The Centers for Medicare & Medicaid Services (CMS) uses a risk adjustment process that is applied to payments made to Medicare Advantage Organizations (MAOs), which administer health plans for Medicare Advantage (MA) beneficiaries. This risk adjustment process creates risk scores designed to account for differences in health status among MA plan enrollees. Risk scores are calculated for all Medicare beneficiaries, regardless of whether the beneficiary was enrolled in Original Medicare (also known as Fee-For-Service Medicare) or MA during the payment year.
For a given payment year, the research risk score files include an Operational Payment Base File and Model Output Files (MOFs) from a mix of six potential risk models. The MOFs contain raw risk scores calculated prior to any adjustments made for payments. Each payment year has specific risk models that are based on that year’s risk adjustment methodology; payment years may use a blend of MOF risk scores as inputs to calculate the operational payment risk scores.
Depending on the payment year, the risk models that are used for MAO payments may include:
- ESRD V21 scores (based on model calibration year)
- Part C V21 scores
- Part C V22 scores
- Part C V23 scores
- Part C V24 scores
- Part D V05 scores
The table below lists the risk models used in each payment year's risk adjustment methodology:
| Model Output Files | 2019 | 2020 | 2021 |
| ESRD V21 | X | X | X |
| V21 | X | - | - |
| V22 | X | X | X |
| V23 | X | - | - |
| V24 | - | X | X |
| RX | X | X | X |
Each MOF raw risk score file is based upon different model coefficients and different diagnosis code inputs for beneficiary clinical information, such as MA encounter data, Original Medicare claims, and the Risk Adjustment Processing System (RAPS), a predecessor system to encounter data collection. As a result, some MOFs will have more than one record per person because the risk model generates raw scores based on multiple sources of clinical diagnoses.
MOFs also contain Hierarchical Condition Categories (HCCs), which represent groups of diagnosis codes for related disease categories.
The research risk score files are designed to be used with other Chronic Conditions Warehouse (CCW) data products that contain a wide range of demographic and Medicare coverage information and can be linked to other CCW files using the Beneficiary ID variable.
What does this file include? (variable highlights)
- Model output raw risk scores
- Model-specific Hierarchical Condition Codes (HCCs/RxHCCs)
- Other model clinical and care setting indicators
- Model-specific disability information
- Demographic variables
Special considerations:
- Reference the CCW Medicare Risk Score Files User Guide for more information on the operational payment risk score base file and model output files.
- Reference the Report to Congress: Risk Adjustment in Medicare Advantage for a discussion of risk adjustment in the Medicare Advantage program.