By Jay Chowdappa, MD, co-founder and CEO, Zynix AI · Updated October 27, 2026
Key takeaways
CMS phased the v28 CMS-HCC model into Medicare Advantage risk scores at 33% in 2024 and 67% in 2025, has used it alone since 2026, and keeps it for 2027. For providers and ACOs, HCC gap closure now means a clinician assessing and documenting each chronic condition, with record support, at a visit during the year, not year-end coding catch-up.
- CMS phased v28 in over three years (33% in 2024, 67% in 2025), has used it for 100% of MA risk scores since 2026 and will keep the same model for 2027.
- v28 increased payment HCCs from 86 to 115, but some diagnosis codes that counted under v24 now map to non-payment HCCs.
- For 2027, diagnoses from audio-only encounters and unlinked chart reviews no longer count toward MA risk scores, with an exception for members who switch plans.
- CMS now plans to audit every eligible MA contract each payment year, so every coded condition must be supported in the medical record.
Understanding the CMS Transition
The Centers for Medicare and Medicaid Services (CMS) has implemented significant changes to the HCC risk adjustment model, transitioning from v24 to v28. This phased transition, completed in 2026, has reshaped how diagnoses are mapped to HCC codes, how RAF scores are calculated, and ultimately how Medicare Advantage and ACO organizations are reimbursed.
Key Changes in the v28 Model
The v28 model introduces several major changes. HCCs were reclassified on ICD-10 data, which increased the number of payment HCCs from 86 to 115, while some diagnosis codes that counted under v24, such as certain "other" and "unspecified" vascular codes, now map to non-payment HCCs. New categories were added to reflect evolving clinical understanding. Coefficient values were recalibrated, meaning that the same diagnosis may generate different reimbursement under the new model.
Impact on Provider Organizations
For ACOs and Medicare Advantage plans, these changes have immediate financial implications. Organizations that have optimized their coding practices around the v24 model may see shifts in their RAF scores — some positive, some negative. The transition demands a comprehensive reassessment of risk adjustment strategies.
AI as the Adaptation Engine
AI-powered risk adjustment tools are well positioned to help organizations navigate this transition. Machine learning models can be rapidly updated to reflect new HCC mappings, identify documentation opportunities under the v28 framework, and quantify the financial impact of coding changes across the entire patient population.
Strategic Recommendations
Organizations should begin preparing now by conducting gap analyses between v24 and v28 coding, training clinicians on new documentation requirements, and implementing AI tools that surface suspected conditions for clinicians to evaluate and document under the evolving framework. Those who adapt proactively will maintain or improve their financial position through the transition.
Where the v28 transition stands in 2026
The transition is complete. For 2026, CMS calculates Medicare Advantage risk scores entirely with the 2024 CMS-HCC model (v28), and it will keep using that model for 2027.
CMS phased the model in over three years. In 2024, risk scores blended 67% from the 2020 CMS-HCC model (v24) with 33% from v28; in 2025 the blend was 33% v24 and 67% v28; and since 2026 CMS has used v28 alone. The model was calibrated with 2018 diagnoses and 2019 expenditures from Original Medicare, which is how it works in practice: conditions documented in one year predict costs, and set risk scores, for the next.
| Payment year | Share from the 2020 model (v24) | Share from the 2024 model (v28) | Other change |
|---|---|---|---|
| 2024 | 67% | 33% | Phase-in begins |
| 2025 | 33% | 67% | Phase-in continues |
| 2026 | None | 100% | Phase-in complete |
| 2027 | None | 100% | Diagnoses from audio-only encounters and unlinked chart reviews excluded, with an exception for members who switch MA organizations |
For 2027, CMS also changed which diagnoses count. It finalized the exclusion of diagnoses from audio-only encounters and from unlinked chart review records, meaning chart reviews not tied to a specific encounter, from Medicare Advantage risk score calculations, with an exception for members who switch from one MA organization to another. For plans and the providers they work with, that puts more weight on documenting conditions during the visit itself.
What it means for documentation and audits
Under v28, accurate risk scores depend on specific, supported documentation more than on catching up at year-end. Each condition that drives a payment HCC should be assessed and documented by the treating clinician during a visit, in terms a reviewer can trace to the medical record.
CMS is also checking more of that documentation. In May 2025 it announced that it would audit all eligible Medicare Advantage contracts for each payment year in newly initiated risk adjustment data validation (RADV) audits, which verify that diagnoses used for payment are supported by medical records, and that it would grow its medical coding team from 40 coders to approximately 2,000.
Technology helps when it supports that discipline: ranking patients with likely undocumented conditions, giving clinicians the open gaps before a scheduled visit, and inviting Medicare Advantage members with undocumented conditions to a visit. The clinician decides what to diagnose and document; software only surfaces the evidence.
HCC gap closure for ACOs under v28
HCC gap closure is the work of making sure every chronic condition a patient actually has is assessed and documented by a treating clinician during the year, so the risk score reflects the patient’s real burden of illness. Under v28, that work starts with the patient list and ends with a supported note, not a coding sweep in December.
For an ACO or a practice with Medicare Advantage contracts, the workflow usually has five steps:
- Find likely gaps. Compare claims history, lab results, medications and prior problem lists with this year’s documentation to list conditions that have not yet been assessed, using v28 mappings so the list reflects what still counts for payment.
- Rank the patients. Put first those with several likely gaps, rising risk or no visit on the calendar.
- Get them seen. Invite patients with no visit booked, preferably to an annual wellness visit or a chronic care visit that is not audio-only, since audio-only diagnoses no longer count for 2027.
- Prepare the clinician. Give the treating clinician a short pre-visit summary of the evidence for each suspected condition. The clinician evaluates the patient and decides what to diagnose and document.
- Check before submission. Have coders confirm that each submitted diagnosis is supported in the medical record, so it holds up in a RADV audit.
Software helps most with the first three steps: a population intelligence layer that joins claims, EHR and lab data can build and rank the suspect list every week, and outreach can fill the visit calendar. The diagnosis itself stays with the clinician.
Frequently asked questions
What is HCC gap closure for ACOs?
HCC gap closure is the process of finding chronic conditions a patient likely has that have not been assessed and documented this year, getting the patient seen, and giving the treating clinician the evidence before the visit. The clinician decides what to diagnose and document, and coders check that each diagnosis is supported. Under v28 the work must happen at real encounters during the year, because unlinked chart reviews and audio-only diagnoses no longer count for 2027.
What is the difference between the v24 and v28 HCC models?
v24 is the 2020 CMS-HCC model and v28 is the 2024 CMS-HCC model. CMS rebuilt v28 on ICD-10 data and more recent costs, which increased the number of payment HCCs from 86 to 115, moved some diagnosis codes to non-payment HCCs and recalibrated coefficients. Because categories and weights changed together, the same diagnosis can carry a different value under v28, and CMS cautions that comparing coefficients across the two models is not appropriate.
Is the v28 transition finished?
Yes. CMS phased the 2024 CMS-HCC model (v28) in at 33% in 2024 and 67% in 2025, and since 2026 it has calculated Medicare Advantage risk scores using only v28. For 2027, CMS keeps the same model. What changes in 2027 is which diagnoses count: those from audio-only encounters and from unlinked chart reviews are excluded, with an exception for members who switch from one MA organization to another.
How should provider organizations adjust their risk adjustment strategy for v28?
Re-run your gap analysis on v28 mappings, because some conditions documented for v24 no longer affect payment and some new categories do. Train clinicians on the specificity v28 rewards, assess chronic conditions at a visit each year, and move the work earlier with pre-visit gap summaries and outreach to patients with no visit booked. Keep every coded condition traceable to the medical record.
Can AI decide which HCC codes to submit?
No. AI can rank patients with likely undocumented conditions, pull supporting evidence from claims, labs and prior notes, and prepare a summary before the visit. The treating clinician then evaluates the patient and decides what to diagnose and document, and coders review what is submitted. That division matters more as CMS expands RADV audits, which check that each diagnosis used for payment is supported by the medical record.
Related reading
- Transforming HCC Risk Adjustment: From Q4 Scramble to Year-Round Success
- Navina Alternatives for ACOs and Value-Based Care Groups in 2026
- Pre-visit planning for value-based care: what to prepare
Source notes
CMS: Announcement of Calendar Year 2024 Medicare Advantage Capitation Rates and Part C and Part D Payment Policies. https://www.cms.gov/files/document/2024-announcement-pdf.pdf
CMS: 2026 Medicare Advantage and Part D Rate Announcement (fact sheet). https://www.cms.gov/newsroom/fact-sheets/2026-medicare-advantage-part-d-rate-announcement
CMS: 2027 Medicare Advantage and Part D Rate Announcement (fact sheet). https://www.cms.gov/newsroom/fact-sheets/2027-medicare-advantage-part-d-rate-announcement
CMS: CMS Rolls Out Aggressive Strategy to Enhance and Accelerate Medicare Advantage Audits (May 21, 2025). https://www.cms.gov/newsroom/press-releases/cms-rolls-out-aggressive-strategy-enhance-and-accelerate-medicare-advantage-audits
CMS: 2025 Medicare Advantage and Part D Rate Announcement (fact sheet). https://www.cms.gov/newsroom/fact-sheets/2025-medicare-advantage-and-part-d-rate-announcement
About the author
Jay Chowdappa, MD is co-founder and CEO of Zynix AI. He is a physician, and ACOs led by him generated $300M+ in shared savings before he started Zynix AI to help care teams follow through on what their data already shows.