CMS-HCC Deep Dive
Medicare Advantage & MSSP model mechanics, RAF score calculation, and the V28 transformation
Source: CMS, MedPAC, RTI International • Last updated March 2026
Key Takeaway
The CMS-HCC model is the financial engine of Medicare Advantage. It translates patient diagnoses into Risk Adjustment Factor (RAF) scores that directly determine how much CMS pays each plan per enrollee per month. Understanding how diagnoses flow through HCC mapping, hierarchies, normalization, and the coding intensity adjustment is essential for any organization in value-based care.
How CMS-HCC Works
Hierarchical Condition Categories
The model classifies ICD-10-CM diagnosis codes into Hierarchical Condition Categories (HCCs). Under V28, approximately 7,770 ICD-10 codes map to 115 HCCs. Each HCC carries a specific cost weight (coefficient) derived from regression analysis of Medicare FFS claims data.
The Hierarchy
When a patient has both a less severe and a more severe manifestation of the same disease (e.g., diabetes without complications and diabetes with kidney complications), the model only counts the more severe condition. This prevents double-counting while accurately reflecting disease burden.
Prospective Model
CMS-HCC is prospective: it uses diagnoses from the prior calendar year (the “data collection year”) to predict costs in the following year (the “payment year”). This means every HCC must be re-documented annually — the “recapture problem” that drives much of risk adjustment strategy.
Disease Interactions
The model includes disease interaction terms that increase the predicted cost when certain combinations of conditions are present together (e.g., diabetes + heart failure). These interactions reflect the clinical reality that comorbid patients cost more than the sum of individual conditions would suggest.
CMS-HCC Sub-Models
CMS runs different model segments for different populations, each with its own coefficients.
Community — Non-Dual Aged
Age 65+, not Medicaid-eligible, living in the community. The largest segment by enrollment.
Community — Full Dual Aged
Age 65+, eligible for both Medicare and Medicaid. Higher average risk scores due to socioeconomic factors.
Community — Partial Dual Aged
Age 65+, partial Medicaid benefits. Between non-dual and full dual in average acuity.
Community — Disabled
Under age 65, eligible for Medicare due to disability. Distinct cost patterns from the aged population.
Institutional
Residing in long-term care facilities. Highest average risk scores and per-capita costs.
New Enrollee
First year in MA without prior diagnosis data. Uses demographics-only scoring until claims history builds.
Programs Using CMS-HCC
Determines monthly capitation payments to MA plans based on enrollee RAF scores
Risk-adjusts ACO benchmarks; 3% cap on risk score growth prevents gaming
Uses CMS-HCC prospective model for benchmark calculation
Uses CMMI-HCC concurrent variant for real-time acuity capture
Will use CMS-HCC with reformulated benchmarks and new risk arrangements
Separate RxHCC model for drug costs uses similar HCC architecture