The Inflation Reduction Act (IRA) introduced several policies that shift financial liability for prescription drug costs across manufacturers, plans, beneficiaries, and the Centers for Medicare  & Medicaid Services (CMS). As of 2025, under the Part D redesign, beneficiary out-of-pocket (OOP) costs are capped at $2,000, while plan and manufacturer liability increase in the catastrophic phase. The IRA also established the Medicare Drug Price Negotiation Program, which allows CMS to negotiate Maximum Fair Prices (MFPs) for select high-spend drugs, with the first MFPs for Part D drugs taking effect in 2026.  

In response to these provisions, CMS implemented updates to its Part D risk adjustment model— the Prescription Drug Hierarchical Condition Category (RxHCC) model—effective in 2025. CMS uses the RxHCC model  to determine the expected prescription drug cost for both Medicare Advantage Prescription Drug plans (MA-PDs) and standalone Prescription Drug Plans (PDPs). Specifically, the RxHCC model determines expected gross drug spending for a particular Part D beneficiary based on certain demographic characteristics and their diagnosed medical conditions. CMS then compares this estimated spend to the average spend for a Part D beneficiary to determine a risk adjustment factor (RAF) for each individual. CMS uses the RAF to adjust Medicare’s direct subsidy payment to a health plan to reflect the relative health status of the beneficiaries enrolled in that plan.  

Impact of Risk Adjustment on Plans 

The RxHCC model directly impacts plan loss ratios. A loss ratio reflects the percentage of total premium that a plan spends on claims for its beneficiaries. Risk adjustment is intended to standardize loss ratios across plans by accounting for differences in beneficiary health status. Without risk adjustment, plans that have a comparatively higher proportion of beneficiaries with complex health conditions could have higher loss ratios (i.e., a higher percentage of premiums are spent on claims).  

Plans evaluate their expected loss ratios when making coverage decisions, such as whether a drug is on formulary or subject to utilization management criteria (i.e., prior authorization or step therapy). Due to its downstream impacts to plan offerings and drug coverage, accurate Part D risk adjustment is critical. 

Even before the implementation of the IRA, stakeholders raised several concerns about the predictive accuracy of the RxHCC model. First, MA-PDs have historically benefited from higher RAFs more than standalone PDPs. As recently noted by the  Medicare Payment Advisory Commission, this dynamic may exist in part because MA-PD plans have more tools (e.g., health risk assessments, medical chart reviews) than standalone PDPs to ensure that diagnoses are documented for a given beneficiary. Additionally, the RxHCC model is calibrated on historic drug costs and is typically updated only every two to three years. Due to the infrequent updates, the RxHCC model may be slow to incorporate the expected cost of new drug therapies.  

Changes to the Risk Adjustment Model  

Starting in calendar year (CY) 2025, CMS made several updates to the RxHCC model to address the implementation of the IRA and the differences in RAFs between MA-PDs and standalone PDPs. The model was modified to reflect changes to the Part D benefit under the IRA and apply separate normalization factors for MA-PDs and standalone PDPs. Combined, these changes are expected to result in stable year-over-year, lower loss ratios.  

Avalere Health analyses estimated that under the revised CY 2025 RxHCC model, loss ratios were lower than the CY 2024 risk adjustment model. For medications in some therapeutic areas like anticoagulants and antidiabetic agents, average loss ratios across MA-PDs and PDPs decreased. Average loss ratios went from values above 1—indicating a plan’s drug costs exceeded its revenues—under the CY 2024 model to under 1—indicating a plan’s drug costs were less than its revenues under the CY 2025 model. While updates to the RxHCC model may better align Part D plan payment with the new benefit design under the IRA, they do not eliminate all Part D plan challenges (e.g., potential downstream impacts of the IRA on patient utilization under an OOP cap).  

Notably, the updates to the RxHCC model do not address the lag in reflecting the costs of emerging drug therapies. The costs of new prescription drug therapies are only incorporated into the RxHCC model after the drug has been covered by Part D, beneficiaries begin to take the drug, and the data used to update the RxHCC model reflects the cost of the new drug. For certain therapeutic areas, this could result in increased loss ratios until the RxHCC model is updated. An Avalere Health analysis of beneficiaries with diagnoses of both Alzheimer’s disease and other forms of dementia found that the RxHCC model consistently underpredicted plan costs, which would lead to increased plan loss ratios. 

For 2026, the RxHCC model will also be updated to reflect MFPs for negotiated drugs. This update will likely lower RAFs for beneficiaries diagnosed with conditions treated by drugs selected for negotiation in 2026—which will increase plan loss ratios for the associated therapeutic areas.  

Next Steps 

As MFPs are incorporated into the risk adjustment model in CY 2026, Avalere Health is prepared to conduct risk adjustment scenario planning to help health plans and other stakeholders understand the financial impact of the updated risk adjustment model. These risk adjustment changes may have downstream effects on Part D plan loss ratios and, in turn, plan availability and formulary design.  

With the release of CY 2026 landscape and formulary files this fall, Avalere Health can help stakeholders assess changes in 2026 Part D plan availability, offerings, and formularies. For more information on how plan availability, offerings, and formularies for 2026 and beyond will impact your organization, connect with us. 

Methodology:  

Analyses referenced in the publication used medical claims and prescription drug event data for FFS and MA beneficiaries across 2019-2022. Avalere estimated plan payments using the RxHCC risk adjustment model, applying v24 (based on the prior benefit structure) and v25 (recalibrated to reflect the redesigned Part D benefit structure under IRA). To estimate gross plan liability on the cost-side, Avalere used its proprietary benefit redesign framework to apply IRA provisions to 100% PDE expenditure data for the plan year and simulated Part D plan stakeholder liability under the redesigned Part D benefit and projected to CY2025 using growth factors based on the Medicare Trustees Report. Avalere then calculated loss ratios by dividing gross/net plan liabilities by risk-adjusted plan payments estimated using the RxHCC model. This was performed separately using v24 and v25, enabling comparison of loss ratios across the two versions. No adjustments were made (i.e. for quality like in the minimum MLRs reported to CMS for Part D). Thus, these represent simple loss ratios. 

Several analyses in this publication utilized 100% percent Medicare fee-for-service (FFS) claims and Medicare Advantage (MA) Encounter data from 2021 and 100% Medicare PDE data from 2022 through a data use agreement (DUA) with CMS, accessed via the Chronic Condition Warehouse (CCW) Virtual Research Data Center. 100% Medicare PDE data for one analysis referenced utilized 2020 PDE and 2019 FFS claims data accessed by Avalere Health via a research collaboration with Inovalon, Inc. by a research focused CMS Data Use Agreement.