For the third installment of our semi-regular series that recommends articles, white papers, or issue briefs of importance to RISE members, we turn to a recent report written by Avalere about the impact of Medicare Advantage (MA) risk adjustment model changes for payment year 2020. The report is essential reading in the wake of the Centers for Medicare & Medicaid Services’ Dec. 19, 2018 release of Part 1 of the 2020 Advance Notice of Methodological Changes for MA Capitation Rates and Part D Payment Policies. Comments about the CMS proposals must be submitted by Feb. 19. The agency intends to publish the final 2020 rate announcement by April 1.
A recent analysis from Avalere estimated the impact of MA risk adjustment model changes for payment year 2019, but the report applies to the CMS 2020 proposal as well.
CMS originally proposed the Payment Condition Count (PCC) model for payment year 2019 but postponed adding condition counts to the model and instead implemented the clinical updates. CMS has now re-introduced the PCC with condition counts and proposed its full implementation for payment year 2020. The proposed changes are included in Part I of the 2020 Advance Notice of Methodological Changes for Medicare Advantage Capitation Rates and Part D Payment Policies (the “Advance Notice”) and, according to CMS, aim to better align risk adjustment payments to the actual costs of providing care to enrollees to conform with requirements of the 21st Century Cures Act.
The full PCC payment model included four new HCCs and a broadened definition of one HCC and added the PCC model to account for the incremental number of HCC conditions each beneficiary has documented.
Avalere captured the impact of the full PCC model while estimating 2019 payment impacts and this analysis applies to the CMS 2020 proposal as no changes were made to the PCC model for 2020. CMS has also asked for comment on a variant of the PCC model that adds three additional HCCs for dementia and lower-severity skin pressure ulcers. The Avalere paper does not analyze the variant of the model including the three additional HCCs.
The consulting firm analyzed MA encounter data for 1.1 million beneficiaries enrolled in 58 contracts from 24 MA organizations in 2015. The analysis compared HCCs, risk scores, and payment under the current model (V22), the full PCC model with the count factor included (Proposed V23), and the final 2019 model (Final V23).
The report found that although the overall average impact of the proposed changes was small, there were winners and losers at the plan level. Overall risk scores in the new version of the model increased only slightly by 0.78 percent in the study, which is less than the 1.1 percent projected by CMS. However, most plans (79 percent) showed an increase in mean risk score ranging from +0.005 to +0.023 (0.31 percent to 1.8 percent) while mean risk scores were lower for five plans ranging from -0.019 to -0.003 (-0.96 percent to -0.18 percent).
“Even small changes in risk scores can have a significant impact on plan reimbursement,” says Sean Creighton, a managing director at Avalere and co-author of the study.
Using an average bid rate of $800, a 0.01 increase in mean risk scores translates to $8 per member per month (PMPM), according to Avalere. For the average size plan in the Avalere study (n=45,760 members), this equates to a potential reimbursement increase of $4.4M per year.
The proposed new count factor had little impact on mean risk scores. Compared to the current payment model, risk scores with the count factor included were identical for patients with 0-3 HCCs, slightly lower for patients with 4-5 HCCs, and slightly higher for patients with 6 or more HCCs. The largest increase was for beneficiaries with 10 or more HCCs, but their prevalence in the population was only three percent.
MA plans with more beneficiaries in the impacted populations could see substantial changes in risk scores and payments, according to the report. The non-linearity of the impacts with patients below 5 HCCs having no change in score or a negative impact should be of concern to health plans as this is not congruent with paying more for enrollees with complex conditions. Plan-specific simulation modeling could help anticipate the financial impact of the upcoming changes.
Editor's note: Sean Creighton is a RISE board member and chair of our risk adjustment policy advisory committee. He will be one of 50 speakers at the 13th Annual RISE Summit in Nashville, March 17-19.