HCC Risk Adjustment · Engine V28

Close chronic-condition recapture gaps the chart can defend

Here is the short answer. Recapture surfaces chronic conditions your documentation already supports but that nobody coded this year. Upcoding invents risk the record cannot back up. The line between them is not volume. It is the note. Suspect analytics that respects that line is the only kind worth running.

What recapture is
A documented chronic condition, re-coded in the calendar year it was assessed.
What it is not
A suspect flag pushed to a claim without MEAT support in the encounter.
The 2026 stakes
CMS-HCC V28 is at 100% weight and RADV extrapolation is live.
The suspect panel, region by region

A geo-grid of suspected-but-undocumented chronic conditions

Read the panel like a map. Each region is a slice of your population. Each tile is a chronic HCC family the analytics flagged as a likely recapture. Green tiles are ready to code today. Amber tiles are real suspects that need a clinician to confirm and document. Red tiles were flagged by an algorithm but the chart does not support them, so they never touch a claim.

Panel suspect map · annual recapture view Illustrative archetype data
Recapture-ready · MEAT present Suspect · needs clinician review No chart support · do not code
North Panel Metro clinics
DMDiabetes w/ complication · confirmed
CKDCKD stage unspecified · confirm stage
CHFHeart failure · MEAT present
COPDSuspect from med list · needs note
-No signal
VASCVascular disease · confirmed
Ready3 confirmed · 2 review
East Panel Rural sites
DMHistorical dx · not addressed this year
CKDLab-suggested · eGFR trend
CHFNo echo, no note · do not code
-No signal
MDDMajor depression · MEAT present
VASCSuspect · confirm laterality
Ready1 confirmed · 3 review
South Panel Multi-specialty
DMDiabetes w/ CKD link · confirmed
CKDStage 4 documented · confirmed
CHFSuspect · confirm systolic vs diastolic
COPDCOPD · MEAT present
AMPUStatus code · confirm from op note
-No signal
Ready3 confirmed · 2 review
West Panel FQHC network
DMSuspect · complication not specified
-No signal
CKDCKD stage 3 · confirmed
CHFAlgorithm only · chart silent
MDDMajor depression · MEAT present
VASCSuspect · confirm active vs history
Ready2 confirmed · 2 review

The archetypes above stand in for real client panels. No client data is shown. The point is the discipline: every amber becomes a gap closed only after a clinician confirms it in the encounter, and every red stays off the claim.

The one distinction that keeps you audit-safe

Closing a gap versus manufacturing risk

Both start from the same suspect list. They diverge the moment you ask what the chart actually says. OIG audits and CMS RADV reviews live in exactly this gap, so the engine has to enforce it, not just suggest it.

Recapture Closes a real gap

The condition is chronic, it exists, and it was addressed this year. You are correcting an omission.

  • The condition is documented in a face-to-face encounter during the current calendar year.
  • The note carries MEAT: it is monitored, evaluated, assessed, or treated.
  • The ICD-10-CM FY2026 code matches the specificity the record supports, including laterality and stage.
  • A provider signs the assessment. The suspect flag informed the review, it did not write the diagnosis.

Upcoding Manufactures risk

The signal came from an algorithm, a stale problem list, or a med the patient no longer takes. The chart is silent.

  • The diagnosis appears only on the claim, never in a current, signed clinical note.
  • It is pulled forward from a prior year with no evidence it was assessed again.
  • Specificity is inflated beyond what the documentation states to reach a higher HCC.
  • It cannot survive a RADV medical-record request or an OIG extrapolated audit.
The gate every suspect has to clear

MEAT is the recapture checkpoint, not a formality

A suspect condition only becomes a codeable diagnosis when the encounter shows at least one element of MEAT. The V28 engine routes every amber tile through this gate before it is ever allowed onto a claim.

M

Monitor

Signs, symptoms, disease progression, or lab trends are being watched over time.

E

Evaluate

Test results, medication response, or exam findings are reviewed and interpreted.

A

Assess

The condition is discussed, ordered, counseled, or recorded in the clinical impression.

T

Treat

Medications, referrals, therapies, or a care plan tie directly to the diagnosis.

Every chronic condition resets on January 1

Risk scores do not carry a diagnosis forward. Under CMS-HCC rules, a chronic condition must be documented and coded again in each calendar year it is assessed. That single fact is why recapture exists, and why the suspect map has to be rebuilt every year rather than copied from the last one.

Jan 1
Clock resets. Prior-year HCCs count for nothing until re-documented.
Q1 to Q2
Suspect analytics surfaces likely recaptures ahead of scheduled visits.
Each visit
Clinician confirms with MEAT. Amber becomes green, or it is dropped.
Dec 31
Open gaps close. Unsupported flags never made it onto a claim.
Why 2026 raises the bar

CMS-HCC V28 makes precise recapture harder and more valuable

The V28 model reached full weight this payment year. It reorganized the condition categories, pruned many codes that used to map to an HCC, and constrained several coefficients. Broad suspect lists that worked under V24 now produce more noise, so the analytics has to be sharper.

100%
V28 weight · PY2026

The three-year V24-to-V28 blend is complete. Payment now runs entirely on the V28 crosswalk, so a code that no longer maps to an HCC quietly stops earning.

Fewer
Payment HCCs

V28 consolidated and removed categories, especially in diabetes and vascular disease. Specificity in ICD-10-CM FY2026 decides whether a condition still lands in a paying HCC.

Live
RADV extrapolation

CMS finalized extrapolated RADV recovery. A pattern of unsupported codes is no longer a single-chart problem, which is exactly why the red tiles never ship.

Guidelines this workflow is built on

Every rule points to a named source

CMS
CMS-HCC Risk Adjustment Model, Version 28

The 2026 payment model, fully phased in at 100% weight, defining which ICD-10-CM codes map to a payment HCC.

CDC / CMS / NCHS
ICD-10-CM FY2026 code set

Effective October 1, 2025. Governs the specificity, laterality, and combination codes a recapture must match.

CMS
Annual chronic-condition recapture rules

The requirement that each chronic HCC be documented and coded again in every calendar year it is assessed.

HHS-OIG
OIG guidance on diagnosis-code integrity

Office of Inspector General audit work and toolkits on high-risk and unsupported diagnosis codes in Medicare Advantage risk adjustment.

CMS
Risk Adjustment Data Validation (RADV)

The finalized RADV rule permitting extrapolated recovery of payments tied to unsupported diagnoses.

Documentation standard
MEAT criteria

Monitor, Evaluate, Assess, Treat. The widely used test for whether an encounter supports coding a chronic condition.

See your panel as a suspect map, with the red tiles already filtered out

The ASP-RCM V28 Risk Adjustment engine builds this geo-grid from your own population, routes every suspect through the MEAT gate, and hands your clinicians a confirm-or-drop worklist. You recapture the gaps the chart supports and leave the rest where it belongs. That is revenue you were already owed, defended for the audit that may follow.

Map my recapture gaps → Talk to a risk-adjustment lead

This page is educational and reflects publicly named 2026 CMS, OIG, and ICD-10-CM guidance current as of publication. Panel figures shown are illustrative archetypes, not client data. It is not coding, billing, or legal advice. Confirm every diagnosis against the signed clinical record before submission.