Edition 1, 2026 · 42 pages · MA & ACO REACH
The V28 migration playbook for risk-bearing health plans.
CMS-HCC moves to V28 at full weight in payment year 2026. Two thousand three hundred codes are re-weighted. Over two hundred are removed entirely. The average modeled RAF drop sits near 9.3 percent. This is the eight-step playbook to land the year inside the new model.
Edition
1 · 2026
Length
42 pages
Audience
MA · ACO REACH
Model
CMS-HCC V28
Phase-in
PY 2024-26
Effective
100% · 2026
Executive summaryFive things to know about V28 before 2026.
V28 is the largest CMS-HCC change since the model was introduced. The arithmetic of the change is settled. The operational discipline to absorb it is not. This playbook is a working document for revenue-cycle leaders, coding directors, and medical economics teams who own the V28 transition inside a risk-bearing organization.
01
Roughly 2,300 ICD-10 codes carry a different weight in V28 than in V24.
02
More than 200 codes have been removed from HCC mapping entirely.
03
Diabetes is split into multiple categories with materially different weights.
04
Phase-in completes in payment year 2026 at full V28 weight.
05
Documentation specificity is the largest controllable lever in absorbing the drop.
THE FRAME
The playbook is not about avoiding the model. It is about converting documentation discipline into recovered RAF inside it.
V24 BASELINE RAF
1.04
Representative MA panel · 12-month average
MODELED V28 RAF
0.94
Same panel · re-mapped to V28 categories
PMPY AT RISK
$1,200
Per member at $12K benchmark
RECOVERABLE WITH PLAYBOOK
~50%
Of modeled drop, inside 12 months
The phase-in arithmetic, year by year
| PAYMENT YEAR | V24 WEIGHT | V28 WEIGHT | OPERATIONAL SIGNAL |
|---|---|---|---|
| PY 2024 | 67% | 33% | Partial signal · most plans absorbed quietly |
| PY 2025 | 33% | 67% | Larger signal · forecasts start to diverge |
| PY 2026 | 0% | 100% | Full impact lands · no V24 floor remaining |
WHY THIS PAPER EXISTS
"The V28 conversation has been reduced to two numbers. Both are real. Neither tells an operator what to do on Monday morning."
OPERATOR NEED 01
Sequenced operating plan
OPERATOR NEED 02
Condition-level priority list
OPERATOR NEED 03
Documentation rubric tuned to V28
OPERATOR NEED 04
QA framework for new failure modes
OPERATOR NEED 05
Monthly RAF trajectory read
WORKING DOC
Eight steps cover the cadence. The worked example illustrates it. Not a primer. Not a sales artifact. The reader is the operator.
INTENDED READER
CMO, CFO, coding director, medical economics lead, REACH program director
ASSUMED FLUENCY
CMS-HCC at working level · prospective vs retrospective · RADV · coding intensity adj.
WHAT THIS IS NOT
Not a primer · not a sales artifact · not a vendor pitch deck
The landscapeWhere the weight actually moved.
Condition-by-condition heatmap. V24 to V28 weight movement. Highest-prevalence categories. Intensity scales with the size of the move on a representative MA panel.
V24 to V28 · condition weight delta
Source: CMS final rule · representative MA panel
Prev V24
Prev V28
RAF V24
RAF V28
Delta
Diabetes with complication
42%
42%
0.302
0.166
-45%
CHF
18%
18%
0.331
0.388
+17%
CKD stage 4
5%
5%
0.289
0.421
+46%
Major depression
14%
14%
0.395
0.309
-22%
Vascular disease
11%
11%
0.288
0.299
+4%
COPD
15%
15%
0.335
0.319
-5%
Morbid obesity
7%
7%
0.250
0.186
-26%
Atrial fibrillation
9%
9%
0.268
0.279
+4%
PATTERN 01 · THE DIABETES SPLIT
Generic codes collapse to the floor
V24 paid the same weight for diabetes regardless of which specified complication was documented. V28 distinguishes the categories. Without specificity in the chart, the capture collapses to the lowest-weight bucket.
PATTERN 02 · THE CKD ACCELERATION
Staging discipline is now paid
CKD stage 4 and stage 5 carry meaningfully higher weight in V28 than V24. The clinical discipline of staging clearly in the note is rewarded directly. CKD is a category where V28 reinforces what was already best practice.
TAKEAWAY
V28 rewards specificity. Re-anchor documentation around the qualifier the model wants, every encounter, no exceptions.
Diabetes consolidation: the encounter code decides everything
MH & SUD · WHAT WAS REMOVED
Unspecified codes dropped from HCC mapping outright
A note that reads "depression" carries nothing under V28. Several unspecified MDD and substance use codes were removed.
MH & SUD · WHAT THE MODEL WANTS
Severity + episode + remission status in the A&P
"MDD, recurrent, severe, without psychotic features" carries HCC 155. Substance use pivots on remission status.
The two-pathway reconciliation map
MA and ACO REACH pathways are not symmetric. Confusing them is the most common source of forecasting error at reconciliation.
| DIMENSION | MEDICARE ADVANTAGE | ACO REACH |
|---|---|---|
| Risk model | CMS-HCC V28 | Normalized CMS-HCC V28 |
| Coding intensity adj. | 5.9% symmetric across plans | REACH-specific · not symmetric |
| Statutory reference | SSA 1853(c)(1)(C)(ii)(IV) | REACH model rules · CIR ratio |
| Year-over-year drift | Stable adjustment | Can move year over year |
Three operational mistakes the playbook is built to avoid
MISTAKE 01
Treating V28 as a coding problem
Coders cannot code what documentation does not support. The fix is upstream: encounter, problem list, assessment, plan.
MISTAKE 02
Confusing drift with drop
Drift is year-over-year decay when chronic conditions are not recaptured. Drop is the model change. The two compound.
MISTAKE 03
Chasing retrospective campaigns
Chart-retrieval chases lift the current year. Prospective discipline lifts every year. Discipline beats sprints.
Top 10 moversThe conditions that drive the drop.
Top 10 condition categories explain ~78% of the modeled V28 RAF drop on a representative MA panel. Focus on the top three alone recovers roughly half.
Top 10 V28 condition movers · RAF $ impact
Source: CMS · ASP-RCM modeling
Net effect ≈ -$430 pmpy on representative panel · before playbook intervention
DIABETES FAMILY
35%
of modeled drop on typical panel
MAJOR DEPRESSION
20%
of modeled drop
MORBID OBESITY + SUD
23%
combined
FOUR FAMILIES TOTAL
78%
of operational target
CONCENTRATION RULE
Train, chart-review, and QA the four categories first. Discipline scales because it concentrates against dollars, not against the full code list.
Cardiovascular & renal: the other side of the story
CHF · SPECIFIED WINS
"Chronic systolic CHF, NYHA III" beats "CHF"
I50.32 carries · I50.9 loses
CKD · BIGGEST GAIN
Stages 4 and 5 = largest single category gain
Clinical staging now paid
THREE MORE FAMILIES
Arrhythmias · Polyneuropathy · Vascular
Specified = reward, generic = penalty
Generic vs specific: the same chart, two different RAF outcomes
| FAMILY | GENERIC NOTE (V28 PENALTY) | SPECIFIC NOTE (V28 REWARD) | HCC |
|---|---|---|---|
| Diabetes | E11.9 · "diabetes, no complication" | E11.22 · "type 2 DM with diabetic CKD" | HCC 38 |
| Depression | "depression" · F32.9 | "MDD recurrent severe w/o psychosis" | HCC 155 |
| CHF | "CHF" · I50.9 | "chronic systolic CHF, NYHA III" · I50.32 | HCC 226 |
| SUD | "substance abuse" | specified type + remission status | varies |
THE PATTERN
V28 rewards specificity and penalizes generality. The fix is upstream of the coder. Every encounter. Every chronic. Every time.
Worked example5,000-life MA plan. Real arithmetic.
Anonymized MA panel · 5,000 lives · V24 RAF 1.04 · modeled V28 RAF 0.94 (-9.6%) · playbook recovered roughly half in 12 months.
The panel snapshot
PLAN TYPE
Mature MA HMO · regional plan
PAYER MIX
Dual-eligible + chronic-care tilt
CODING TEAM
Experienced · high audit agreement
THE REAL GAP
Upstream · documentation, not coding
AVG AGE
74
DIABETES PREV
41%
CHF PREV
17%
CKD STAGE 3+
12%
MDD PREV
14%
The six-step recommendation, sequenced
STEP 01
Baseline V24 and modeled V28
Same twelve months of claims, two HCC mappings, side-by-side comparison.
STEP 02
Rank conditions by RAF $ impact
Panel-level dollar delta. Top fifteen categories own the priority list.
STEP 03
One-page cheat sheets per category
Specificity language the V28 model rewards, written for the clinician.
STEP 04
Rebuild coding QA rubric
V28 categories. Structured kick-back rule. Defined turnaround clock.
STEP 05
Prospective gap-closure visits
Members whose prior-year HCCs have not refreshed. Address the full problem list.
STEP 06
Real-time V28 EHR dashboard
Capture by category, surfaced in the workflow. No more reconciliation surprise.
PROJECTED RAF
0.98 - 1.00
LANDED RAF
0.99
RECOVERED PREMIUM
~$3M
DIABETES SHARE
~40%
THE FRAME
MDD and obesity kept losing inside V28 even after recapture. Expected. Modeled. The work was discipline, not magic.
Where the RAF lift came from
THE WIN
Diabetes recapture drove ~40% of the lift
CHF and CKD lifts ranked second and third. Total recovered premium: ~$3M at the panel level on a $12K PMPY benchmark.
WHAT DID NOT MOVE
MDD and morbid obesity kept losing weight
Expected. Modeled. The point of the playbook is not to fight the model, it is to harvest where V28 rewards specificity and accept where it does not.
V28 capture · top 8 categories · 5,000 lives
refreshed monthly
Diabetes with peripheral angiopathyHCC 38 · target capture 85%
0.302
0.166
-45%
CHF systolic specifiedHCC 226 · target capture 92%
0.331
0.388
+17%
CKD stage 4HCC 327 · target capture 88%
0.289
0.421
+46%
Major depression recurrent severeHCC 155 · target capture 78%
0.395
0.309
-22%
COPD specifiedHCC 280 · target capture 90%
0.335
0.319
-5%
Vascular disease specifiedHCC 264 · target capture 86%
0.288
0.299
+4%
Atrial fibrillation specifiedHCC 238 · target capture 91%
0.268
0.279
+4%
Morbid obesity with comorbidityHCC 48 · target capture 82%
0.250
0.186
-26%
Aggregate panel RAF · V24 1.04 · V28 0.94 · post-program 0.99
5,000-Life MA PlanAnonymized · representative engagement
WORKED EXAMPLE
RAF and dollars before vs after the playbook.
Panel size
5,000
5,000
V24 RAF
1.04
1.04
Modeled V28 RAF (no action)
0.94
0.94
Post-playbook V28 RAF
0.94
0.99
Implied PMPY at $12K benchmark
$11,280
$11,880
Panel-level recovered premium
$0
$3.0M
Edition 1 · 2026
ASP-RCM HCC AI
V28 is not a coding problem. It is a documentation discipline problem with a coding consequence. The plans that landed inside the new model are the ones that rebuilt the documentation guide around V28 specificity twelve months before the close. The rest are still trying to recover lost ground.
Senior partner · ASP-RCM HCC team
Implementation checklistLand the year inside V28.
Sequenced so the first six items deliver the bulk of recovery. The remaining items lock in discipline and feed the next year's program.
PITFALL 01
Suspect-list governance without a clinical validator
Adds reach submission without MEAT support. Becomes RADV exposure.
PITFALL 02
Retrospective chases crowding out prospective work
Lifts the current year, costs every year after. Imbalance kills the program.
PITFALL 03
Year-end gap visits with a chief-complaint visit type
Misses the chronic problem list. The recapture does not happen.
ESCALATION RULE · WRITE IT DOWN
Any condition with a panel-level RAF dollar impact above $50K that drifts more than two percentage points below target in a given month is escalated to the medical director and the coding lead together, not separately.
01
Run V24 and V28 baseline on the same population.
Twelve months of claims. Map every HCC. Lock the comparison.
02
Rank conditions by panel-level dollar impact.
Focus on top fifteen categories. Sort by RAF $ delta.
03
Rewrite documentation cheat sheets for V28.
One sheet per high-impact category. Specificity language anchored.
04
Run a V28-anchored coding audit on a random sample.
Set the kick-back rules to V28 categories. Track agreement.
05
Schedule prospective gap closure visits.
Panel members with prior-year HCCs that have not refreshed.
06
Update HCC mapping inside the EHR and the dashboard.
Real-time V28 view of capture rate by category.
07
Stand up monthly RAF trajectory tracking.
Rolling twelve-month V28 RAF compared to plan target.
08
Reconcile to the CMS final RAF at year end.
Explain the variance. Feed back into the next program cycle.
Capability stackThe HCC Coding AI under the playbook.
THE PLAYBOOK
The operating cadence
Eight steps. Twelve months. One model. Discipline beats sprints.
THE HCC AI
The assist, not the autonomous coder
Deterministic V28 list + doc citations + specificity-gap flag. Coder still owns the code.
MEAT framework, anchored at the encounter
M
MONITORED
E
EVALUATED
A
ASSESSED
T
TREATED
RADV RULE
Problem-list diagnosis without note-body MEAT = top RADV downgrade driver. Defense begins at the encounter, not at chart retrieval.
RADV defensibility, indexed at submission
REGULATORY ANCHORS
Retrieval window: ~25 weeks per RADV cycle
Best-one rule: 42 CFR 422.310(e)
Attestation: credentialed provider signature
DOS alignment: diagnosis date must match
Chart selection runs in the same pipeline as prospective gap closure. Records indexed at submission. Not reconstructed under audit pressure two years later.
ACO REACH runs on a different math
| MA | REACH |
|---|---|
| 5.9% symmetric coding-intensity adj. | CIR benchmarked vs peer cohort |
| Stable across plans | Asymmetric · moves year over year |
Confusing MA and REACH math underestimates the variance the REACH participant sees at reconciliation.
Layer 04 · AI
HCC Coding AI · V28 condition extractor · specificity gap flag
▼
Layer 03 · LLM Gateway
Single audited choke point · prompt registry · cost meter
▼
Layer 02 · Platform
HCC AI Dashboard · V28 capture by category · trajectory tracker
▼
Layer 01 · HIPAA-eligible AWS
AES-256-GCM PHI · row-level RBAC · PHI access log
GlossaryThe vocabulary of V28.
CMS-HCC
The risk adjustment model CMS uses to pay Medicare Advantage and ACO REACH.
V24
The prior CMS-HCC model. Phased out across PY 2024 to 2026.
V28
The current CMS-HCC model. Full weight in PY 2026.
RAF
Risk Adjustment Factor. The per-member multiplier applied to the benchmark.
HCC
Hierarchical Condition Category. The grouped condition bucket that carries a weight.
Phase-in
The blended weighting of V24 and V28 across PY 2024 to 2026.
Common questionsFrequently asked: V28 migration.
What is V28 and how does it differ from V24?
V28 is the CMS-HCC risk adjustment model phased in from payment year 2024 through 2026. It re-weights roughly 2,300 ICD-10 codes, removes more than 200 codes from HCC mapping entirely, and consolidates several condition categories. The headline effect for most Medicare Advantage plans is a RAF score that is 8 to 10 percent lower than what V24 would have produced on the same population.
Who is affected by the V28 transition?
Every Medicare Advantage organization, every ACO REACH participant, and every risk-bearing provider group paid on CMS-HCC. The Part D model is separate. Commercial HHS-HCC follows a different schedule. If your contract references CMS-HCC RAF, you are inside V28.
Why do diabetes complications matter so much in V28?
V24 paid the same for diabetes with any specified complication. V28 splits diabetes into multiple condition categories with materially different weights. Without specificity in documentation and coding, the RAF impact for diabetic populations is one of the largest single drivers of the V28 drop. Capturing the specific complication is a documentation discipline before it is a coding discipline.
Is the V28 transition phased or a cliff?
Phased. Payment year 2024 blended 67 percent V24 and 33 percent V28. Payment year 2025 blends 33 percent V24 and 67 percent V28. Payment year 2026 runs at 100 percent V28. The full impact lands in the 2026 payment year for most populations.
What is the 8-step migration timeline?
Baseline the V24 RAF and the V28 RAF on the same population. Identify movers, removers, and stable conditions. Re-anchor documentation training around V28 specificity. Stand up a coding QA rubric tuned to V28. Run prospective gap closure on the V28 condition list. Rebuild claim-side rules for V28 mappings. Track RAF trajectory monthly. Reconcile to CMS final RAF annually.
How does the HCC Coding AI help?
The Coding AI runs against the chart and produces a deterministic V28 condition list with rationale, supporting documentation citations, and risk flags. It is not autonomous coding. It is a coder-augmentation tool that compresses the chart review cycle and surfaces the V28-specific specificity questions every condition needs.
What does the worked example show?
A 5,000-life MA plan with a V24 RAF of 1.04 saw a modeled V28 RAF of 0.94, a 9.6 percent drop. After the 8-step playbook ran, the post-program V28 RAF landed at 0.99. On a $12,000 per-member-per-year benchmark, that is roughly $3M in recovered premium across the panel.
Does ASP-RCM replace our coders?
No. The playbook augments the in-house coding team with V28-specific QA rubrics, condition lists, and chart-review acceleration. The coder still owns the code. The playbook owns the framework and the AI assist.
Want this playbook applied to your panel?
Send twelve months of de-identified claims. We run V24 and V28 on the same population, deliver a written condition-by-condition delta, and a recovery estimate in dollars. Yours to keep.