On March 11, a national health insurer agreed to hand back $117.7 million to Medicare — the newest data point in a federal enforcement campaign that keeps setting records even as the underlying problem keeps getting bigger.
According to the Justice Department’s announcement, Aetna Inc. will pay $117,700,000 to resolve allegations that it submitted — or failed to withdraw — “inaccurate and untruthful” diagnosis codes for its Medicare Advantage enrollees to inflate the payments it received from the government. The case turns on a mechanic that has become central to how the privatized side of Medicare is financed: the insurer ran a “chart review” program that combed patient records for additional diagnoses, the government contends, then submitted the codes that raised its payments while declining to delete codes its own reviewers could not substantiate. Prosecutors also alleged that, from 2018 through 2023, Aetna billed for “morbid obesity” in patients whose recorded body-mass index did not support the diagnosis. Aetna did not admit wrongdoing; the government’s release notes the claims “are allegations only and there has been no determination of liability.”
Read in isolation, the settlement describes a vigorous accountability apparatus. Read against the arithmetic of the program it polices, it describes something closer to a rounding error. Federal auditors estimate that one narrow slice of the same billing practice drains tens of billions of dollars from taxpayers every year. The government’s single largest attempt to claw those dollars back — a decade-long case against the industry’s biggest player — collapsed last year when a court-appointed referee found prosecutors had not proved it. This is the paradox at the center of Medicare Advantage in 2026: the recoveries are the largest on record, and they are recovering a fraction of what the record says is being lost.
How the diagnosis machine works
Medicare Advantage — Part C — lets seniors trade traditional Medicare for a private plan. The government does not reimburse those plans for services rendered, as it does under fee-for-service Medicare; instead, the Centers for Medicare & Medicaid Services (CMS) pays each plan a fixed monthly amount per enrollee, “risk-adjusted” so that sicker members command higher payments. As Assistant Attorney General Brett A. Shumate noted in the Aetna announcement, “the government pays private insurers over $530 billion each year” to care for Medicare Advantage members. The diagnoses that set those payments come from the insurers themselves.
That design creates a documented, one-directional incentive. Every additional diagnosis code an insurer reports can raise its payment; every code it removes lowers it. Federal watchdogs have flagged two tools as especially vulnerable to that pull. The first is the chart review, in which plans hire coders or analytics vendors to re-examine medical records and surface diagnoses a treating physician never billed. The second is the health risk assessment (HRA) — often an in-home visit arranged by the plan or a contractor rather than the patient’s own doctor. A December 2019 review by the HHS Office of Inspector General found the asymmetry in stark terms: “Over 99 percent of chart reviews in our review added diagnoses,” and diagnoses that appeared only on chart reviews — with no accompanying record of a visit, test, or treatment — drove an estimated $6.7 billion in payments for 2017 alone. The Aetna allegations describe that asymmetry in miniature: use the review to find codes worth adding, ignore the same review when it flags codes that should come off.
A record run of settlements
The Aetna deal did not arrive in a vacuum. It is the latest in an accelerating sequence of Medicare Advantage risk-adjustment recoveries. In 2023, Cigna paid $172 million to resolve similar allegations. In December 2024, the Justice Department announced that New York’s Independent Health would pay up to $98 million, and its former coding vendor’s CEO would contribute, to settle claims of invalid diagnosis codes. In March 2025, Seoul Medical Group and related parties agreed to pay more than $62 million over unsupported codes. This June, the in-home assessment vendor Matrix Medical Network and HealthFair settled for $56.5 million.
The high-water mark came in January, when five Kaiser Permanente affiliates agreed to pay $556 million — the largest Medicare Advantage risk-adjustment settlement to date. The government alleged that, from 2009 to 2018, Kaiser mined patients’ histories to identify diagnoses that had not been submitted, then sent “queries” urging physicians to add those conditions through record “addenda” — sometimes months, and in some cases more than a year, after the visit. All told, prosecutors said, the practice added roughly half a million diagnoses and generated payments “in the range of $1 billion.” Kaiser, which did not admit liability, said in a public statement that it settled “to avoid the delay, uncertainty, and cost of prolonged litigation.”
The recoveries lean heavily on insiders. Nearly every one of these cases began as a whistleblower suit under the False Claims Act, which lets private relators sue on the government’s behalf and keep a share of any recovery. The Kaiser relators — two former employees — are set to divide about $95 million. The Aetna whistleblower, identified in court records as a former risk-adjustment coding auditor at the company, is to receive just over $2 million. The pattern is telling: the government’s enforcement pipeline runs substantially on the testimony of the coders and auditors who watched the machine operate from the inside.
The number beneath the settlements
Stack those settlements against the government’s own estimates of the leakage, and the scale gap becomes the story. In an evaluation released in October 2024 (report OEI-03-23-00380), HHS-OIG found that diagnoses reported only on health risk assessments and HRA-linked chart reviews — and on no other record of care — generated an estimated $7.5 billion in risk-adjusted payments for 2023, tied to 1.7 million enrollees who had no follow-up visit, test, or treatment for the conditions billed. In-home assessments accounted for nearly two-thirds of that sum. And the payments were concentrated: just 20 Medicare Advantage companies drove 80 percent of the $7.5 billion. CMS, notably, concurred with only one of the inspector general’s three recommendations to tighten the practice.
That $7.5 billion is itself a subset of a far larger figure. In its March 2025 report to Congress, the nonpartisan Medicare Payment Advisory Commission (MedPAC) estimated that Medicare will spend about 20 percent more on Advantage enrollees than it would have spent covering the same people under traditional Medicare — roughly $84 billion in 2025. MedPAC attributed about $44 billion of that to “favorable selection” (healthier-than-average people enrolling) and roughly $40 billion to coding intensity: the tendency of plans to document diagnoses more aggressively than fee-for-service providers do. Even after CMS applies a mandatory across-the-board coding adjustment, MedPAC found, Advantage risk scores still run about 10 percent higher than those of comparable traditional-Medicare beneficiaries, and the intensity varies enormously by insurer — a 26-percentage-point spread separated the highest and lowest coders among the ten largest plans. The Committee for a Responsible Federal Budget projects that, left unchecked, MA overpayments will exceed $1.2 trillion over the coming decade.
Set the enforcement numbers beside the estimates and the disproportion is hard to miss. The record-shattering $556 million Kaiser settlement amounts to less than 1.5 percent of the roughly $40 billion that MedPAC pegs to coding intensity in a single year. The entire multi-year run of risk-adjustment settlements — Cigna, Independent Health, Seoul Medical, Kaiser, Aetna, Matrix — totals under $1 billion. The recoveries are real. They are also, by the government’s own accounting, a fraction of one year’s estimated overpayment.
Why the biggest case fell apart
The gap is not merely a matter of prosecutorial appetite. It reflects how genuinely hard these cases are to win at scale — a lesson the Justice Department learned in its most ambitious attempt. For more than a decade, the government pursued UnitedHealth Group, the nation’s largest Advantage insurer, in a case filed by whistleblower Benjamin Poehling in 2011 and joined by the DOJ in 2017. Prosecutors alleged Medicare paid the company more than $7.2 billion from 2009 through 2016 on the strength of chart-review codes, and that it would have paid $2.1 billion less had unsupported codes been deleted.
That theory did not survive. As KFF Health News reported, a court-appointed special master, Suzanne Segal, recommended dismissing the case, writing that “a mere possibility of an overpayment is not enough for the government to carry its burden.” The government, she found, had not compared the disputed codes against the underlying medical records to show any specific diagnosis was unsupported; its case “depends entirely on speculation.” She noted that CMS audits had found roughly 89 percent of the company’s codes were supported, and that CMS had known about — and, in 2014, backed away from regulating — the very chart-review practices it later attacked. UnitedHealth said the ruling confirmed “there was no evidence to support the DOJ’s claims we were overpaid or that we did anything wrong.” The recommendation can be appealed, and a separate, more recent Justice Department inquiry into the company’s coding remains open; those are allegations, not findings.
The ruling exposes the core enforcement problem. To win a False Claims Act case, the government must prove specific codes were both unsupported and knowingly false — a record-by-record burden across millions of encounters. Insurers counter, not without force, that aggressive coding reflects real illness in older populations, that traditional Medicare under-documents by comparison, and that CMS blessed their methods for years. Those defenses are why the settlements cluster around cases with damning internal evidence — queries, addenda, coders instructed to add but not subtract — and why the tens of billions MedPAC identifies as systemic overpayment largely sit beyond the reach of fraud law.
A new lever — and a familiar fight
Faced with the limits of litigation, CMS appears to be testing a different tool: its own regulatory leverage over the plans. In what health-policy researchers called a potential first, Elevance Health — which covers about 2 million Medicare members — wired $342,209,085.30 to CMS on May 27, court records show, after the agency threatened in February to bar the insurer from enrolling new members over what it called “substantial and persistent noncompliance” with rules requiring accurate billing and the return of overpayments. As KFF Health News reported, Elevance described the money as a “remittance of the total overpayment amount” estimated by government audits, while contesting the enforcement action as “unprecedented” and denying wrongdoing in a related, still-pending 2020 False Claims Act suit filed when the company was known as Anthem. In an April quarterly filing with the Securities and Exchange Commission, Elevance put its “current best estimate” of potential exposure in the matter at roughly $935 million — a figure that, against a payment of $342 million, suggests how much remains contested.
Whether the Elevance approach scales is the open question. “The payment Elevance is making here is not trivial,” Brookings Institution researcher Matthew Fiedler told KFF, but he cautioned that denting the overpayment problem would require CMS to collect “many similar payments” from “every” Advantage insurer. History counsels caution: in 2014, CMS drafted a rule to restrict chart reviews and then abandoned it amid what an official described as an industry “uproar.” This January, CMS revived the idea, proposing in its 2027 Advance Notice to curb chart-review-driven diagnoses while holding payment rates nearly flat. CMS Administrator Mehmet Oz framed the change as “protecting taxpayers from unnecessary spending that is not oriented towards addressing real health needs.” Insurer shares fell and the industry’s trade group, AHIP, warned the plan “could result in benefit cuts and higher costs for 35 million seniors.” A final decision is expected by early April.
What to watch
The through-line connecting these threads is not partisan; the overpayment problem has been documented across administrations, flagged by a nonpartisan commission, and pursued by Justice Department lawyers under multiple attorneys general. The question now is structural: whether enforcement and rule-making can close a gap that whistleblower settlements, however large, plainly cannot. Three tests will tell. First, whether CMS finalizes its chart-review limits this spring or, as in 2014, retreats under pressure. Second, whether the Elevance remittance proves to be a template other insurers follow or an outlier they litigate for years. Third, whether the government appeals the UnitedHealth ruling and, if it does, whether it can finally meet the burden the special master said it never carried.
Until then, the numbers point in opposite directions at once. Recoveries are climbing toward records. The estimated baseline of overpayment is climbing faster. A $556 million check makes headlines; a $40-billion-a-year coding gap does not. For a program that now covers more than half of all Medicare beneficiaries, the distance between those two figures is, increasingly, the whole story.
Note on sourcing and reply: Settlement figures and allegations are drawn from Justice Department announcements and court records; the resolved claims are allegations only, and the settling companies did not admit liability. Estimates of program-wide overpayment are attributed to HHS-OIG, MedPAC, and CMS. Statements from Aetna, Kaiser Permanente, UnitedHealth, Elevance, and AHIP are included where the companies have addressed the matters publicly; each disputes that its coding practices were improper.