TL;DR

The CMS-HCC V28 model finishes its three-year phase-in in payment year 2026. For most Medicare Advantage plans and risk-bearing groups, the all-in effect is roughly a 3% to 4% RAF compression if nothing changes. This guide is a 90-day playbook for closing that gap: 30 days of audit, 30 days of operational changes, and 30 days of audit defense and monitoring.

What V28 actually changed

V28, officially the 2024 CMS-HCC Model, restructures how diagnoses translate into Hierarchical Condition Categories. Three structural changes matter most:

1. More HCC categories, but fewer ICD codes per HCC. V28 expanded the set from 86 categories under V24 to 115 categories. The ICD-10-to-HCC crosswalk shrank from roughly 9,700 codes mapped under V24 to about 7,770 under V28. Codes that previously triggered an HCC may no longer do so.
2. Coefficient changes. Many categories saw their RAF coefficients reduced. Diabetes without complications, vascular disease, and major depression took some of the largest cuts. Several renal categories were elevated or split into more granular tiers.
3. Constrained groups. V28 introduces or tightens "constrained" groups, where multiple related conditions share a single coefficient. The clinical effect: documenting more conditions in a constrained group does not increase RAF; only the highest-weighted member counts toward payment.

The phase-in blends V24 and V28 across three payment years. By calendar year 2026, the model is fully V28. Plans relying on V24-era capture playbooks will see RAF erode quietly as the blend shifts.

The financial impact, in plain terms

The CMS Office of the Actuary's 2024 Advance Notice estimated an aggregate average decrease of roughly 3.12% in MA risk scores from V28, before behavioral effects. Plan-level analysis from major actuarial firms places the realized range between 2% and 6% depending on member mix and prior coding intensity.

~3.12%

Average MA risk score reduction projected by CMS OACT under V28

$18M

Annual revenue exposure on a 50,000-life MA plan from a 3% RAF compression at $12K benchmark

15-25%

Share of previously-captured HCCs at risk under V28 in typical MA panels

Higher-acuity populations and dual-SNP plans tend to be hit harder, because diabetes, vascular disease, and depression appear at higher frequencies in those panels. The effect is also more pronounced for plans that had aggressive coding intensity adjustments under V24.

The 90-day playbook

Days 1 to 30: Audit

The goal of the first 30 days is to know exactly where you stand under V28, by HCC category and by provider.

• Re-run the last 12 months of submitted claims under both V24 and V28 mappings. The delta is your V28 exposure. Most teams find that 15% to 25% of previously-captured HCCs are at risk.
• Identify non-mapping diagnoses. Pull every ICD-10 that triggered an HCC under V24 but does not under V28. Group by frequency. The top 50 codes typically explain 80% of the loss.
• Map the gap to provider, member, and payer. Risk teams should know, by member: which conditions need a 2026 face-to-face encounter to recapture, which have already been captured in 2025 carry-forward, and which are no longer codable at all.
• Stress-test your documentation. For each at-risk HCC, sample 20 charts and ask: would this clinical note support the diagnosis if pulled in a RADV audit? Documentation that worked under V24's coefficient logic may not survive V28's tighter mappings.

Days 31 to 60: Operationalize

The audit produces a worklist. The next 30 days are about turning that worklist into a daily provider-facing workflow, before the submission window for plan year 2026 closes.

• Pre-claim flagging at the point of care. Every encounter for a member with an at-risk HCC should surface, in the EHR or coder workspace, the specific condition that needs documentation and the clinical criteria required. Generic prompts ("review chronic conditions") do not move the needle. Specific prompts ("Patient has documented CKD Stage 3a in 2024. Confirm staging today, or document Stage 3 unspecified.") do.
• Targeted provider re-education. Pick the top 10 V28 changes affecting your panel. Build a 5-minute briefing for each. Deliver in 1:1 huddles, not all-hands. Diabetes coders need different context from cardiology coders.
• Coordinated outreach for unconfirmed conditions. A member with three at-risk HCCs should get one well-prepared encounter, not three separate calls from three teams. The shared canonical record matters most here.
• Submission-window awareness. V28-eligible diagnoses must be documented and submitted within the CMS sweep windows. Build a "days remaining" view per member, and prioritize closures by days-to-window-close, not by RAF size alone.

Days 61 to 90: Defend and monitor

Capture without documentation is exposure. The final 30 days build the audit posture.

• Chain-of-custody per HCC. Every captured HCC should have an attached, retrievable evidence packet: the source clinical note, the supporting labs or imaging, the provider attestation if relevant, and the coder review record. Treat this as a system requirement, not a manual workflow.
• Trumping logic checks. V28's constrained groups create new opportunities for redundant capture. Audit your coder workspace to make sure trumping is applied correctly: if HCC X trumps HCC Y, only X should hit the claim.
• Real-time submission reconciliation. Track every submitted diagnosis through the 277CA, 002, MAO-004, MMR, and MOR cycles. Rejections from MAO-004 are an early signal that documentation will not survive RADV.
• Monthly RAF realization tracking. Compare submitted RAF to recognized RAF month-over-month. A widening gap is the single best leading indicator of a future RADV finding.

Questions to ask your tooling

Most plans run on a stack of point vendors that each see only their slice of the problem. As you operationalize V28, the right tooling questions are:

• Does the system show me, per member, which HCCs are at risk under V28 and why?
• Does it apply V28 trumping logic in real time, or only post-claim?
• Can I see the full submission lifecycle from 277CA through MOR per diagnosis in one place?
• Does it surface documentation gaps before the encounter, or only flag rejections after?
• Can I run a parallel V24 versus V28 simulation on historical claims without an analyst ticket?

If the answer to most of these is "no," the gap will widen as 2026 progresses.

"Plans that win the V28 transition treat capture as an operating-cadence problem, not a coding problem. The right diagnoses, in the right charts, at the right time, with the right documentation, repeated every encounter."

Sources and further reading

• CMS Office of the Actuary: Annual Advance Notice for Medicare Advantage Capitation Rates (risk score impact projections)
• CMS-HCC Risk Adjustment Model: Software and technical documentation (V24 and V28)
• CMS Final Rule for CY 2024 Medicare Advantage Capitation Rates and Part C and D Payment Policies (Federal Register, CMS-4201-F)

Frequently asked questions

When does CMS V28 take full effect?

The CMS-HCC V28 model is fully effective for payment year 2026. Payment years 2024 and 2025 used a blended model: roughly one-third V28 in 2024 and two-thirds V28 in 2025.

What is the average RAF impact of V28?

CMS's Office of the Actuary projected an average risk score reduction of roughly 3.12% in the 2024 Advance Notice. Real-world plan-level impact ranges from 2% to 6% depending on member mix and prior coding intensity.

Which HCC categories were cut most under V28?

Categories with the largest coefficient reductions or ICD-10 unmappings include diabetes without complications, vascular disease, and several mental health categories. Renal categories were generally elevated or split into more granular tiers.

Does V28 affect ACOs the same way as Medicare Advantage plans?

ACOs in MSSP and REACH use risk adjustment models that differ from Medicare Advantage but draw on similar HCC logic. V28 changes meaningfully affect ACOs subject to risk score normalization or financial benchmarking based on HCCs.

What is the single highest-leverage change a risk team can make for V28?

Pre-claim flagging at the point of care, tied to V28-aware trumping logic and submission-window awareness. Most other changes such as provider re-education and outreach are amplified or wasted by whether this works.

Run a V28 simulation on your own population.

Pelica's Risk Adjustment Copilot applies V28 trumping logic in real time, surfaces at-risk HCCs at the point of care, and tracks every diagnosis through the full submission lifecycle. Across Pelica deployments, customers have lifted RAF by roughly +0.4 in two quarters with no new headcount.