A main source of benchmark uncertainty in the Accountable Care Organization Realizing Equity, Access, and Community Health (ACO REACH) model is the retrospective trend adjustment (RTA), a retrospectively calculated factor that corrects for any significant (greater than 1%) differential between the prospective adjusted United States Per Capita Cost (USPCC)¹ trend and the actual REACH National Reference Population² trend.

For ACOs going into performance year (PY) 2025, the trend gap created by the low 2023-2024 adjusted USPCC trend in the PY2025 Kidney Care Choice (KCC)/REACH Rate Book combined with the high emerging 2023-2024 trend in the Reference Population is creating a significant RTA tailwind on the PY2025 benchmark. Based on our estimate of final PY2024 Reference Population claims levels, the PY2025 RTA tailwind is estimated to be about 5.2% for the aged and disabled population,³ significantly increasing the likelihood that the final PY2025 RTA will be a favorable benchmark adjustment and potentially in excess of the RTA risk corridor (which would act to dilute the RTA tailwind for ACOs).

What is the RTA?

If the emerging expenditure trend in the Reference Population deviates by more than 1% from the prospective adjusted USPCC trend in the same timeframe, then the Centers for Medicare and Medicaid Services (CMS) will implement a retrospective trend adjustment (RTA) to the benchmarks of REACH ACOs.⁴ In recent years (since PY2023), the RTA has been a two-year trend correction and will again be so in 2025, i.e., the 2025 RTA will reflect the trend differential from 2023 to 2025.

Given the retrospective nature of the RTA, it is a major source of uncertainty for REACH ACOs. It has materially decreased the prospectively calculated financial benchmarks in every performance year (PY) from PY2021 to PY2023. In PY2024, CMS introduced a symmetric risk corridor to help alleviate the risk associated with the RTA. With the risk corridors, REACH ACOs will only have their benchmarks adjusted for portions of the RTA in excess of 4% (i.e., 100% of the RTA up to +/-4%, 50% of the RTA between +/-4% and +/-8%, and 0% of the RTA in excess of 8%).⁵

CMS interim RTA estimates

CMS releases the National Reference Population Data Report monthly, which includes historical and year-to-date (YTD) eligibility and claims data of the Reference Population, as well as interim forecasts of the RTA. As the performance year develops, the interim RTA gradually converges to the final value, with runout through March of the following calendar year.

Using the December 2024 report as an example, this report includes claims incurred and paid through November and estimates an interim RTA as the ratio between the adjusted USPCC 2022-2024 trend in the PY2024 Rate Book and the estimated 2022-2024 Reference Population trend. To estimate the 2022-2024 Reference Population trend, CMS in turn estimates ultimate 2024 Reference Population expenditures based on the following methodology:

• Use claims incurred through October, the prior month, and paid through November.⁶
• Determine completion factors for each incurred month based on prior years’ data, using the year with the same number of cumulative processing days. For PY2024, January and February use completion factors from 2018, and March to October use completion factors from 2019.
• Apply the same YTD-to-year-end (YE) seasonality as in 2022.

The most recent RTA report (i.e., data through November 2024) estimates PY2024 RTA to be 1.010 for aged and disabled (AD) beneficiaries. The implied 2023-2024 Reference Population trend is about 8.3%. The exclusion of significant, anomalous, and highly suspect (SAHS) catheter costs lowered 2023 expenditure by about 1.0%. The 2023-2024 trend would have been 7.3% if catheter costs were included in 2023. Note that this trend is estimated by the CMS default methodology described above. To estimate a full year 2023-2024 trend requires an independent estimate of claims seasonality and completion of the emerging 2024 experience.

Milliman RTA estimates

Milliman has developed an independent model to estimate potential PY2024 RTA outcomes, which we use to help sensitivity-test financial performance for clients. Our approach incorporates Milliman’s experience with reserving and forecasting. The main methodology differences between our approach and the CMS default methodology include that we:

• Use claims incurred through November and paid through November, utilizing all data available.
• Develop independent estimates of completion factors for each incurred month. Our methodology considers both historical completion patterns of the Reference Population and adjusts them for the timing of CMS payment processing days in 2024.⁷ We consider a range of scenarios; our final estimate uses completion patterns from 2022 and 2023.
• Develop YTD-to-YE seasonality factors based on historical seasonality of the Reference Population, adjusted to the monthly working days for each month in 2024. We consider a range of scenarios; our final estimate uses seasonality patterns from 2023.

Our estimated Reference Population PY2024 ultimate incurred per beneficiary per month (PBPM) rate is higher than CMS’s estimate (based on the default methodology), primarily due to higher completion factor estimates. This results in a higher 2023-2024 Reference Population trend (9.0%) and a higher estimate for the PY2024 RTA (1.017), shown in Figure 1 below.

We make monthly updates to the full-year RTA forecast using this methodology and multiple scenarios can be considered based on different historical precedents for CMS processing speed and the monthly seasonality pattern. For example, older precedents for completion patterns and year-end seasonality patterns would suggest a lower completion and annualization of 2024 YTD experience, causing around a -1.3% decrease in the final RTA through the remainder of the year.

PY2025 RTA tailwind

The adjusted USPCC trends are restated by CMS every year. This restatement, combined with the rolling two-year nature of the RTA, can create interesting dynamics. Figure 1 shows the RTA calculation for PY2023 through PY2025.

For PY2023, the Reference Population (Ref Pop) trend is about 4% lower than the adjusted USPCC in 2021-2022 and about 0.7% higher in 2022-2023. The two effects partially offset, resulting in an overall RTA of 0.9668. About -1% of this RTA adjustment arose because CMS carved out certain catheter expenses from 2023 costs. In other words, without the carve-out, the RTA would have been 0.9762.

For PY2024, the Reference Population trend is lower than the adjusted USPCC in 2022-2023 due to cost restatement, but showing higher trends than the adjusted USPCC in 2023-2024. The two effects partially cancel out, resulting in a forecasted RTA of 1.017, outside of the 1% corridor.

For PY2025, CMS released the PY2025 KCC/REACH Rate Book in November 2024, and restates the 2023-2024 adjusted USPCC trend even lower, thereby increasing the estimated RTA tailwind. Using the current estimates (i.e., assuming the Reference Population trend is approximately 9.0% in 2023-2024, including the catheter cost exclusion, and is 4.2% in 2024-2025, same as the adjusted USPCC trend), the PY2025 RTA tailwind is forecasted to be 1.052.⁸ In this case, the applied RTA is 1.046 due to the application of the 4%-8% risk corridor, resulting in an RTA insufficiency of 0.6%, where the effective benchmark trend is lower than the actual (national) expenditure trend.⁹

A similar table for the end-stage renal disease (ESRD) RTA can be found in the Appendix.

A similar occurrence (i.e., prospective trend estimate being lower than actual emerging trend) is happening in the Medicare Shared Savings Program (MSSP) as well. CMS recently published the MSSP PY2024 Accountable Care Prospective Trend (ACPT) for 2023-2024, which is significantly lower than the emerging 2023-2024 actual trend. For new agreement periods beginning on or after January 1, 2024, the ACPT accounts for one-third of the benchmark trend from the third benchmark year (BY3) to the performance year (PY). Therefore, an understated ACPT will likely hurt MSSPs in 2024 if left unchanged.

Conclusion

Based on the latest available data, the ACO REACH 2023-2024 Reference Population emerging experience is trending higher than the adjusted USPCC trends provided in the PY2025 KCC/REACH Rate Book, creating a 5.2% RTA tailwind on benchmarks for REACH ACOs going into PY2025. However, this large RTA tailwind as it currently stands (with no impact on 2024-2025 trends) would yield a triggering of the RTA corridor, which would result in ACOs receiving a lower actual RTA than the RTA calculation of prior performance years would have produced (that is, sharing some of the RTA tailwind with CMS). We will be monitoring emerging 2025 experience monthly to evaluate whether the tailwind grows or shrinks throughout the year. If you are interested in understanding the monthly movement of our PY2025 RTA estimate, please reach out. We encourage REACH ACOs to understand the implications and incorporate the RTA into their financial projections, while keeping in mind that CMS still retains authority to make program rule changes for PY2025.

---

Additional Milliman publications related to RTA:

• The Retrospective Trend Adjustment in Direct Contracting and ACO REACH: A brief explanation
• The retrospective trend adjustment in ACO REACH: What’s new in PY 2024?
• The retrospective trend adjustment: Looking back, looking forward

¹ The adjustment includes adding hospice cost and removing the uncompensated care cost (UCC) from the USPCC.

² The definition of the National Reference Population can be found in Appendix 2 of the Milliman white paper found at https://www.milliman.com/en/insight/aco-reach-risk-scores-performance-year-development. It is also discussed in the article found at https://www.milliman.com/en/insight/retrospective-trend-adjustment-direct-contracting-aco-reach. We will use the term “Reference Population” for short hereafter.

³ Based on CMS guidance, claims for the two urinary catheter codes related to SAHS billing activity are removed from 2023 Reference Population expenditures for the calculation of the PY2023 RTA right before PY2023 final settlement. We assume the same carve-out will be made to 2023 expenditures (but not 2025 expenditures) when calculating the PY2025 RTA. The RTA tailwind would be around 4.2% if catheter claims are not carved out of 2023 expenditures.

⁴ CMS. ACO Realizing Equity, Access, and Community Health (REACH) Model: PY2025 Financial Settlement Overview, page 4. Retrieved December 27, 2024, from https://www.cms.gov/priorities/innovation/files/aco-reach-model-py25-financial-settlement-overview.pdf.

⁵ For more information on the RTA corridors, see: https://www.milliman.com/en/insight/rta-aco-reach-whats-new-py-2024.

⁶ This is the default setting in CMS’s RTA report, although users can manually change the incurred-through and paid-through dates.

⁷ CMS adds claims to the National Claims History (NCH) weekly on Fridays.

⁸ Calculations shown in this paragraph are to illustrate the general magnitude of the tailwind. The actual 2023-2024 and 2024-2025 Reference Population trends will likely be different from what’s shown in Figure 1, resulting in different PY2025 RTA estimates.

⁹ See RTA insufficiency discussed in the white paper at https://www.milliman.com/en/insight/rta-aco-reach-whats-new-py-2024.

Appendix