CMS-HCC V28: What Changed and Why It Matters for 2026

The V24 to V28 Transition Timeline

CMS-HCC V28 is the current version of CMS's Hierarchical Condition Category risk adjustment model, fully implemented for payment year 2026, that replaced the legacy V24 framework with a rebuilt ICD-10-native architecture featuring 115 HCCs, constrained disease hierarchies, updated ICD-10 diagnosis mappings, and recalibrated coefficients to more accurately predict Medicare Advantage beneficiary healthcare costs.

The CMS-HCC V28 model represents the most significant change to risk adjustment methodology in over a decade. CMS introduced V28 through a three-year phased transition designed to give Medicare Advantage plans time to adapt their operations. That transition is now complete.

• Payment Year 2024: CMS calculated RAF scores using a 67% V24 / 33% V28 blend. Plans operated under both models simultaneously, with V24 still carrying the majority weight.
• Payment Year 2025: The blend shifted to 33% V24 / 67% V28. V28 became the dominant model, and plans that had not adapted their coding and documentation practices began seeing material RAF impacts.
• Payment Year 2026: 100% V28. The transition is complete. All RAF scores are calculated exclusively under V28 rules, V28 HCC mappings, and V28 coefficients. V24 is no longer a factor in payment calculations.
• Why the Phase-In Mattered: CMS used the blended approach to smooth the financial impact of V28's significant structural changes. Plans that used the transition years to retrain coders, update analytics, and adjust documentation workflows are now positioned for V28. Plans that delayed are experiencing the full impact simultaneously.

Key Structural Changes

V28 restructured the CMS-HCC model at every level — from the number of condition categories to the way ICD-10 codes are mapped and how coefficients are assigned.

• HCC Count: 86 to 115: V28 expanded the number of Hierarchical Condition Categories from 86 to 115, organized into 26 disease families. This allows greater clinical granularity in distinguishing condition types and severity levels.
• ICD-10 Codes Mapped: 9,797 to 7,770: Despite more HCCs, fewer ICD-10 codes qualify for risk adjustment. CMS removed 2,294 codes from the crosswalk and added 268 new ones. The net reduction means codes that previously generated HCC value may no longer do so.
• ICD-10 Native Construction: V24 was originally built on ICD-9 and later crosswalked to ICD-10. V28 was constructed directly from ICD-10 classification, resulting in more clinically accurate mappings that align with current coding standards.
• Disease Family Reorganization: V28 reorganized conditions into 26 disease families, restructuring how conditions relate to each other hierarchically. Some V24 hierarchies were split, merged, or fundamentally redesigned.
• Updated Interaction Terms: Disease interaction factors were recalculated based on current claims data and the new HCC structure. Some V24 interactions were removed, others were modified, and new interaction definitions were added.
• New Demographic Segments: V28 introduced updated demographic segments with refined coefficients reflecting current utilization patterns across age, sex, dual-eligibility, and institutional status combinations.

The Constraining Methodology

Constraining is the single most consequential change in V28. It fundamentally alters how CMS assigns financial value to conditions within disease families.

• Definition: Under constraining, related HCCs within the same disease family are required to carry identical regression coefficients regardless of their clinical severity level. This means a severe diabetes complication generates the same risk weight as a moderate diabetes complication.
• Purpose: CMS implemented constraining to eliminate the financial incentive for severity upcoding within disease families. Under V24, plans could increase RAF scores by documenting (or over-documenting) more severe manifestations of conditions. Constraining removes that differential.
• Diabetes Example: Under V24, diabetes with complications carried a significantly higher coefficient than diabetes without complications — creating a financial incentive to find and document complications. Under V28, all diabetes HCCs (except pancreas transplant status) share the same coefficient of approximately 0.166.
• Clinical vs. Financial Distinction: Constraining does not eliminate clinical distinction. The hierarchy still identifies which HCC is the most severe, and CMS still reports severity levels. The change is purely financial — different severity levels within a family no longer generate different payment amounts.
• Affected Disease Families: Constraining applies to disease families where CMS identified significant variation in coding patterns between severity levels. Not all disease families are constrained — some retain differentiated coefficients across severity levels.
• Strategic Implication: Constraining shifts the value equation from "documenting the most severe version of each condition" to "documenting every condition across disease families." The financial return now comes from capturing conditions across multiple disease families rather than maximizing severity within a single family.

Condition-Specific Changes

Four disease areas experienced the most significant changes under V28, affecting the majority of Medicare Advantage populations.

Diabetes

• Constraining Impact: All diabetes HCCs now share a single coefficient of approximately 0.166. Under V24, diabetes with complications carried coefficients of 0.302-0.368 depending on severity. This represents the largest single-condition revenue reduction in V28.
• Mapped Code Reduction: Numerous diabetes ICD-10 codes were removed from the V28 crosswalk, particularly unspecified and peripheral vascular manifestation codes. Plans with large diabetic populations face compounding impact from both constraining and code reduction.

Congestive Heart Failure (CHF)

• Category Restructuring: V28 reorganized CHF into new HCC categories that distinguish heart failure types differently than V24. The mapping of specific ICD-10 codes to CHF HCCs changed, requiring updated coding practices.
• Documentation Specificity: V28 requires greater specificity in CHF documentation — systolic vs. diastolic distinction, acuity level, and ejection fraction documentation affect which HCC a CHF diagnosis maps to.

Chronic Kidney Disease (CKD)

• Staging Realignment: CKD HCCs were realigned with more granular staging under V28. The stage-to-HCC mapping changed, meaning the same CKD stage may map to a different HCC than under V24.
• ESRD Interaction: The relationship between CKD HCCs and End-Stage Renal Disease (ESRD) model components was updated, affecting dual-model calculations for members with advanced kidney disease.

Dementia

• Expanded Categories: V28 expanded dementia-related HCCs to capture more clinical specificity, including new categories for different dementia types and severity levels.
• Coding Opportunity: The expanded dementia categories represent one of the few areas where V28 creates new HCC capture opportunities that did not exist under V24. Providers documenting dementia type and severity may generate HCCs that V24 did not recognize.

Revenue Impact

CMS projected a 3.12% aggregate reduction in MA risk scores under V28. The actual impact at the plan level varies dramatically based on population composition and coding practices.

• Aggregate Projection: CMS estimated that V28 would reduce total MA risk scores by approximately 3.12% at the national level. This translates to billions in reduced CMS payments across the MA program.
• Plan-Level Variation: Early industry data shows individual plan RAF declines ranging from 5% to 30% for plans with identical patient populations. The variation depends on condition mix, coding specificity, and the proportion of revenue historically driven by within-family severity coding that constraining now equalizes.
• Diabetes-Heavy Plans: Plans with large diabetic populations and historically high diabetes complication capture rates face the steepest declines. A plan where 60% of members have diabetes-related HCCs may see disproportionate RAF reduction from constraining alone.
• Offsetting Opportunities: Plans that systematically expand cross-family condition capture — documenting conditions in disease families that were previously deprioritized — can partially offset the constraining-driven declines. The value has shifted from depth (severity within families) to breadth (conditions across families).
• Multi-Year Adjustment: Plans that used the 2024-2025 transition to adjust coding and documentation practices entered 2026 with partially adapted workflows. Plans that delayed face the full V28 impact in a single year.

What Organizations Should Do Now

With V28 now the sole model, organizations must ensure every aspect of their risk adjustment operations is calibrated to V28 rules.

• Validate V28 Crosswalks: Confirm that all coding and analytics systems use the current V28 ICD-10-to-HCC crosswalk. Any system still referencing V24 mappings will produce incorrect HCC assignments and inaccurate RAF projections. Use the RAF Score API to validate mappings in real time.
• Retrain on Constraining: Educate coders and providers that within-family severity no longer drives differential payment. Redirect documentation and coding effort toward capturing all active conditions across disease families rather than pursuing the most severe manifestation within a single family.
• Recalibrate Revenue Models: Update all revenue forecasting, budgeting, and bid models to reflect V28 coefficients, constrained disease families, and updated interaction terms. Models still using V24 assumptions will systematically overestimate revenue.
• Prioritize Cross-Family Capture: Implement analytics that identify members with documented conditions in only one or two disease families when clinical indicators suggest conditions in additional families. The highest-ROI activity under V28 is expanding the number of captured disease families per member.
• Audit ICD-10 Code Validity: Review submission data for ICD-10 codes that mapped to HCCs under V24 but no longer do under V28. These codes generate zero risk adjustment value and should be evaluated for replacement with codes that map to V28 HCCs where clinically appropriate.
• Monitor Coding Quality Continuously: Deploy real-time dashboards that track V28-specific metrics: constrained-family capture rates, cross-family breadth scores, V28 HCC mapping accuracy, and interaction term capture rates. Quarterly retrospective audits are insufficient under V28's compressed value landscape.