Scrutiny of Medicare Advantage risk adjustment coding is at an all-time high
Why a rigorous, programmatic approach to deleting unsupported HCCs is no longer optional.
Sometimes the sound of silence is threatening: it’s the noise the other shoe makes when it’s about to drop.
In recent months, government watchdogs across three separate federal agencies announced sanctions or findings that have rattled risk adjustment program managers across the country. The announcements vary in terms of the specific infractions they allege, but a common thread unites them: Scrutiny of risk adjustment practices among Medicare Advantage plans is growing, and perhaps has never been higher.
At stake is the future of business as usual for risk adjustment teams. While the sanctions may apply only to individual payers, the message they convey is meant for every Medicare Advantage plan: Auditors and law enforcement are increasing their oversight of risk adjustment programs in an effort to bring down costs for patients and taxpayers.
The policy context
Warming to the theme of an “affordability crisis,” policymakers and lawmakers in Congress have taken up the task of runaway healthcare costs.
In March of this year, the Joint Economic Committee published a report blaming rising Medicare Part B premiums on overpayments made to MA plans. Chairman David Schweikert summarized the state of affairs as one where “seniors who stay in traditional Medicare are effectively subsidizing the system.” JEC researchers put a $13.4 billion price tag on the total increase in Medicare Part B premiums in 2025 alone. That comes out to an additional $212 per enrollee—not an insignificant sum for senior citizens on fixed incomes.
The JEC report appears as the HHS Office of Inspector General (OIG) is close to finishing an audit series that has returned abundant evidence for the claim that unsupported diagnosis codes are a principal driver of MA overpayments.
The politics of the issue have shifted in response to the mounting evidence. What was once primarily a concern for compliance teams has now become a highly visible problem for lawmakers and regulators to solve.
Three channels of enforcement
Various federal agencies and departments have taken action in response to MA overpayments stemming from unsupported diagnosis codes.
- The Department of Justice: DOJ announced in March 2026 that Aetna agreed to pay $117.7 million to resolve False Claims Act allegations related to Medicare Advantage coding. The government alleged that for payment years 2018 through 2023, Aetna "knowingly submitted or failed to delete or withdraw inaccurate and untruthful diagnosis codes for morbid obesity" to increase CMS payments. The case was brought in part by a former Aetna risk adjustment coding auditor. Aetna stated it disagreed with the allegations and did not admit liability.
The phrase "failed to delete or withdraw" is significant. It means that the government is examining not just what codes were submitted, but what plan officials knew about insufficiently supported claims, and what they chose to do—or not do—in response to them. - Centers for Medicare and Medicaid: In a separate action, CMS sanctioned Elevance Health in February 2026, suspending new enrollment in its MA plans effective March 31, 2026. The agency found that Elevance spent seven years submitting corrections for unsupported diagnosis codes using encrypted USB flash drives instead of the required electronic systems, and had repeatedly failed to report and return identified overpayments within the 60-day window required by federal regulation. The Elevance case adds a procedural dimension to compliance evaluation: Even a plan that catches its errors can create additional liability by failing to manage the corrections in compliance with regulations.
- The OIG: Auditors at the HHS watchdog are finishing an audit series targeting high-risk diagnosis codes. Over the past four months alone, the agency published findings against three MA plans. These three audits found a total of more than $21 million in estimated overpayments related to unsupported diagnosis codes.
The common thread: A failure to delete
In each case, the infractions committed by the named MA plans involved a failure to identify and remove diagnosis codes that lacked adequate documentation.The Aetna settlement names this explicitly. And the ongoing OIG audits consistently show that high-risk diagnoses are submitted as active conditions when the medical record supports them only as historical conditions.
As the audit findings stack up, they suggest that RAF scores inflated by unsupported codes have become common practice across plans of different sizes, in different regions, and with different coding vendors.
Risk adjustment programs have historically concentrated resources on capture: identifying supported diagnoses, improving provider documentation, closing coding gaps. That work remains important. But the recent enforcements make it clear that deletion requires equal operational investment.
In short, government watchdogs have forcibly shifted MA plans’ strategy from additional diagnosis captures to diagnostic accuracy.
The emphasis on RAF accuracy
Going forward, MA plans that want to avoid punitive audit outcomes will need to adopt a programmatic approach to code deletion across three compliance pillars:
- Clinical accuracy: The active vs. historical HCC distinction is the most consistent abuse finding across OIG audits. Plans need consistently applied standards for when high-risk conditions (e.g. cancers, resolved acute events, conditions in remission) qualify as active for HCC submission purposes.
- Process: Corrections to submissions must be made through CMS’s required channels and completed within required regulatory timeframes. For example, the 60-day repayment window for identified overpayments is a legal obligation, not a suggestion. Plans need workflows that connect code review decisions to compliant submission corrections without unnecessary delay.
- Audit responsiveness. If regulators inquire, plans need to show proof of work for how they identify and review codes for submission and deletion. Documentation of addition and deletion decisions, and the clinical rationales behind them, will be a crucial compliance asset going forward.
How Charta strengthens RAF accuracy
Charta’s risk adjustment engine uses advanced AI technology to deliver higher RAF accuracy at lower costs, along with audit-ready decision and documentation trails.
Charta helps MA achieve RAF accuracy and RADV readiness with four AI-enabled features:
- Universal document legibility: Charta AI can read and analyze any document type contained within encounter notes (including handwritten notes, faxes, and image files) to extract evidence for incremental HCC codes, and to identify codes that lack sufficient clinical documentation support. Document legibility enables Chata to efficiently and intelligently match charts to claims, and to facilitate the location of supporting evidence through AI automation.
- Standardized AI reasoning: Charta AI models are trained on CMS risk adjustment models and further customized to each payer’s organizational policies. After reviewing patient encounter documentation for possible code additions and to identify codes that are unsupported by the evidence, Charta AI applies standardized reasoning to supply a recommendation for HCC code addition or code deletion.
- Layered human review: Trained and certified human coders review each AI recommendation for incremental and unsupported codes alongside citations of the supporting documentation to confirm, reject, or modify the recommendation. Human coders can also flag edge cases for additional human QA by a supervisor.
- Document citations: Because every coding decision contains a trail of human decisions and AI-supplied source documentation, risk adjustment by Charta is always RADV-ready.
Learn more
To learn more about Charta’s risk adjustment solutions, request a demo from a member of our risk adjustment solutions team today.