Proposed Revisions to Part D Risk Adjustment Model in 2025

  • This page as PDF


Due to increasing the portion of plan payments subject to risk adjustments in 2025, CMS proposed aligning the risk adjustment model with Part D redesign.

On January 31, the Centers for Medicare & Medicaid Services (CMS) released the Calendar Year (CY) 2025 Medicare Advantage (MA) and Part D Advance Notice. The Advance Notice is an annual regulatory document that describes the agency’s proposed payment and coverage policies for MA and Part D plans for the upcoming plan year.

CMS proposed updates to the Part D risk adjustment (RxHCC) model to reflect the impact of the new 2025 benefit structure per the Inflation Reduction Act (IRA). Under the IRA benefit changes, plan liability will increase across a variety of therapeutic areas, particularly for drugs with spending that falls primarily in the catastrophic phase and for low-income subsidy (LIS) enrollees. The portion of plan payments that are risk adjusted will increase substantially relative to pre-IRA Part D redesign levels. This will make the accuracy of the risk adjustment model even more important for plans when designing their formularies and benefits.

CMS proposed the following modifications to the risk adjustment model, which are likely to significantly impact plan payments:

  • Update the model to reflect the new 2025 IRA benefit structure
  • Apply separate normalization factors to MA Prescription Drug Plan (MA-PD) vs. standalone Prescription Drug Plan (PDP) risk scores
  • Remove 95 diagnosis codes
  • Update data years to reflect 2021 diagnoses and 2022 expenditures, which would reflect the most recent year of complete spending data available.
    • CMS also presented a model reflecting the current data years used for the model (2018 diagnoses and 2019 expenditures)

CY 2025 RxHCC Model Recalibration to Reflect IRA Benefit Redesign

The risk adjustment model is calibrated so the mean risk score is 1.0. For example, a risk score of more than 1.0 indicates higher predicted plan liability and spending for a given beneficiary relative to average, while a risk score below 1.0 indicates lower predicted liability and spending.

Each demographic and condition category (also known as RxHCC) in the model has a risk score factor representing the additional contribution of that category to the total risk score for the beneficiary. In recalibrating the model for CY 2025, the relative risk score factors for each category changed from the previous version of the model. The changes in the relative factors were not uniform, with some factors having larger changes than others. While the percentage change between the new and previous model is notable, it is important to also examine the magnitude of the risk score factor. The sum of the factors creates the total risk score for an individual and, when applied to the plan bid, determines the direct subsidy payment amount.

The RxHCC risk model is designed to generate a separate risk score for 8 subsets of beneficiaries depending on whether they are institutionalized, low-income (LIS), older or younger than age 65, or enrolled in Medicare fewer than 12 months (new enrollee). The 8 model segments are Community, Non-Low Income, Age ≥ 65; Community, Non-Low Income, Age<65; Community, Low Income, Age ≥ 65; Community, Low Income, Age < 65; Institutional; Non-Low Income (new enrollee); Low-Income (new enrollee); and Institutional (new enrollee). Each model has unique risk score factors for the demographic and condition category components.

The top 10 percentage changes (positive or negative) in relative risk score factors from the current to the proposed model (Community, Non-LIS, ≥ 65 Model) are presented in Table 1.

Table 1: Top 10 Largest Proposed Changes to CY 2025 RxHCC Relative Risk Score Factors, Community, Non-LIS, ≥ 65 Model


RxHCC Relative Risk Factor, Current Model Relative Risk Factor, Proposed CY 2025 Model Percent Change
RXHCC225 Cystic Fibrosis* 2.109 8.025 281%
RXHCC41 Lysosomal Storage Disorders* 1.468 4.566 211%
RXHCC154 Amyotrophic Lateral Sclerosis and Other Motor Neuron Disease* 0.431 1.054 145%
RXHCC153 Myasthenia Gravis and Other Myoneural Disorders* 0.658 1.533 133%
RXHCC157 Chronic Inflammatory Demyelinating Polyneuritis* 1.865 4.327 132%
RXHCC18 Secondary Cancer of Lung, Liver, Brain, and Other Sites 1.202 2.623 118%
RXHCC19 Leukemias and Other Hematologic Cancers 1.202 2.623 118%
RXHCC100 Immune Thrombocytopenic Purpura* 0.157 0.334 113%
RXHCC96 Acquired Hemolytic, Aplastic, and
Sideroblastic Anemias
0.368 0.775 111%
RXHCC395 Stem Cell, Including Bone Marrow, Transplant Status/Complications 2.111 4.362 107%

*Represented RxHCCs also have the largest percentage changes in relative risk score factors in the Community, LIS ≥ 65 Model.

Six of the 10 RxHCCs with the largest percentage changes in relative factors in the Community, Non-LIS, ≥ 65 Model also had the largest changes in the Community, LIS, ≥ 65 Model. Notable increases in the LIS model not represented in Table 1 include RXHCC160 Huntington Disease; RXHCC21 Lymphomas and Other Hematologic Cancers; RXHCC316 Psoriasis, Except with Arthropathy; and RXHCC355 Narcolepsy and Cataplexy.

Different Risk Score Normalization Factors Will Be Calculated for PDPs vs. MA-PDs

CMS proposed estimating normalization factors separately for the PDP and MA-PD populations; previously, both populations were used to estimate a single normalization factor. The split for the CY 2025 model reflects the widening differences in the risk score growth rates between fee-for-service (FFS) and MA populations. The model tends to overpredict risk scores for MA-PDs and underpredict PDPs; for PDPs, this means actual liability is higher than predicted liability. As a result, using these underpredicted risk scores in the bid process means that PDP bids could be lower than the actual liability these plans will incur.

Additionally, the IRA is expected to disproportionately impact PDPs relative to MA-PDs. Separate normalization factors are one tool to mitigate some of these varying impacts of the IRA and address differences in risk score rates between FFS and MA.

Updates to Diagnosis Codes

Of the more than 5,000 diagnosis codes in the RxHCC model, 95 codes were cut, and 2 codes were added across 10 RxHCCs. The most notable cuts were to RxHCC112 Dementia, Except Alzheimer’s Disease, which had a 69% reduction (from 100 codes to 31) (see Table 2).

Table 2. Proposed RxHCCs Code Removal in CY 2025 Model

>RxHCCs with Proposed Removed Codes in CY 2025 Number of ICD-10 Codes Cut in CY 2025* % Reduction from Previous Number of Codes
RxHCC112 Dementia, Except Alzheimer’s Disease 69 69%
RxHCC188 Coronary Artery Disease 9 10%
RxHCC22 Prostate, Breast, Bladder, and Other Cancers and Tumors 4 1%
RxHCC96 Acquired Hemolytic, Aplastic, and Sideroblastic Anemias 4 12%
RxHCC191 Ventricular Septal Defect and Major Congenital Heart Disorders 4 15%
RxHCC261 Dialysis Status, Including End Stage Renal Disease 3 12%
RxHCC83 Rheumatoid Arthritis and Other Inflammatory Polyarthropathy 1 0.2%
RxHCC99 Immune Disorders 1 2%

Note: 1 code was added to RxHCC148 Mild or Unspecified Intellectual Disability/Developmental Disorder to make 38 diagnosis codes in the RxHCC; 1 code was added to RxHCC84 Systemic Lupus Erythematosus and Other Systemic Connective Tissue Disorders to make 55 diagnosis codes in the RxHCC.

Part D plans consider the accuracy of risk-adjusted payments when designing their benefits and formularies. For example, risk adjustment that underpredicts spending for a given therapeutic area or beneficiary group can create incentives for plans to narrow formularies for certain conditions in order to mitigate financial losses. With risk adjustment being applied to a larger dollar value of Part D plan payments under the IRA, the proposed model changes are likely to substantially affect plan payments, formularies, and benefit designs. It will be critical for stakeholders to assess the impacts of these model changes on their patient populations alongside other IRA impacts (e.g., changes in beneficiary utilization) in 2025 and beyond.

To learn more about payment and coverage policies for MA and Part D plans and how they affect your business, connect with us.

Webinar | A Closer Look at Patient Support On June 6 at 2 PM ET, Avalere experts will explore how potential implications of the Inflation Reduction Act (IRA)’s out-of-pocket cap, in addition to other key regulatory and policy activities shaping benefit design and patient cost-share (e.g., EHB), could impact patient commercial and foundation assistance. Learn More
From beginning to end, our team synergy
produces measurable results. Let's work together.

Sign up to receive more insights about Federal and State Policy
Please enter your email address to be notified when new Federal and State Policy insights are published.

Back To Top