Medicare Advantage Risk Adjustment Payment Authority Failure Through Unsupported Diagnosis Codes Submitted Without Corresponding Clinical Documentation at Aetna

Medicare Advantage is the private plan alternative to traditional Medicare. CMS pays plan sponsors a risk-adjusted capitated rate per enrolled beneficiary based on documented health status. A beneficiary with documented chronic conditions produces a higher risk score and a higher payment than a healthy beneficiary. The diagnosis code is the instrument through which health status is represented to CMS. A code asserts that a clinician documented that condition during a qualifying encounter. CMS accepts submitted codes as the evidentiary basis for the risk score. The code authorizes the payment. The clinical documentation is the condition that makes that authorization legitimate. The code does not carry the documentation. It represents that the documentation exists.

Medicare Advantage regulations also require the plan sponsor to submit a written attestation to CMS certifying that the risk adjustment data is accurate and truthful. This is a second credential layer: the diagnosis code authorizes individual payments; the attestation certifies the aggregate submission’s integrity. Neither layer requires independent correspondence verification at reliance. Each vouches for the other. Both are submitted by the entity with direct financial interest in the outcome.

On March 11, 2026, DOJ and HHS-OIG announced that Aetna had agreed to pay $117.7 million across two settlement agreements to resolve False Claims Act allegations related to Medicare Advantage risk adjustment. The settlement incorporated allegations from two distinct qui tam complaints and closed a multi-year federal investigation.

The first agreement, $106.2 million, concerns Aetna’s payment year 2015 chart review program. Aetna hired coders to examine medical records from providers and identify additional diagnosis codes supported by documentation. Aetna submitted those additional codes and received higher payments. The government alleged that the same chart reviews also identified previously submitted codes that were not supported by the records — codes Aetna was required to delete and refund to CMS. Aetna did not. The chart review program was used to add supported codes. It was not used to delete unsupported ones. The mechanism operated in a single direction.

The second agreement, $11.5 million, concerns payment years 2018 through 2023. The government alleged that Aetna knowingly submitted or failed to delete morbid obesity diagnosis codes assigned to members whose recorded BMI fell below the clinical threshold for that diagnosis. The second agreement resolved a qui tam complaint filed in January 2024 by Mary Melette Thomas, a former Aetna risk-adjustment coding auditor. The settlement includes the allegation that Aetna falsely certified in writing to CMS that the submitted data was accurate and truthful while its own review processes had identified specific codes as unsupported.

The diagnosis code functions as a payment credential in the Medicare Advantage risk adjustment system. CMS has no mechanism at payment authorization to verify that the documentation the code represents is present in the patient record. The payment follows the code. The failure condition is the absence of a correspondence requirement at the point of reliance. Whether the clinical encounter occurred, whether the condition was documented, whether the documentation meets the standards required to support the code — none of these are verified by CMS at payment. They are represented by the submitting plan sponsor, whose financial interest runs in one direction: a higher-acuity code produces a higher payment.

The chart review program in the 2015 settlement illustrates the exploitation. Aetna built an internal process that examined the same medical records from two directions — additions yielded higher payments; deletions would have triggered refunds. Both directions produced findings simultaneously. Aetna acted on one and not the other. The architecture permitted chart review without requiring symmetric action on its findings. A rational profit-maximizing sponsor inside that architecture would capture the upside and ignore the downside. Aetna structured its operations accordingly.

The morbid obesity specimen illustrates a simpler form of the same failure. Morbid obesity requires a BMI above a defined clinical threshold; BMI is a calculation from height and weight; Aetna possessed the BMI data. Aetna submitted the codes anyway. The sponsor’s own data contradicted the credential the sponsor submitted. The architecture does not evaluate sponsor-held contradictory data at payment. Aetna’s own coding auditors — including the relator — identified unsupported codes as part of the sponsor’s internal integrity function. Those findings were overridden at the submission level. The architecture permits the override and does not require internal integrity findings to be encoded in the credential the relying party evaluates.

Aetna agreed to pay $117.7 million without admitting liability. A CVS Health spokesperson stated that Aetna “continues to disagree with the DOJ’s industry-wide allegations, and this settlement should not be seen as an acknowledgment of liability.” The response parallels Aetna’s 2023 response to the HHS-OIG audit, in which the company’s chief compliance officer disputed the methodology as reflecting “an apparent expectation for perfect coding.” The 2026 settlement covers conduct periods overlapping the 2023 audit findings. The audit was disputed. The settlement followed three years later.

The $117.7 million settlement is the second-largest individual Medicare Advantage upcoding resolution in the catalog, following Kaiser Permanente’s $556 million settlement in January 2026. MedPAC estimated in its March 2025 Report to Congress that systemic Medicare Advantage overpayments for 2025 would total approximately $84 billion above what the same beneficiaries would cost under traditional Medicare — approximately $40 billion attributable to coding intensity (the mechanism the Aetna settlement addresses) and approximately $44 billion to favorable selection. The Joint Economic Committee reported in March 2026 that MA overpayments have produced approximately $82 billion in cumulative Part B premium increases since 2016, borne by all Part B beneficiaries including those in traditional Medicare who did not elect MA; the 2025 single-year premium impact was $13.4 billion. The Medicare Hospital Insurance Trust Fund solvency horizon was revised to 2033. The named-entity settlement addresses individual sponsor conduct within a program-wide architectural failure.

• Aetna settled FCA allegations for $117.7 million across two agreements announced March 11, 2026: $106.2 million for a payment year 2015 chart review program that added supported codes while ignoring unsupported ones, and $11.5 million for 2018–2023 morbid obesity codes submitted for members whose recorded BMI fell below the clinical threshold
• Medicare Advantage risk adjustment operates through two credential layers — the transaction-level diagnosis code and the aggregate-level written sponsor attestation — neither of which requires independent correspondence verification at the point of reliance; both are submitted by the entity with direct financial interest in the outcome
• The chart review mechanism documents the exploitation: the same internal process identified both additions that raised payments and deletions that would have triggered refunds; Aetna acted on the additions and not the deletions; the architecture permitted chart review without requiring symmetric action
• The morbid obesity specimen documents a simpler form: Aetna submitted codes while possessing BMI data that disqualified them; the sponsor’s own data contradicted the credential; the architecture does not evaluate sponsor-held contradictory data at payment
• The case originated from a qui tam by a former Aetna coding auditor whose function was internal integrity verification; her findings were overridden at submission; HHS-OIG’s 2023 audit of Aetna Contract H5521 had already identified similar conduct and extrapolated $25.5 million in overpayments for 2015–2016, which Aetna disputed; the 2026 settlement covers conduct periods overlapping that disputed audit, indicating that regulatory findings did not alter the conduct pattern
• The named-entity settlement sits within a program-wide architectural failure: MedPAC estimated $84 billion in systemic MA overpayments for 2025 ($40 billion from coding intensity); the Joint Economic Committee documented $13.4 billion in single-year Part B premium impact borne by all beneficiaries including those in traditional Medicare; the Medicare Hospital Insurance Trust Fund solvency horizon was revised to 2033; the diagnosis code as payment credential is structurally distinct from provider enrollment and cybersecurity compliance credentials documented elsewhere in the catalog — the correspondence gap is built into the payment architecture