0% of the RTA in excess of 8%.

Potential Implications

While this adjustment helps stabilize REACH ACO benchmarks in the case of sizable RTAs in future years, our expectation is that it is unlikely that the RTA will exceed 4% in future performance years as the market continues to recover and stabilize after the lingering impacts of COVID-19 on healthcare utilization and trends.

Health equity changes

Modification to Calculation of Health Equity Benchmark Adjustment (HEBA) Score

Rule Change

CMS is revising the calculation of the HEBA Score for PY2024 to be calculated as:

• 1/3 Weight on National-Based Area Deprivation Index (ADI)
• 1/3 Weight on State-Based ADI
• 1/3 Weight on Dual Medicare-Medicaid Status/Low-Income Subsidy Status

Potential Implications

This modified HEBA score calculation includes a greater weight on dual status (as well as incorporating low-income subsidy status) and takes into account state-based ADI. This change attempts to address a number of concerns raised by ACOs that the nationwide ADI does not sufficiently recognize health equity needs for some populations.

The specific impact of these calculation changes on ACOs will vary significantly and will need to be assessed on a case-by-case basis.

Adjustment to Health Equity Benchmark Adjustments (HEBA)

Rule Change

The HEBA adjustment to the benchmark will be modified to be:

• +$30 per beneficiary per month (PBPM) for beneficiaries with a HEBA score in the top decile
• +$20 PBPM for beneficiaries in the second decile
• +$10 PBPM for beneficiaries in the third decile
• -$10 PBPM for beneficiaries in the bottom three deciles

Potential Implications

This adjustment expands access to positive HEBAs to include increased bonus payments for those in the second and third deciles (who previously did not receive any adjustment). Additionally, the downward HEBA was reduced to only impact beneficiaries in the bottom three deciles instead of the bottom five deciles. In general, this change will increase the net impact of the HEBA on REACH ACOs (and remove the net-neutral impact of this adjustment across all REACH ACOs) but its effect on a specific ACO will depend on the changing HEBA scores and will need to be assessed for each ACO individually.

This change should increase the benchmark for most ACOs, all else equal. ACOs with beneficiaries in the 30th to 50th percentile and the 70th to 90th percentile of health equity scores will see the largest impact.

High needs population ACO changes

Updated Eligibility Criteria for High Needs Population ACO

Rule Change

In addition to the current criteria for alignment to a High Needs Population ACO, CMS is now including beneficiaries with at least 90 Medicare-covered days of home health services utilization or at least 45 Medicare-covered days in a skilled nursing facility (SNF) within the previous 12 months.

Potential Implications

This will likely increase the number of aligned beneficiaries for most High Needs Population ACOs. CMS expects that the expanded eligibility criteria will more effectively identify beneficiaries with complex healthcare needs.

Introduction of Risk Score Caps to High Needs Population ACOs

Rule Change

Currently, the 3% symmetric risk score cap is not applied to High Needs Population ACOs. Starting in PY2024, CMS will apply this cap.

Potential Implications

In past documentation,² CMS has stated that the cap would not be applied to High Needs Population ACOs, but that, if the observed level of coding growth for High Needs Population ACOs was greater than the capped coding growth observed for the Standard and New Entrant ACOs, then they may begin applying the cap in PY2024 and onward. This change implies that the observed coding growth for High Needs Population ACOs was higher than expected and, if that is the case, then some ACOs will likely reach the 3% risk score cap. The CIF is calculated after the risk score cap is applied, so this change could reduce the impact of the CIF for High Needs Population ACOs.

Beneficiary alignment changes

Updated Minimum Beneficiary Alignment Rules

Rule Change

The minimum beneficiary alignment for New Entrant ACOs will be reduced for PY2025 from 5,000 to 4,000 and the minimum for High Needs Population ACOs will be reduced from 1,200 to 1,000 for PY2025 and from 1,400 to 1,250 for PY2026.

Additionally, CMS will introduce a 10% alignment buffer such that an ACO’s beneficiary count can temporarily drop below the minimum (not more than 10% below the minimum) and continue its participation in the model (this is only allowed once during an ACO’s participation in the model).

Potential Implications

These changes make it somewhat easier for smaller organizations to meet the criteria for participation in ACO REACH. This aligns with the CMS goal of increasing the number of beneficiaries in accountable care relationships.

This white paper presents the authors’ preliminary thoughts on the ACO REACH PY2024 model updates and not those of Milliman.

---

¹ CMS. ACO REACH Model Performance Year 2024 (PY2024) Model Update – Quick Reference. CMS Innovation Center. Retrieved September 6, 2023, from https://innovation.cms.gov/innovation-models/reach-py24-model-perf.

² ACO Realizing Equity, Access, and Community Health (REACH) and Kidney Care Choices Models: PY2023 Risk Adjustment (cms.gov). Retrieved September 7, 2023 from https://www.cms.gov/priorities/innovation/media/document/aco-reach-py2023-risk-adjustment.