Risk Adjustment

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

Medicare Advantage

Determines monthly capitation payments to MA plans based on enrollee RAF scores

MSSP

Risk-adjusts ACO benchmarks; 3% cap on risk score growth prevents gaming

ACO REACH (Standard)

Uses CMS-HCC prospective model for benchmark calculation

ACO REACH (High Needs)

Uses CMMI-HCC concurrent variant for real-time acuity capture

LEAD Model (2027)

Will use CMS-HCC with reformulated benchmarks and new risk arrangements

Part D (RxHCC)

Separate RxHCC model for drug costs uses similar HCC architecture