Risk Adjustment & HCC Coding

The Purpose of Risk Adjustment

Risk adjustment is a payment mechanism that addresses a fundamental challenge in capitated insurance: if plans are paid the same amount regardless of member health, they have a financial incentive to avoid enrolling or retaining sicker members. Risk adjustment prevents this by increasing plan payments for members with chronic conditions and complex medical needs.

In Medicare Advantage, ACA Marketplace plans, and Medicaid managed care programs, capitation means the plan receives a fixed payment per member per month, accepting full financial risk for that member's care. Without adjustment, that fixed payment would inadequately cover costs for sicker populations—yet plans might still profit by attracting only the healthiest enrollees. Risk adjustment tilts payment toward medical need, creating a level playing field.

The CMS-HCC Model: From Diagnosis to Risk Score

The Centers for Medicare & Medicaid Services (CMS) uses the Hierarchical Condition Category (HCC) model to measure member health status and calculate a Risk Adjustment Factor (RAF) score. Here is how it works:

1. Diagnosis collection. Providers submit claims and encounter records containing ICD-10 diagnosis codes for each member during the benefit year (typically calendar year for Medicare Advantage).
2. HCC mapping. CMS maps those diagnosis codes into HCCs—clinically meaningful groupings that predict future healthcare costs. One diagnosis code may map to an HCC; multiple codes may map to the same HCC.
3. Hierarchies applied. The model enforces hierarchies: if a member has both a less severe and more severe form of a condition, only the more severe HCC is counted. For example, diabetes without complications (HCC 19) rolls up into Type 2 diabetes with complications (HCC 18) if both are present; only HCC 18 counts.
4. RAF calculation. CMS assigns a risk weight to each applicable HCC (e.g., congestive heart failure might have a weight of 0.45, meaning it adds 45% to the member's base payment). The RAF is the sum of all applicable weights plus a demographic factor. An RAF of 1.0 represents average risk; 0.8 represents 20% below average; 1.2 represents 20% above average.
5. Payment adjustment. The plan's capitation payment is multiplied by the RAF. A member with RAF 1.2 generates 20% higher payment than a member with RAF 1.0, reflecting higher expected costs.

The HCC model is recalibrated annually by CMS based on claims and encounter data, and the diagnosis codes captured drive the RAF. Accurate, timely capture of diagnoses is therefore central to appropriate payment and plan sustainability.

Documentation and Compliance: MEAT/TAMPER at the Concept Level

To ensure HCC coding is accurate, providers and plans follow documentation standards often abbreviated as MEAT (for clinician documentation) and TAMPER (for auditor and plan verification). While these acronyms are checklists, the underlying principle is straightforward:

• Medical record support. Every coded diagnosis must be clearly documented in the member's medical record—visit notes, test results, imaging reports, medication lists, or assessment summaries—created by the treating provider during or shortly after clinical encounter.
• Specificity. Documentation should support the most specific, accurate diagnosis. Specificity matters because more specific diagnoses often map to more accurate HCCs. For example, "Type 2 diabetes with stage 2 chronic kidney disease" is more specific and may map to a different HCC than "diabetes unspecified."
• Encounter timing. The diagnosis must be documented during the benefit year in question. A diagnosis documented on January 5, 2024, counts for 2024 payment; one documented on January 2, 2025, counts for 2025. Encounter data (claims and face-to-face visits) capture these diagnoses.
• Clinical validity. Coding should reflect the member's actual clinical status. A diagnosis must be actively managed, mentioned in the assessment or plan, or supported by objective findings—not listed speculatively or from past history alone.

Encounter data—both claims (institutional and professional) and supplemental documentation—feed the HCC model. Accurate encounters ensure accurate HCCs and appropriate RAF calculation.

Payment Models and Annual Reconciliation

CMS calculates RAF scores and determines plan payments in advance, based on diagnoses captured in the prior-year claims. For example, diagnoses submitted in 2023 inform 2024 payment. During the benefit year, plans and providers continue documenting; these 2024 diagnoses will inform 2025 payment.

This structure creates several important dynamics:

• Lag and reconciliation. Plans receive prospective payment based on prior-year risk profiles, then submit current-year encounter data knowing it will affect future years' payment. CMS may reconcile prior-year payments if data audits reveal coding errors.
• Incentive alignment and risk. Accurate, complete capture of diagnoses benefits both the plan (through appropriate RAF) and providers (through payment reflecting actual case complexity). Plans and providers have aligned incentives to ensure diagnoses are documented thoroughly and accurately during each encounter.
• Encounter completeness. Not all encounters are captured on claims. Plans may supplement with additional visit documentation to reflect the full clinical picture. CMS distinguishes between claims and supplemental encounters; both feed HCC models for some programs.

RADV Audits and Compliance Risk

Risk Adjustment Data Validation (RADV) audits are periodic compliance reviews conducted by CMS, state agencies, or contracted auditors. In a RADV audit, auditors randomly sample a subset of members' medical records for a given benefit year and verify that the diagnoses submitted for HCC coding are actually supported by the medical record and documented according to clinical standards.

RADV audits can identify several types of findings:

• Unsupported diagnoses: A diagnosis was coded but the medical record shows no clinical evidence of that condition.
• Incorrectly specified diagnoses: The wrong ICD-10 code was used, or a less specific code was used when a more specific code was warranted.
• Timing or documentation defects: The diagnosis was recorded but not during an actual encounter in the benefit year, or documentation is too vague to support the clinical conclusion.
• Fraud indicators: Patterns suggesting intentional or systematic overcoding, such as identical diagnoses appearing across many members with no clinical basis.

If RADV findings indicate systemic error or overcoding, CMS may adjust plan payment retroactively or impose penalties. Plans therefore invest in concurrent coding audits and encounter capture programs to detect and correct errors before external audit, and in provider training to ensure accurate documentation practices. See Policy Changes for current CMS RADV initiatives.

The Ethical Line: Specificity vs. Upcoding

The boundary between accurate coding and improper upcoding can be nuanced. Here is how compliance-conscious organizations navigate it:

• Appropriate: A provider documents "Type 2 diabetes with chronic kidney disease stage 2" based on current labs and clinical assessment; the coder selects the specific HCC that reflects this documented status. The medical record supports the specificity claimed.
• Appropriate: A plan identifies that many encounters document a diagnosis in assessment but not in the coded diagnosis field, and systematically adds missing documented diagnoses to the encounter record, supported by the provider's contemporaneous clinical documentation.
• Questionable: A provider systematically codes diagnoses to the highest-severity version (e.g., always selecting "with complications" variants) without clear clinical documentation supporting the severity level.
• Improper: A coder infers a diagnosis not documented by the provider, or assigns a diagnosis that is documented only in prior history without active evidence it is being managed in the current year.

Regulatory guidance and court cases have clarified that coding must reflect current, actively managed diagnoses supported by the medical record. Specificity and appropriate hierarchies are expected; systematic upgrades without clinical basis are not. Plans and providers operate under a compliance posture that emphasizes accuracy and transparency with auditors, rather than maximization of RAF.

Risk Adjustment in Other Programs

Medicare Advantage uses the HCC model as described above. The CMS-HCC version 28 (and evolving versions) applies.

ACA Marketplace plans also use a risk adjustment mechanism, funded through the ACA's risk-adjustment program. Plans with sicker-than-average enrollees receive transfers from the central program; plans with healthier-than-average enrollees contribute to it. The model and HCC mapping differ slightly from Medicare Advantage but serve the same purpose: leveling the payment field across health risk.

Medicaid managed care risk adjustment varies by state and program. Some states use HCC-based models; others use simpler demographic factors or condition-based categories. State Medicaid agencies define the risk-adjustment methodology in their managed-care contracts.

Commercial plans are not subject to the same CMS risk-adjustment framework, though insurers use their own internal risk models for pricing and population management.

Related Topics

Risk adjustment intersects with several professional functions. See also:

• Health-Plan Operations — How plans organize clinical and financial functions.
• Providers & Billing — How providers submit encounters and claims.
• Star Ratings & Quality Measurement — Related CMS metrics and their data sources.
• CMS RADV Program — Official regulatory framework and audit procedures.

Glossary