Overview
Risk adjustment exists to solve a structural problem in prospective payment: if plans or providers are paid the same per-member-per-month regardless of member acuity, sicker members become financially unattractive and plans would rationally avoid them. Risk adjustment corrects this by scoring each member's expected cost using documented diagnoses and demographics, then scaling payment accordingly. It is the financial mechanism that makes capitated, value-based payment sustainable in populations with heterogeneous health status.
The dominant US framework is the hierarchical condition category (HCC) model. CMS-HCC, used in Medicare Advantage, and HHS-HCC, used in the ACA marketplace, each map ICD-10-CM diagnosis codes to condition categories arranged in clinical hierarchies. A member's Risk Adjustment Factor (RAF) is the sum of category weights plus demographic weights (age, sex, Medicaid dual-eligible, institutionalized status) and disease interactions. A healthy 68-year-old woman might score 0.4; a 74-year-old with diabetes with complications, CHF, and COPD might score 2.8 — seven times the base capitation.
Risk adjustment reshapes RCM in three distinct ways. First, diagnosis capture becomes a year-round workflow rather than an annual billing formality: conditions must be documented and coded every calendar year, even if stable, because CMS-HCC resets annually. Second, clinical documentation improvement programs add a risk-capture dimension — coders and physicians are trained on MEAT criteria (Monitor, Evaluate, Assess, Treat) that demonstrate the condition was actively addressed during the encounter. Third, coding audit stakes rise: adding unsupported HCC codes inflates capitation and is a False Claims Act violation; CMS Risk Adjustment Data Validation (RADV) audits recover tens of millions in per-plan overpayments annually and sustained False Claims Act settlements with MA plans have reached hundreds of millions.
The operational discipline is specific. Coders scrub prior-year HCC codes each January — suspect codes that cannot be re-validated should be dropped, and dropped codes require chart abstraction to re-establish the diagnosis in the current year. Face-to-face encounter requirements apply: a lab result or nursing note is not sufficient to support an HCC; the condition must be addressed by an acceptable provider type at a face-to-face visit. Telehealth was permitted during the public health emergency and remains for parity-classified services under current CMS guidance.
For providers in value-based contracts, under-coding is nearly as damaging as over-coding. A diabetic patient whose neuropathy is treated but not coded to specificity (E11.42 Type 2 diabetes mellitus with diabetic polyneuropathy) drops the RAF from roughly 0.45 to 0.10, taking hundreds of dollars per member off the table for a condition the practice actually managed. Programs that invest in MEAT-criteria physician education and year-round gap-closure reports typically lift RAF accuracy by 8–15%, translating directly to shared-savings dollars.
Formula
Risk Adjustment is calculated as:
Risk Adjustment Factor (RAF) = Σ(HCC category weights) + demographic weights + disease interactions. Payment = Base capitation × RAF × geographic adjustment.Industry benchmark
CMS Medicare Managed Care Manual Ch. 7. CMS-HCC v28 model (phased in through 2025). HHS-HCC model maintained by CCIIO. RADV audit methodology published annually. Industry: well-managed MA populations typically operate at RAF 1.0–1.2; institutionalized SNP populations RAF 2.5+.
Worked example
A Medicare Advantage plan has 12,000 attributed members. Average RAF = 1.05. Base CMS capitation = $1,050 PMPM. Monthly MA revenue = 12,000 × $1,050 × 1.05 = $13.23M. After a structured year-round risk-capture program, RAF accuracy rises to 1.14 (9-point lift from previously un-coded but clinically active conditions — diabetic complications, CKD stage 3, depression). New monthly revenue = 12,000 × $1,050 × 1.14 = $14.36M — an incremental $1.13M/month or $13.6M annualized, against program cost of ~$1.8M.
Frequently asked questions — Risk Adjustment
Do HCC codes from last year carry forward?
No. CMS-HCC resets every calendar year. A condition documented in 2025 must be re-documented and coded again in 2026, supported by a face-to-face encounter where the provider addresses the condition per MEAT criteria. This is the most common source of lost risk score year-over-year.
What's the difference between CMS-HCC and HHS-HCC?
CMS-HCC is used for Medicare Advantage and Part C payment; HHS-HCC is used for ACA marketplace risk-transfer pool. Condition-category definitions overlap substantially but weights, hierarchies, and model versions differ; they are not interchangeable and each member population uses its designated model.
How is risk adjustment audited?
CMS conducts annual Risk Adjustment Data Validation (RADV) audits: a statistical sample of members is selected, medical records are requested, and coders verify each submitted HCC against documentation. Unsupported codes trigger extrapolated recoveries — small per-chart errors scale to multi-million-dollar plan-level settlements.
Is adding HCC codes fraud?
Adding unsupported HCC codes to inflate capitation is a False Claims Act violation. DOJ has reached multi-hundred-million-dollar settlements with MA plans. Legitimate programs capture diagnoses already supported in the record; they never add codes to meet a target.
Disclaimer
This glossary entry is operational reference for revenue-cycle and medical-billing professionals. It is not legal, clinical, or contractual advice. Industry benchmarks cite named public sources where available; always verify against the current guidance from the authority body before relying on a number in a contract, policy, or compliance filing.