CY2027 RxHCC Model Changes: What CMS Actually Finalized
CMS released the CY2027 Rate Announcement on April 6, 2026, finalizing the most significant Part D risk adjustment update in years. This page covers what is final — not what was proposed or speculated. Every claim below cites the Rate Announcement directly.
The five finalized changes
- Separate model segments for MA-PD and PDP populations. For the first time, the disease coefficients themselves differ by plan type — not just the normalization factors. CMS finalized this to improve predictive accuracy for both MA-PD and standalone PDP populations (Rate Announcement, Attachment VIII; CMS fact sheet).
- Recalibration on 2023 diagnoses / 2024 expenditures. The updated model uses more recent data years and reflects the IRA's CY2027 Part D benefit design, including the phased-in manufacturer discount for specified small manufacturers and gross-drug-cost adjustments for maximum-fair-price (MFP) drugs in IPAY 2026.
- Audio-only diagnoses excluded. Diagnoses collected from audio-only services no longer count toward Part D risk scores.
- Unlinked chart review records (CRRs) excluded. Diagnoses from chart review records that cannot be linked to an encounter are excluded. CMS modified the final policy: the exclusion does not apply to members who switch from one MA plan in the diagnosis year to another plan in the payment year.
- Separate normalization factors continue. For the 2023/2024 calibration: 1.109 for MA-PD plans, 1.005 for PDPs, calculated by multiple linear regression over average 2020–2024 FFS risk scores with a pre/post-COVID flag. PACE organizations use a 50/50 blend of the 2023/2024 and 2018/2019 calibrations, with a 1.237 normalization factor on the latter.
Why the MA-PD / PDP split matters commercially
Until 2027, a plan could model its Part D risk revenue with one coefficient set. That ended. The same diagnosis — heart failure, diabetes with complications, hepatitis C — now pays differently depending on whether the member is in an MA-PD or a standalone PDP. Plans that priced 2027 bids off a single coefficient set mis-priced. See the actual numbers in our 2027 coefficients table and the segment mechanics in MA-PD vs PDP segments.
What did not change
- The hierarchical structure of the model is intact — the same disease hierarchies apply across all RxHCC models (Table VIII-13; see our hierarchy explainer).
- The MA coding-pattern adjustment remains at the statutory minimum of 5.90% (Part C context, unchanged from CY2026).
- PACE organizations transition gradually via the 50/50 blend rather than a hard cutover.
What plans should do now
- Re-run 2027 bid assumptions with the correct plan-type segment — the coefficient deltas are material for oncology, hepatitis C, and cardiac categories.
- Audit documentation pipelines for audio-only and unlinked-CRR diagnoses that will no longer count.
- Model the normalization asymmetry: MA-PD scores divide by 1.109, PDP by 1.005 — a ~10-point relative swing between plan types before any coding change.
*CuraFi runs population-level 2027 impact analyses for MA-PD plans, PDPs, and ACOs. Talk to us.*
Frequently asked questions
Is the CY2027 RxHCC model final?
Yes. CMS finalized it in the Rate Announcement released April 6, 2026, following the January 26, 2026 Advance Notice.
Where are the official coefficients?
Tables VIII-1 through VIII-12 of the CY2027 Rate Announcement, plus machine-readable files in the 2027 model software on CMS.gov. We reproduce the continuing-enrollee tables here.
Does this affect RADV audits?
Indirectly. The audio-only and unlinked-CRR exclusions change what counts as a valid diagnosis for payment — documentation that fed those categories is now both non-payable and a compliance flag. See RADV preparation for Part D.