The Centers for Medicare and Medicaid Services (CMS) continues to refine how it pays Medicare Advantage Organizations (MAOs) and stand-alone Prescription Drug Plans through updates to Part C (CMS-HCC) and Part D (RxHCC) risk adjustment models. This paper examines the potential impacts of the proposed and finalized 2027 changes—including the exclusion of audio-only telehealth and unlinked chart review records (CRRs) from risk score calculations, recalibrated coefficients, and independent MAPD/PDP model calibrations—and quantifies their expected financial impact across plan types, markets, and member populations.
While CMS did not ultimately implement the full slate of proposed 2027 Part C model changes, the analysis offers a forward-looking view of what MAOs could face if similar proposals are finalized in future years. For Part D, the absence of 2027 Maximum Fair Price (MFP) drug adjustments in the model creates a notable mismatch between projected and actual plan costs—with meaningful implications for conditions like diabetes, cystic fibrosis, and severe persistent asthma.
Highlights include:
- Part C impacts vary widely by plan type and population: We expect Dual and non-Full Dual to face the largest average risk score decreases, while Institutional and New to Medicare members could see risk score increases on average.
- Part D divergence between MAPDs and PDPs: Fully independent model calibrations and separate normalization factors are expected to drive materially different outcomes by plan type.
- MFP drug misalignment creates overprediction risk: Because 2027 MFP drug pricing was not incorporated into the RxHCC model calibration, risk scores for affected conditions could overstate expected plan costs, while conditions outside the MFP scope could be comparatively underpredicted.
- Unlinked CRR exclusion impact may be muted in practice: Though the proposed removal of unlinked CRRs could reduce risk scores by an estimated 1.5% on average, MAOs that proactively link supplemental records to valid encounters should see significantly lower realized impacts.