Today, the Centers for Medicare & Medicaid Services (CMS) released Part I of the 2019 Advance Notice of Methodological Changes for Medicare Advantage Capitation Rates and Part D Payment Policies (the Advance Notice), which contains key information about proposed updates to the Part C Risk Adjustment Model and the use of encounter data.
The 2019 Advance Notice is being published in two parts this year due to requirements in the 21st Century Cures Act, which mandated certain changes to the Part C risk adjustment model and a 60 day comment period for these changes. Changes to other payment methodologies proposed for the following calendar year that are typically contained in the Advance Notice only require a 30 day comment period and will be released in accordance with that statutory deadline. The payment policies for 2019, proposed in both in Part I and Part II of the Advance Notice, will be finalized in the annual Rate Announcement. To be assured consideration, comments on the proposals announced today should be submitted by March 2, 2018.
2019 Part C Risk Adjustment Model proposal
The 21st Century Cures Act amended the Social Security Act by, in part, requiring CMS to make improvements to risk adjustment for 2019 and subsequent years. In response to these requirements, we are proposing changes to the CMS-HCC Risk Adjustment model that is used to pay for aged and disabled beneficiaries enrolled in Medicare Advantage plans. These proposals reflect changes to risk adjustment required by the 21st Century Cures Act, including an evaluation of adding mental health, substance use disorder, and chronic kidney disease conditions to the risk adjustment model and making adjustments to take into account the number of conditions an individual beneficiary may have, as well as a variety of additional technical updates. Further, the 21st Century Cures Act requires that CMS fully phase in the required changes to the risk adjustment model by 2022. We are therefore proposing to begin the phase in of this new model in 2019, starting with a blend of 75% of the risk adjustment model used for payment in 2017 and 2018 and 25% of the new risk adjustment model proposed.
For 2019, CMS is proposing a model that includes additional mental health, substance use disorder, and chronic kidney disease conditions in the risk adjustment model.
With respect to taking into account the number of conditions an individual beneficiary has, in Part 1 of the Advance Notice we describe a proposed new risk adjustment model and discuss an alternative model. The model we are proposing – the “Payment Condition Count model” – takes into account the number of conditions that a beneficiary has, only among the conditions that are included in the payment model. The model discussed as an alternative – the “All Condition Count model” – takes into account all conditions that a beneficiary has, including both those in the payment model and those not in the model.
The charts below indicate the range of contract-level impacts of each of these CMS-HCC Risk Adjustment models on Medicare Advantage risk scores. Overall, while the experience of individual plans would vary, the Payment Condition Count model is projected to increase MA risk scores by 1.1%, while the All Condition Count model would decrease MA risk scores by -0.28%. Under the Payment Condition Count model, the change in MA contracts’ risk scores is generally positive and less varied than the All Condition Count model. The change in MA contracts’ risk scores under the All Condition Count model is more varied, with both negative and positive changes.
Payment Condition Count model – Percent change in MA contract-level risk scores
This graph displays the estimated percent change in payment for 446 Medicare Advantage contracts. Each line indicates the Payment Condition Count model’s estimated impact on payment for 1 contract.
All Condition Count model – Percent change in MA contract-level risk scores
This graph displays the estimated percent change in payment for 446 Medicare Advantage contracts. Each line indicates the All Condition Count model’s estimated impact on payment for 1 contract.
Using Encounter Data
The model we are proposing in Part 1 of the Advance Notice also makes technical updates, including calibrating the model with more recent data, selecting diagnoses with the same method used for encounter data, and supplementing encounter data used in payment with inpatient data submitted to the historical risk adjustment data collection system (the Risk Adjustment Processing System (RAPS)).
CMS calculates risk scores using diagnoses submitted by Medicare FFS providers and by Medicare Advantage organizations. Historically, CMS has used diagnoses submitted into CMS’ RAPS by Medicare Advantage organizations. In recent years, CMS began collecting encounter data from Medicare Advantage organizations, which also includes diagnostic information. In 2016, CMS began using diagnoses from encounter data to calculate risk scores, by blending 10% of the encounter data-based risk scores with 90% of the RAPS-based risk scores. For 2017 and 2018, CMS continued to use a blend to calculate risk scores, by calculating risk scores with 25% encounter data and 75% RAPS in 2017, and 15% encounter data and 85% RAPS in 2018. For 2019, CMS proposes to calculate risk scores by adding 25% of the risk score calculated using diagnoses from encounter data and FFS diagnoses with 75% of the risk score calculated with diagnoses from RAPS and FFS diagnoses. CMS is also proposing to implement the phase-in of the new risk adjustment model by calculating the encounter data-based risk scores exclusively with the new risk adjustment model, while maintaining use of the current 2018 risk adjustment model for calculating risk scores with RAPS data.