Background
The Medicare fee-for-service (FFS) Global and Professional Direct Contracting (GPDC) program is an accountable care organization (ACO) model run by the Centers for Medicare and Medicaid Services (CMS) Innovation Center (CMMI). CMMI provides DCEs with a prospective financial benchmark prior to the beginning of each performance year. However, CMS guidance directs that, if “Adjusted USPCC trend differs by at least 1% from the observed expenditure trend in the DC National Reference Population…CMS may apply a retrospective trend adjustment to the benchmark that reflects this difference.”1 In performance year (PY) 2021, CMMI applied a significant Retrospective Trend Adjustment (RTA) and, through the first three quarters of PY 2022, CMMI is estimating an RTA will apply. The RTA was applied because the actual PY Adjusted U.S. Per Capital Cost (Adjusted USPCC) for the DC National Reference Population came in lower than CMMI’s preliminary Adjusted USPCC, which was used to set preliminary financial benchmarks.
The application of the RTA for PY 2021 and PY 2022 has garnered much attention from DCEs because it has meaningfully decreased financial benchmarks relative to prospective estimates from CMMI. In this paper, we explain how the RTA is determined, and what may be in store for 2023.
CMMI has not published data on the RTA as of the date of this publication. This paper relies on publicly available information and hypotheses about the RTA.
What effect has the RTA had?
In PY 2021 and PY 2022, the RTA decreased financial benchmarks for DCEs. The Adjusted USPCC used for the preliminary DCE financial benchmarks was substantially higher than actual national expenditures for the DC National Reference Population. Figure 1 illustrates the RTA for PY 2022 using the most recent available data at the time of publication. For example, CMMI set the preliminary Adjusted USPCC for the Aged and Disabled assuming annual trends in expenditures per beneficiary per month (PBPM) would be 5.1% from PY 2019 to PY 2022. Actual emerging trends are much lower. The RTA adjusts benchmarks for the resulting difference between projected and actual expenditures.
Figure 1: RTA development illustration
The effects of the RTA for PY 2022 DCE financial benchmarks will only be known after the performance year is complete because the final RTA will be set using all performance year data. The RTA decreases financial benchmarks because expenditures for the reference population are lower than initial projections. If the lower trend that CMMI observes for the DC National Reference Population is similar to the observed DC assigned beneficiary expenditure trend, then the average financial impact on DCEs of the RTA will be limited. However, we expect significant regional variation from national average expenditure patterns. Therefore, the effect of the RTA will vary across DCEs, with some benefiting and some being adversely affected.
How is the RTA calculated?
Calculation
On a monthly basis, CMMI releases to participating DCEs an updated Excel file containing expenditures for the DC National Reference Population. Users can interface with the Excel file and measure the emerging RTA. The RTA is the ratio of two trend measurements. The denominator is the prospective trend from PY 2019 to PY 2022 based on the Adjusted USPCC.2 The numerator is the emerging trend, measured as the year-to-date expenditures PBPY of the DC reference population (adjusted for incurred but not reported expenditures) expressed relative to the baseline year 2019 year-to-date. For example, if prospective total cumulative trend (not annualized) from PY 2019 to PY 2022 is 10% and the expenditures for the DC National Reference Population show a total cumulative trend (not annualized) of 8%, then the RTA would be calculated as 1.08 / 1.10 = 0.9820.
What is the reference population?
Central to the GPDC model is CMMI’s reliance on a DC National Reference Population to establish regional rates for the Kidney Care Choices (KCC)/GPDC rate book as well as to monitor and understand emerging trends. This reference population is defined as beneficiaries who meet five criteria: beneficiary is alive, beneficiary is enrolled in Medicare Part A, beneficiary is enrolled in Medicare Part B, beneficiary is not enrolled in a managed care plan (e.g., Medicare Advantage), Medicare is not the secondary payer, and the beneficiary is a U.S. resident.
There are material differences between the populations of beneficiaries assignable to a DCE or to the ACO Realizing Equity, Access, and Community Health (REACH) model and the DC National Reference Population used to set DCE and ACO REACH financial benchmarks. For example, the DC National Reference Population includes beneficiaries who do not have a Primary Care Qualified Evaluation and Management (PQEM) service, which all beneficiaries assigned to a DCE or REACH ACO must have, and the DC National Reference Population will include some beneficiaries that have no claims. Furthermore, the assignment period for performance years 2021 through 2022 overlaps with the public health emergency, possibly exacerbating these differences.
What might happen in 2023?
In PY 2023, the GPDC program will continue forward as the redesigned ACO REACH model. Current GPDC participants must maintain a strong compliance record and agree to meet all the ACO REACH model requirements by January 1, 2023, to continue participating in the ACO REACH model as ACOs. The ACO REACH financial benchmark is calculated in a way that is almost identical to the GPDC financial benchmark.
It appears likely at this time that an RTA will be applied to REACH ACOs’ financial benchmarks for PY 2023. The PY 2023 ACO REACH rate book published by CMMI in August 2022 shows an Adjusted USPCC for the Aged and Disabled population of $1,094.23, 5.4% higher than the preliminary $1,038.09 Adjusted USPCC used for PY 2022. Our review of emerging expenditure data leads us to believe that PY 2022 expenditures PBPM for the DC National Reference Population in PY 2022 will be materially lower than the $1,038.09 Adjusted USPCC. If the PY 2022 Adjusted USPCC does come in substantially lower than $1,038.09, then the amount of growth required for DC National Reference Population expenditures in order to remain within 1% of $1,094.23 would be substantially higher than 5.4%. As such, we feel the application of an RTA for PY 2023 is likely.
Conclusion
For the past two years, CMMI’s initial estimates of DCE financial benchmarks have been overstated due to high preliminary estimates of trend from the Adjusted USPCC, and the same appears likely to hold true for REACH ACOs in PY 2023. DCEs and REACH ACOs should use initial PY 2023 financial benchmarks with caution, and expect variation to occur. Monitoring the emerging PY 2023 data for ACO attributed beneficiaries and for the ACO REACH National Reference Population will be important for understanding PY 2023 ACO REACH financial performance.
Glossary
Retrospective Trend Adjustment (RTA): DC preliminary benchmarks rely on a prospective trend. The RTA is an adjustment made retrospectively to DCE financial benchmarks to account for actual emerging trends observed in the DC National Reference population.
Adjusted USPCC: CMMI adjusts the USPCC to exclude uncompensated care payments (UCC) and include hospice expenses.
DC National Reference Population: All Medicare FFS beneficiaries enrolled in Medicare Part A and Part B. Beneficiaries must reside in the United States and Medicare must be listed as the primary payer for the beneficiary to be included in the reference population.
1 CMS. PY2022 Financial Operating Guide: Overview, p. 16. Global and Professional Direct Contracting Model. Retrieved November 30, 2022, from https://innovation.cms.gov/media/document/gpdc-py2022-fin-op-guide-ovw.
2 CMS adjusts the USPCC to exclude uncompensated care payments (UCC) and include hospice expenses. See https://innovation.cms.gov/media/document/gpdc-py2023-ratebook-dev.