HCC Risk Adjustment / V28 / RADV 2026

Coding to Survive RADV: the MEAT behind every 2026 HCC

Since the CMS RADV final rule, a single unsupported diagnosis is no longer a single dollar. CMS extrapolates the error rate across your whole contract. The chart that survives is the one where every HCC carries its MEAT.

The short answer: a code is not documentation. To hold up under an extrapolated RADV audit, each 2026 HCC needs a face-to-face note showing the condition was Monitored, Evaluated, Assessed, or Treated in the service year, signed by an eligible provider, and coded to FY2026 ICD-10-CM specificity.
Extrapolated
RADV recovery is projected across the contract, not limited to the sampled charts. One weak note scales.
CMS RADV Final Rule
PY2018+
Extrapolation applies to payment year 2018 and every year after, including 2026 dates of service.
CMS-4185-F2
No FFS Adjuster
CMS finalized removal of the Fee-for-Service Adjuster from RADV recovery calculations.
CMS RADV Final Rule
Oct 1 2025
The ICD-10-CM FY2026 code set is in effect. Laterality and specificity are validated at FY2026 rules.
ICD-10-CM FY2026
The four-letter standard

MEAT is what an auditor reads for, line by line

MEAT is the documentation standard reviewers use to decide whether a diagnosis was actively managed at the encounter, not just copied forward on a problem list. Every HCC in the matrix below is scored against these four.

M
Monitor

Signs, symptoms, labs, or disease progression tracked over the visit.

E
Evaluate

Test results reviewed, response to treatment examined, findings interpreted.

A
Assess

The condition addressed in the assessment, ordering, or clinical decision-making.

T
Treat

Medication, therapy, referral, or a plan tied to the diagnosis by name.

The survival matrix

Common HCC drivers, and the MEAT RADV will demand

Left is what the coded claim asserts. Right is what an extrapolated RADV audit needs to see in the note to let it stand. The flag is the gap we watch fail most often on real chart review.

CMS-HCC V28

Diabetes with chronic complication

E11.22 - Type 2 DM with diabetic CKD
Claim asserts a linked, active diabetic complication driving the risk score.
MEAT the auditor needs
M Monitor

Recent A1c, eGFR trend, or urine albumin noted.

E Evaluate

Provider reviews the labs and links CKD to diabetes.

A Assess

Assessment states diabetic CKD, not just "DM" plus "CKD" separately.

T Treat

Metformin dose, ACE/ARB, or nephrology referral tied to it.

Fails when: the note lists diabetes and CKD as unrelated problems. The causal link that makes it the combination HCC is never stated.
CMS-HCC V28

Chronic heart failure

I50.32 - Chronic diastolic (HFpEF) heart failure
Claim asserts an active, chronic, typed heart failure.
MEAT the auditor needs
M Monitor

Weight, edema, dyspnea, or BNP tracked at the visit.

E Evaluate

Echo or EF result reviewed to support the diastolic type.

A Assess

Documented as chronic diastolic HF, not unspecified "CHF".

T Treat

Diuretic, GDMT, fluid or sodium plan named.

Fails when: only "CHF" is documented. Unspecified I50.9 does not carry the same HCC, and the type cannot be inferred by the coder.
CMS-HCC V28

COPD

J44.1 - COPD with acute exacerbation
Claim asserts a chronic respiratory condition, exacerbation adds specificity.
MEAT the auditor needs
M Monitor

Oxygen saturation, spirometry, or symptom status noted.

E Evaluate

Inhaler technique or exacerbation frequency reviewed.

A Assess

Stable vs exacerbated stated for FY2026 code selection.

T Treat

Bronchodilator, steroid, or oxygen order documented.

Fails when: "history of COPD" is charted with no active management. History-of language reads as resolved, not current.
CMS-HCC V28

Major depressive disorder, recurrent

F33.1 - MDD, recurrent, moderate
Claim asserts an active recurrent mood disorder at defined severity.
MEAT the auditor needs
M Monitor

PHQ-9 or symptom check documented this encounter.

E Evaluate

Response to current medication reviewed.

A Assess

Recurrent and severity stated, not just "depression".

T Treat

Antidepressant, therapy, or referral tied to the dx.

Fails when: coded from the med list alone. An SSRI on the profile is not proof the condition was addressed at the visit.
CMS-HCC V28

Chronic kidney disease, stage 4

N18.4 - CKD stage 4 (severe)
Claim asserts a specific, staged renal condition.
MEAT the auditor needs
M Monitor

eGFR value supporting stage 4 documented.

E Evaluate

Trend reviewed, dialysis planning discussed.

A Assess

Stage stated explicitly, not "CKD" unspecified (N18.9).

T Treat

Nephrology referral, med adjustment, or diet plan.

Fails when: the stage is only in a lab flag, never in the provider's assessment. The eGFR alone does not stage the code for RADV.
CMS-HCC V28

Morbid obesity with BMI

E66.01 + Z68.42 - BMI 45.0 to 49.9
Claim pairs the clinical diagnosis with a documented BMI band.
MEAT the auditor needs
M Monitor

Current BMI recorded at the encounter.

E Evaluate

Weight-related risk or comorbidity reviewed.

A Assess

Provider states "morbid obesity", not just a BMI number.

T Treat

Counseling, plan, or referral by the provider.

Fails when: only the Z68 BMI is charted by staff. Under coding guidelines the clinical E66 diagnosis must come from the provider.
CMS-HCC V28

Atherosclerotic disease with complications

I70.25 - Atherosclerosis of extremities with ulceration
Claim asserts vascular disease with a documented complication.
MEAT the auditor needs
M Monitor

Ulcer site, pulses, or wound status examined.

E Evaluate

ABI or vascular study reviewed.

A Assess

Laterality and ulcer site coded to FY2026 specificity.

T Treat

Wound care, antiplatelet, or vascular referral.

Fails when: laterality is missing. Under ICD-10-CM FY2026, an unspecified-side code where a side is documented is a validation risk.
Why one chart matters

How extrapolation turns a note into a number

The mechanics of the CMS RADV final rule are the reason MEAT discipline is a financial control, not a coding nicety.

Sample

CMS pulls a sample

A subset of enrollees and their HCCs is drawn for the contract year.

Validate

Charts get MEAT-tested

Each HCC must be supported by a compliant record for the date of service.

Error rate

Unsupported = error

Codes without MEAT are removed and an error rate is calculated.

Extrapolate

The rate scales up

Under the final rule, that rate is projected across the full contract population.

Cite it by name

The 2026 guidelines this page is built on

CMS-4185-F2

CMS RADV Final Rule

Establishes extrapolation of RADV audit findings for payment year 2018 and forward, and removes the Fee-for-Service Adjuster from recovery calculations.

Documentation standard

MEAT: Monitor, Evaluate, Assess, Treat

The industry-recognized framework auditors apply to confirm a diagnosis was actively managed at a face-to-face encounter in the service year.

Effective Oct 1 2025

ICD-10-CM FY2026

The code set and Official Guidelines governing specificity, laterality, and combination coding for all 2026 dates of service.

Risk model

CMS-HCC Model V28

The risk-adjustment model whose HCC mappings and constrained categories determine which documented conditions carry a 2026 risk score.

Build MEAT into the coding, not into the appeal

Our HCC and risk-adjustment engine (V28) flags every HCC that lacks its Monitor, Evaluate, Assess, or Treat evidence before the claim goes out, so an extrapolated RADV audit finds a defensible chart instead of a recovery. That is the difference between coding to bill and coding to survive.

Pressure-test your HCC documentation
ASP-RCM Solutions / Risk Adjustment Practice / Frisco