The RAF Audit You're Already Losing Money On—And How To Stop

Two major CMS changes landed simultaneously in 2026 — and most internal medicine practices are only paying attention to one of them.

According to RAAPID Inc., CMS-HCC Model V28 became fully operative on January 1, 2026, eliminating 2,027 valid ICD-10-CM diagnosis codes from risk scoring and projecting a 3.12% reduction in average Medicare Advantage RAF scores nationwide. At the same time, CMS expanded RADV audits from roughly 60 MA plans per year to all 550+ eligible contracts.

The math is unforgiving. According to Invene, every 0.1 increase in RAF score generates approximately $1,040 in annual revenue per member — meaning documentation gaps are not an administrative nuisance. They are a measurable revenue problem compounded by audit exposure on both ends: undercoding loses reimbursement, overcoding triggers clawbacks.

This is the dual pressure ForeSee ESP is built to solve. Learn how Perfect Rhythm can help your practice close HCC gaps and protect RAF revenue.

What You Need to Know About How Internal Medicine Practices Capture Missing RAF Scores with ForeSee ESP

Two regulatory forces landed simultaneously in 2026, and internal medicine practices caught in the middle are losing revenue from both directions at once.

CMS-HCC Model V28 went to 100% implementation on January 1, 2026 — cutting valid ICD-10-CM codes from 9,797 to 7,770 and projecting a 3.12% reduction in average MA risk scores nationally, according to RAAPID Inc. At the exact same moment, CMS expanded RADV audits from roughly 60 Medicare Advantage plans per year to all 550+ eligible contracts, scaling its medical coder workforce from 40 to approximately 2,000 by September 2025, per an official CMS press release.

The math on what you're leaving on the table is straightforward. According to Invene, the 2025 base rate is approximately $10,402 per member annually — meaning every 0.1 increase in RAF score generates roughly $1,040 in additional revenue per patient. Across a 300-patient MA panel, modest RAF improvements compound fast.

The problem most practices haven't solved: up to 80% of clinically actionable patient data lives in unstructured formats — consult PDFs, scanned notes, transcripts — that EHR-native tools cannot process, according to Outsource Strategies International. That's where documented chronic conditions disappear from RAF calculations entirely.

ForeSee ESP addresses both failure modes simultaneously:

• Surfaces missing HCC codes using NLP to mine unstructured records, with InstaVu® linking every suggested diagnosis back to its exact source chart page
• Flags unsupported diagnoses via the Delete Suspects report, removing codes that create RADV clawback exposure before auditors find them
• Automates V28 mapping, eliminating revenue loss during the model transition

As RAAPID Inc. put it directly: "The same year V28 hit 100 percent, RADV audits resumed at scale. The model change and the enforcement change arrived together." Practices that treat HCC coding as an annual clean-up exercise are now operating in a fundamentally different regulatory environment.

Key Strategies and Best Practices

HCC coding gaps rarely happen because a condition went untreated. According to Inferscience, they happen because the clinical evidence wasn't documented with the MEAT criteria — Monitoring, Evaluation, Assessment, Treatment — required to support the diagnosis. That's a process problem, not a clinical one. And process problems have process solutions.

Three disciplines separate practices that capture RAF revenue from those that don't:

• Mine unstructured data before every encounter. Up to 80% of clinically actionable patient data lives in unstructured formats — PDFs, consult notes, audio transcripts — that EHR-native tools can't process, according to Outsource Strategies International. UPMC Health Plan demonstrated exactly this problem when it deployed NLP specifically to surface HCC-eligible evidence buried in unstructured records that existing tools had missed entirely.
• Link every suggested diagnosis to its source documentation. ForeSee ESP's InstaVu® feature does this automatically — clicking any suggested diagnosis takes the coder directly to the exact chart page that supports it. That's not just efficiency. That's RADV defensibility.
• Audit what you've already coded, not just what you're missing. ForeSee ESP's Delete Suspects report flags previously captured diagnoses that lack current clinical evidence. With CMS now auditing all 550+ eligible MA contracts annually, an unsupported code isn't just inaccurate — it's a liability.

According to VBC Risk Analytics, providers who receive actionable, evidence-based suspect lists close HCC gaps at 3–4x the rate of those receiving generic coding reminders. The difference between a reminder and a source-linked, audit-ready suggestion is the difference between a nudge and a workflow.

With 55% of Medicare beneficiaries now enrolled in Medicare Advantage — and the Congressional Budget Office projecting that figure hits 63% by 2034, per Medisys Inc. — this isn't a niche compliance exercise. It's the core revenue engine for internal medicine practices, and it runs on documentation discipline.

Common Mistakes to Avoid

Most internal medicine practices are not failing at HCC coding because they are careless. They are failing because their tools were built for a different problem.

Mistake #1: Treating EHR-native coding tools as a complete solution. According to Outsource Strategies International, up to 80% of all clinical data lives in unstructured formats — notes, PDFs, consult letters — that standard EHR tools cannot process for risk adjustment. A specialist's CKD Stage 3 finding buried in a PDF generates zero RAF credit if your system never surfaces it.

• Coding only for recapture, never for discovery. Most platforms chase last year's codes. ForeSee ESP's Disease Discovery Engine maps 7,903 HCC ICD codes across 115 HCC categories, surfacing new and progressive conditions that recapture-only tools routinely miss, according to an independent compliance review by Impakter.
• Ignoring unsupported codes until an audit arrives. Submitting a diagnosis without current clinical evidence is not just a revenue risk — under expanded RADV enforcement, it is a clawback waiting to happen.
• Delaying the V28 workflow transition. V28 eliminated over 2,200 previously valid diagnosis codes. Practices still running V24 logic are coding on a model CMS no longer recognizes.

The fix is not working harder on coding. It is closing the gap between what is clinically documented and what is actually submitted — in real time, at every encounter.

Practical Next Steps

Three concrete actions move the needle on RAF capture — and none of them require a full workflow overhaul to start.

• Audit your V28 exposure first. V28 removed over 2,200 diagnosis codes from risk scoring entirely, according to HOM RCM. Any practice still relying on V24 logic for diabetic complication codes is bleeding reimbursement silently. Run a gap analysis before the next encounter cycle.
• Switch from annual cleanup to encounter-level documentation. RAF scores reset every January 1st. Every chronic condition not re-documented with MEAT-compliant specificity disappears from your risk score — regardless of how well it was captured last year. This is a process problem, not a coding problem.
• Deploy tools built for unstructured data. According to VBC Risk Analytics, providers who receive actionable, evidence-based suspect lists close HCC gaps at 3–4x the rate of those receiving generic coding reminders. ForeSee ESP's automated V24-to-V28 mapping and NLP-driven gap surfacing deliver exactly that.

The AI coding market is moving fast — Grand View Research valued it at nearly $3.4 billion in 2025, projecting over $10 billion by 2033. Practices that build these workflows now won't be playing catch-up when AI-assisted documentation becomes standard of care.

HCC coding gaps rarely exist because a condition went untreated. According to Inferscience, they exist because documentation failed to capture the clinical evidence — the MEAT criteria — required to support the diagnosis. That is a solvable operational problem, not a clinical one.

HOM RCM reports that AI-assisted documentation programs achieve more than 98% accuracy while identifying missed HCCs that can increase RAF scores by up to 40%. The math on even a fraction of that improvement — across a panel of 300 Medicare Advantage patients — is not incremental. It is practice-changing.

The practices that will absorb the V28 revenue cut and survive expanded RADV audits are not the ones working harder. They are the ones working with better tools — surfacing what is already in the chart, linking every code to its source, and proactively removing what cannot be defended.

• V28 is live and cutting RAF scores by a projected 3.12% — every unaddressed gap compounds that loss
• RADV audits now cover 550+ MA plans annually — unsupported codes are no longer a theoretical risk
• 80% of clinically actionable data lives in unstructured text — EHR-native tools cannot see it

ForeSee ESP addresses all three simultaneously: NLP-powered gap discovery, InstaVu® source linking for audit defense, and the Delete Suspects report for code integrity. Discover how Perfect Rhythm's HCC coding solutions can transform your practice's revenue capture.

---

Frequently Asked Questions

How do internal medicine practices capture missing RAF scores?

Practices capture missing RAF scores by identifying clinically documented but uncoded chronic conditions in patient records — particularly in unstructured notes and consult PDFs — and ensuring every diagnosis meets MEAT criteria before submission. AI-powered tools like ForeSee ESP automate this process at the encounter level rather than as an annual retrospective exercise.

What is an HCC code and why does it affect Medicare Advantage reimbursement?

Hierarchical Condition Categories (HCC codes) are diagnosis codes that CMS uses to calculate a patient's Risk Adjustment Factor (RAF) score, which determines per-member reimbursement under Medicare Advantage. Missing or underdocumented HCC codes directly reduce RAF scores and the revenue a practice receives for managing those patients.

What changed with CMS-HCC Model V28?

V28 became fully operative on January 1, 2026, reducing valid ICD-10-CM diagnosis codes from 9,797 to 7,770 and expanding HCC categories from 86 to 115. CMS projected the transition would reduce average Medicare Advantage RAF scores by approximately 3.12% nationally, creating measurable revenue pressure for practices that have not updated their coding workflows.

What is a RADV audit and how does it affect my practice?

Risk Adjustment Data Validation (RADV) audits are CMS reviews that verify submitted HCC diagnoses are supported by clinical documentation. CMS expanded these audits from approximately 60 Medicare Advantage plans annually to all 550+ eligible contracts — meaning unsupported diagnoses now carry real clawback risk for any practice billing under Medicare Advantage.

How does ForeSee ESP reduce RADV audit exposure?

ForeSee ESP's InstaVu® feature links every suggested diagnosis directly to its source page in the clinical record, providing immediate audit-ready documentation. The Delete Suspects report flags previously captured diagnoses that lack current clinical evidence, allowing practices to proactively remove unsupported codes before they create liability. Get started with Perfect Rhythm to strengthen your audit defenses today.