CMS Stars ratings drive Medicare Advantage bonuses, market positioning, and member growth. The measures depend on clinical and administrative data spanning years. Most US Medicare Advantage plans compute Stars against claims data warehouses today, but the architecture is starting to shift toward FHIR-based computation as the FHIR data investments from CMS-0057-F mature. Five FHIR-based Stars measurement patterns are worth knowing in 2026. For FHIR for analytics and Stars guides on this site, these are the practical patterns.
1. FHIR Measure Resource for CMS Quality Measures
The most direct pattern. CMS quality measures (the underlying spec for Stars) can be expressed as FHIR Measure resources with CQL-defined logic. The Measure resource captures the population criteria, numerator definition, denominator definition, and exclusions. Running the measure against FHIR clinical data produces MeasureReport resources with the computed results.
The pattern uses standard FHIR resources end to end. Implementations that store clinical data as FHIR resources can compute Stars measures directly without translating to a parallel claims-data model.
2. Hybrid FHIR Plus Claims Warehouse
A pattern that uses FHIR for clinical signal and the existing claims warehouse for administrative data. The Stars measure logic queries both sources, combining clinical resources (for measures like blood pressure control, A1c management, cancer screening) with claims-derived signals (encounter counts, service utilization).
The hybrid handles the transition state most large payers occupy in 2026: extensive existing claims warehouse infrastructure plus growing FHIR clinical data investments. The two stay separate but work together.
3. FHIR Subscription-Driven Real-Time Measurement
A pattern where FHIR Subscriptions feed near-real-time measurement updates. A new Observation arrives (BP reading from a primary care visit), and a Subscription notification updates the relevant Stars measure for that member. The dashboard reflects the latest measurement state without waiting for batch processing.
The pattern enables intervention timing. Care managers can see members who are about to fall out of measure compliance and intervene before the deadline. The trade-off is the operational complexity of running Subscription-driven measure updates at scale.
4. Da Vinci Risk-Based Contract IG Integration
A pattern that uses the emerging Da Vinci Risk-Based Contract IG to share Stars measure data between payers and providers. The pattern fits value-based care arrangements where the provider is responsible for measure performance and needs visibility into the same measures the payer is tracking.
Adoption is limited in 2026 because the IG is newer and provider-side support varies. The pattern signals where the Da Vinci portfolio is heading for value-based care.
5. AI-Augmented Measure Identification
A pattern that uses ML on FHIR clinical data to identify members who are likely to fall out of measure compliance based on care patterns. The model output is not the measure computation itself (which uses deterministic logic from the CMS spec); it is the prioritization layer that surfaces which members need attention.
For Stars measures with large eligible populations, the prioritization layer matters because care management capacity is limited. AI on FHIR data identifies the highest-leverage interventions.
The HEDIS Overlap
Stars measures and HEDIS measures overlap substantially. Most of the patterns above apply to HEDIS computation with minor adjustments. The differences come from measure-specific logic (HEDIS has measures Stars does not, and vice versa) and from reporting context (HEDIS is part of NCQA accreditation; Stars is CMS).
For HEDIS-specific platform options, the Best FHIR analytics platforms for HEDIS measure computation covers the leaders. For the risk adjustment patterns that complement Stars and HEDIS measurement, the Top 5 risk adjustment patterns using FHIR clinical data covers the parallel layer.
How to Pick the Pattern for Your Plan
The choice usually depends on the existing data architecture. Plans with strong existing claims warehouses use the hybrid pattern (Pattern 2) as the transition state. Plans with FHIR-native data investments use the FHIR Measure resource pattern (Pattern 1) directly. Plans with real-time intervention capability layer in Subscription-driven measurement (Pattern 3). AI augmentation (Pattern 5) applies regardless of the foundational pattern.
The patterns are not mutually exclusive. Production Stars programs in 2026 typically combine several: hybrid data architecture, FHIR Measure resource for the computation, AI prioritization for care management. The exact mix depends on the plan's specific capabilities and priorities.