Private Insurers Collected $462,000,000 for Strokes a Federal Audit Couldn’t Confirm Ever Happened. Every One of the 97 Cases Auditors Checked Failed.
Every year, Medicare Advantage insurers get paid more for sicker patients. That is by design — it keeps plans from cherry-picking healthy enrollees. But it also means a single diagnosis code, entered once by a doctor and never checked against a hospital record, can be worth real money to the company managing the plan. In a report published May 28, 2026, the HHS Office of Inspector General went looking for that gap in exactly one condition: acute stroke. It found the gap was enormous.
Auditors identified 240,401 Medicare Advantage enrollees nationwide who were coded in 2020 as having suffered an acute stroke, based only on a physician's office record — with no corresponding hospital admission anywhere in that year's data. They pulled a random sample of 97 of those cases for a hands-on look at the medical charts. The result: all 97, without exception, lacked medical-record support for the stroke diagnosis. Extrapolated across the full pool, OIG estimates CMS overpaid Medicare Advantage organizations $462,000,000 in 2021 alone because of it.
RealClearInvestigations’ Jeremy Portnoy turned the finding into the July 6, 2026 installment of “Waste of the Day,” the recurring waste-and-fraud column produced with OpenTheBooks.com. This page walks through what OIG found, why the incentive structure invites it, who has already paid to settle similar allegations, and who in government is now on the hook to fix it.
- $462,000,000 — HHS-OIG's estimated net overpayment to Medicare Advantage organizations for 2021, tied to unsupported acute stroke diagnosis codes · Source: HHS OIG Report A-02-23-01020
- 97 of 97 — sampled enrollees whose acute stroke diagnosis had zero supporting hospital or medical-record documentation — a 100% failure rate on the sample OIG pulled · Source: HHS OIG, May 2026
- 240,401 — Medicare Advantage enrollees nationwide flagged with an unsupported acute stroke code in 2020, the base population OIG sampled from · Source: HHS OIG
- 2020 — the year OIG first flagged this identical acute-stroke coding gap in traditional Medicare data — six years before this report found it still costing hundreds of millions · Source: HHS OIG 2020 report
- $556,000,000 — Kaiser Permanente's January 2026 False Claims Act settlement over separate Medicare Advantage upcoding allegations — the largest such settlement on record · Source: DOJ
Medicare Advantage is the privatized alternative to traditional Medicare: instead of the government paying doctors and hospitals directly, it pays a private insurer a fixed monthly amount per enrollee, adjusted up or down based on how sick CMS believes that enrollee to be. That “risk adjustment” system relies on diagnosis codes the insurer’s network submits to CMS. A stroke diagnosis is a high-value code — it signals an enrollee likely to need expensive, ongoing care, and it raises the monthly payment accordingly.
OIG’s audit tested whether those stroke codes held up. Investigators cross-referenced every acute stroke diagnosis Medicare Advantage plans submitted in 2020 against that same enrollee’s inpatient and outpatient hospital records for the year. An acute stroke is not a subtle event — it is, by definition, the kind of medical emergency that generates a hospital record. When 240,401 enrollees showed a stroke code with no hospital record anywhere to back it up, auditors pulled a random sample of 97 for a full chart review.
Every one of the 97 sampled cases failed. Reporting on the underlying charts, RealClearInvestigations found individual examples that illustrate why: one stroke diagnosis had been entered by a pharmacist, who is not permitted to diagnose a condition for Medicare purposes; one enrollee coded for an acute stroke was, per the record, sent home without any stroke treatment; and in at least 68 cases, the diagnosis date on file was for a stroke the patient had actually suffered years earlier — recoded as if it had happened in 2020, the year that generated the payment.
OIG’s recommendation is narrow and specific: CMS should build an automated check that blocks payment on an acute stroke code when there is no corresponding inpatient or outpatient hospital record for that enrollee in the same service year. As of the report’s release, CMS had not told OIG whether it agreed or disagreed with the recommendation.
This is not a story about a hacked computer system or a rogue billing clerk. It is a story about an incentive built into the plumbing of a $1-trillion-plus federal program. CMS pays Medicare Advantage organizations a base rate per enrollee, then adjusts that rate upward for every diagnosis code on file that signals higher expected cost. The insurer does not need to falsify a claim for a service that never happened — it only needs a diagnosis code to exist somewhere in an enrollee’s file, whether or not a hospital, a specialist, or anyone qualified to make that diagnosis ever confirmed it.
That is precisely the gap this audit exploited: of 774,000 Medicare Advantage enrollees who carried some form of acute stroke code in 2020, more than 240,401of them — roughly a third — had no hospital record to match. A physician’s office note, a chart-review vendor hired by the insurer, or in at least one sampled case a pharmacist, can be enough to trigger the higher payment, months or years before anyone checks whether the underlying medical event ever occurred.
This pattern has a name in health-policy circles: upcoding. It is the same mechanism behind a string of federal settlements over the last two years involving diagnoses ranging from HIV to morbid obesity to, now, acute stroke. Government auditors and independent budget analysts have separately estimated that Medicare Advantage overpayments tied to this kind of diagnosis inflation could cost taxpayers on the order of $1 trillion or more over the coming decade — the stroke-code finding is one specific, documented slice of a much larger pattern.
The most damning number in this story might not be $462,000,000— it might be the year 2020. HHS-OIG audited this exact vulnerability once before, in a report examining acute stroke diagnosis codes submitted by traditional Medicare providers that fed into Medicare Advantage risk scores, and found the same basic problem: stroke codes without hospital corroboration, driving up payments. That report is now six years old. The May 2026 report found the identical gap, at nine-figure annual cost, still uncorrected.
“It is time CMS faithfully executes its duty to audit these plans and ensure they are billing the government accurately for the coverage they provide to Medicare patients.”
Dr. Mehmet Oz · CMS Administrator · CMS press release, March 3, 2026
To be fair to the current CMS leadership, that quote comes from an actual policy shift: in March 2026, CMS announced it would audit all roughly 550 eligible Medicare Advantage contracts every payment year going forward — up from a historical practice of auditing only a small sample — and said it would grow its review staff from about 40 auditors to roughly 2,000. Whether that expanded audit force closes the specific stroke-code gap OIG identified is, as of this writing, still an open question CMS has not answered.
OIG’s May 2026 report is a multi-plan statistical sample, not a case against a single company, and it does not name which Medicare Advantage organizations submitted the 97unsupported stroke diagnoses in its sample. But the broader pattern of diagnosis-code inflation across the industry is not in dispute — it is documented in a growing list of federal settlements and separate, plan-specific audits.
In January 2026, Kaiser Permanente affiliates agreed to pay $556,000,000— the largest Medicare Advantage risk-adjustment False Claims Act settlement on record — to resolve allegations that from 2009 to 2018 the health system pressured physicians to add diagnoses to patient charts after the fact through “addenda,” generating roughly $1 billion in improper payments from about half a million added diagnoses. Kaiser’s settlement, like the others below, includes no admission of wrongdoing.
Aetna, a CVS Health subsidiary, separately agreed on March 11, 2026 to pay $117,700,000 to resolve allegations it ran a chart-review program that added diagnosis codes its own reviewers could not substantiate, and, from 2018 through 2023, submitted or failed to delete morbid-obesity codes for patients whose recorded body mass index didn’t support the diagnosis. In a separate, smaller audit released in 2025, OIG specifically flagged Coventry Health and Life — another CVS Medicare Advantage subsidiary, covering roughly 200,000 enrollees — for an estimated $6,995,522in overpayments tied to high-risk diagnosis codes that included acute stroke, sepsis, and prostate cancer. Patrick Jeswald, CVS’s chief compliance officer for Medicare, disputed that audit’s methodology in a formal response to OIG.
$556,000,000 — Kaiser Permanente, January 2026, largest Medicare Advantage risk-adjustment settlement on record.
$117,700,000 — Aetna/CVS, chart-review and morbid-obesity coding allegations.
$6,995,522 — Coventry Health and Life (CVS), 2018–2019 audit including acute stroke codes.
$2.1 billion — the amount the DOJ alleges UnitedHealth Group should have refunded under a still-contested, unresolved federal case; a special master has so far sided with UnitedHealth, and DOJ is pressing the case forward. This one remains disputed, not settled.
Two federal officials sit directly over this problem today. Dr. Mehmet Oz, the CMS Administrator, oversees the agency OIG says failed to prevent this specific overpayment and is now the one implementing the audit expansion meant to catch it going forward. T. March Bell, confirmed by the Senate in December 2025 as the seventh HHS Inspector General, now leads the office that produced both the 2020 warning and the 2026 report confirming the same gap was never closed.
Dr. Mehmet Oz — CMS Administrator. Announced the March 2026 expansion of Medicare Advantage audits from roughly 60 plans a year to all ~550 eligible contracts annually, and the growth of CMS's coding-review staff from about 40 to roughly 2,000.
T. March Bell — HHS Inspector General, confirmed by the Senate December 18, 2025. His office issued the May 2026 report; CMS had not stated whether it agrees with the office's recommendation as of publication.
Patrick Jeswald — CVS's chief compliance officer for Medicare. Formally disputed OIG's audit methodology on behalf of Coventry Health and Life in that plan's separate stroke-inclusive audit.
None of the officials above has been accused of personal wrongdoing in connection with this report — Oz and Bell are the government's own accountability apparatus, not defendants, and the audit expansion is a genuine response, even if its stroke-code fix is still pending. The accountability question this page raises is squarely about the private insurers collecting the money and the multi-year lag between a federal audit finding the gap and CMS actually closing it.
Strip away the acronyms and this is a simple story: a federal audit checked whether 97 stroke diagnoses insurers were paid extra for actually happened, and found that not one of them could be confirmed. Extrapolated across the full flagged population, that gap is worth an estimated $462,000,000 a year to the private companies running Medicare Advantage plans — on top of the $556,000,000, $117,700,000, and other settlements already paid for the same broader upcoding pattern in different diagnoses.
What makes this specific report notable isn’t just the dollar figure — it’s the recurrence. OIG flagged this precise vulnerability in traditional Medicare stroke-code data back in 2020. Six years, one change of administration, and one still-pending federal upcoding lawsuit later, the same coding gap surfaced again, larger, in a different corner of the same program. CMS says it is now auditing more plans, more often, with a far bigger staff. Whether that translates into the specific stroke-code fix OIG recommended — and whether the money already collected is ever clawed back — is the part of this story still being written.
Established, by federal audit: HHS-OIG sampled 97 Medicare Advantage acute stroke diagnoses; all 97 lacked supporting medical-record documentation; OIG estimates a resulting $462,000,000 net overpayment for 2021; OIG flagged the same underlying gap in a 2020 report.
Established, by settlement (different diagnoses, same industry pattern): Kaiser Permanente ($556,000,000) and Aetna/CVS ($117,700,000) have both resolved separate Medicare Advantage upcoding allegations; Coventry Health and Life (CVS) was separately audited for $6,995,522 in overpayments including stroke codes.
Still open: which specific insurers submitted the 97 unsupported stroke diagnoses in OIG's May 2026 sample; whether CMS adopts OIG's recommended payment-blocking fix; and the outcome of DOJ's separate, contested $2.1 billion case against UnitedHealth Group.
This is the kind of item OpenTheBooks’ “Waste of the Day” series exists to catalog: not a single dramatic heist, but a documented, structural gap in how a trillion-dollar federal program pays private companies — one that keeps reappearing because it keeps going unfixed. This page will be updated if CMS responds to OIG’s recommendation or if a specific insurer is identified in connection with the sampled cases.
