MedPAC's V28 Coding Intensity Report: What the First Year Actually Showed
In 2026, MedPAC published its first look at Medicare Advantage coding intensity under the V28 risk model, using 2024 risk scores — the first year V28 was partially in effect. It is the most current authoritative read on whether the new model is doing what CMS designed it to do. (Source: MedPAC, "Medicare Advantage coding intensity during the first year of the V28 risk model implementation," June 2026)
The headline numbers
- Under the old V24 model, 2024 MA risk scores ran 18% higher than demographically similar fee-for-service beneficiaries.
- Under V28, that gap fell to 9% — an 8.8-percentage-point reduction in coding intensity.
- 2024 payments were blended two-thirds V24, one-third V28. MedPAC estimates MA plans would have been paid $10 billion more in 2024 had payment been based solely on V24.
The model is working as designed: less payment lift from coding differences, stable plan rebates, and continued enrollment growth.
The effect was not evenly distributed
MedPAC ranked MA organizations by their V24 coding intensity and measured the V28 effect for each tier:
| V24 coding intensity | Share of enrollee-years | V28 effect |
|---|---|---|
| Below 0% | 0.2% | +0.3% |
| 0–5% | 9.1% | −7.9% |
| 5–10% | 4.2% | −5.0% |
| 10–15% | 19.3% | −5.5% |
| 15–20% | 30.8% | −8.5% |
| 20–25% | 20.0% | −11.5% |
| Above 25% | 1.5% | −23.3% |
The plans that coded most aggressively under V24 saw the largest corrections. The outlier tier — organizations with coding intensity above 25% — saw a 23-point swing.
A small number of codes still drive the gap
MedPAC's most operationally useful finding: coding differences are concentrated, not diffuse.
- Diagnosis codes excluded from V28 reduced total MA–FFS coding-rate differences by 30%. Just 14 excluded codes accounted for 23% of total differences — concentrated in vascular disease, mental health, hematological, and immunity disorders.
- Of codes still included in V28, 70% of the remaining differences persist — concentrated in diabetes, congestive heart failure, COPD, and morbid obesity. 29 codes account for 52% of total remaining differences.
Why this matters for Part D
V28 governs the medical (Part C) side, but the lesson transfers directly to RxHCC: CMS is systematically removing the diagnoses that contributed most to MA–FFS coding differences, and the remaining gap concentrates in a short list of chronic conditions. Expect the same directional pressure in Part D risk adjustment — the CY2027 RxHCC model already excludes audio-only diagnoses and re-derives coefficients on newer data.
The plans that win under this regime are not the ones that code more. They're the ones whose coding is documented well enough to survive exactly this kind of analysis.
*Read next: Aligning MedPAC and CMS coding intensity estimates — why the two agencies' numbers differ, and why both are right.*