Page 80

80

selected drugs would incentivize higher utilization, which the commenter believed would result in misaligned costs for plans. Response: CMS appreciates the commenters’ concerns about the impact of MFPs on plan liability estimates. With regard to the comments about use of rebates in the data, as stated in the CY 2026 Advance Notice, we substituted the gross drug costs for the ten selected drugs with their agreed-upon MFPs (prior to any additional price concessions for CY 2026, if applicable15), adjusted for inflation to the calibration year. For purposes of re-mapping PDEs in the model, the MFP is used as an effective new gross drug cost. Using the agreed-upon MFPs that CMS published rather than the gross drug costs on the PDE records for these drugs allows the model to more accurately reflect plan liability for CY 2026. Substituting MFPs for gross drug costs for 2026 creates stability as the PDE records reflecting these prices will eventually be incorporated as we continue to update the data years. We can account for MFPs because the amount for 2026 is known in advance. Rebates for non-selected drugs in 2026 are not known and will likely differ from rebates in the data currently available to CMS. We will continue to consider if and how to make adjustments for rebates and similar accountancies in future recalibrations. With regard to the comments about the impact of MFPs on utilization, we do not believe it is appropriate to model expected changes in behavior for incorporation into the model. We believe that modeling future behavior would result in error in the model and inaccurate predictions of relative costs. Instead, we believe it is prudent to calibrate on the most recent available data and wait to account for changes in utilization in future iterations of the model. We believe that it is still more appropriate to reflect MFPs in the model calibration than to not reflect them. If the model did not use the agreed-upon MFPs, it would likely overestimate the expected plan liability for conditions that are treated with these drugs. This would not only overestimate relative costs for RxHCCs with conditions that are prevalently treated using these drugs, but it would also likely underestimate relative costs for RxHCCs for which treatment for the conditions is not associated with these drugs. As a result, we are finalizing the models reflecting MFPs as proposed. Comment: A few commenters expressed concern about the updated RxHCC model’s impact on beneficiaries taking high-cost specialty drugs. Some of these commenters expressed appreciation for CMS providing predictive ratios for the model, but also expressed a belief that the model does not fully predict costs or risk for beneficiaries by using historical data. One of these commenters said that the IRA benefit changes increase the proportion of revenue from risk- adjusted payments, so the financial impact on mismatches between risk scores and expected costs could become greater, which the commenter believed could result in plans being disincentivized to cater to populations with high variability in drug costs. Another of these commenters suggested that the model should be refined to address high-cost RxHCCs,

15 Please see the Medicare Drug Price Negotiation Program: Revised Guidance for further discussion of MFP methodology for initial price applicability year 2026, including how agreed-upon MFPs are prior to any additional price concessions.