The Audit Landscape Shifted Permanently
RADV audits in Medicare Advantage underwent a structural transformation between 2025 and 2026. CMS moved from periodic audits targeting select contracts to annual audits of all 550+ MA contracts. Payment year 2020 audits began in February 2026 with quarterly cadence, meaning new audit cycles launch every three months. The agency scaled its audit workforce from approximately 40 certified coders to plans for approximately 2,000, supplemented by AI-assisted pattern detection for audit targeting.
These changes aren’t temporary enforcement spikes. They’re permanent infrastructure investments. CMS built the workforce, the technology, and the operational cadence to sustain universal annual audits indefinitely. Plans that treated RADV as an occasional risk they’d deal with if selected now face it as a continuous operational reality.
What’s Different About the Current Audit Structure
Three technical changes make the current RADV framework more consequential than previous versions. First, variable sample sizes. CMS now selects between 35 and 200 enrollees per contract for medical record review. Larger samples produce more statistically reliable findings, which strengthens CMS’s position in disputes and recoupment calculations.
Second, the quarterly cadence means plans may face overlapping audit cycles. A plan responding to a PY 2020 audit in Q1 2026 could receive a PY 2021 audit notification in Q2. Managing concurrent audits with different enrollee samples, different documentation packages, and different submission deadlines requires operational infrastructure that most plans didn’t have when audits were periodic.
Third, CMS confirmed that AI assists its audit targeting and review process. The agency uses AI as a “medical coder support tool” to identify coding patterns warranting investigation. This means CMS is scanning population-level data for the statistical signatures of problematic coding practices before selecting specific contracts for deeper review.
The OIG Audit Results That Preview RADV Exposure
While RADV and OIG audits are technically separate processes, the OIG’s March 2026 findings provide a preview of what plans should expect when RADV results arrive. Three simultaneous audits found error rates between 81% and 91% across 857 sampled enrollee-years. The most common failures were history-of conditions coded as active, acute diagnoses carried forward without current management documentation, and missing MEAT evidence (Monitoring, Evaluation, Assessment, Treatment).
These patterns exist in most plans’ submitted data. The OIG findings aren’t unique to the three organizations audited. They reflect industry-wide documentation and coding practices that will produce similar results when RADV auditors apply the same standards to other contracts.
What Preparation Actually Requires
Plans serious about radv audits in 2026 need three capabilities operational before the next audit notification arrives: pre-submission defensibility scoring that catches weak codes before they reach CMS, a unified evidence trail system where every submitted code has documented MEAT validation, and an audit response infrastructure that can assemble medical record packages for 35 to 200 enrollees within CMS’s five-month submission window.
Plans that build these capabilities now are managing RADV as the continuous function it has become. Plans that wait are betting they can build them under deadline pressure after the notification arrives.