A federal watchdog has concluded that HumanaChoice, a Medicare Advantage plan run by Humana, Inc. under contract H5216, collected government payments tied to diagnosis codes that patients’ own medical records did not back up. The finding comes from a compliance audit covering payment years 2020 and 2021, built on a sample of enrollee records that mostly failed to hold up under review. Humana has disagreed with some of the findings and with every recommendation the auditors made.
The Audit Behind the $130.9 Million Figure
The Department of Health and Human Services’ Office of Inspector General examined HumanaChoice, a preferred provider organization based in Louisville, Kentucky, that Humana, Inc. administers under CMS contract number H5216. As of December 2021, HumanaChoice covered 1,730,650 enrollees across multiple states, and CMS paid the plan roughly $37.8 billion for the 2020 and 2021 payment years the audit reviewed.
Within that spending, auditors identified 68,701 enrollee-years tied to diagnosis codes flagged as high-risk for miscoding, associated with $165,672,058 in payments. From that pool, they sampled 220 enrollee-years and checked each diagnosis against the medical record behind it. Medical records failed to support the diagnosis codes in 178 of those 220 enrollee-years, producing $669,237 in overpayments inside the sample alone. Extrapolating that failure rate across the full pool, the inspector general’s office estimated HumanaChoice received at least $130.9 million in overpayments for 2020 and 2021, under Report A-05-24-00010, issued September 10, 2026 and posted to the inspector general’s site five days later.
Free download: How to read the Annual Notice of Change, spot the cost, drug and network changes, and decide whether to keep or switch. Get the free 2027 plan-change checklist.
How Medicare Advantage Ties Payments to Diagnosis Codes
Medicare Advantage works differently from traditional Medicare in one structural way: CMS pays each private plan a set monthly amount per enrollee, calculated by multiplying a base rate tied to the plan’s bid against a regional benchmark by an individual risk score. That risk score comes from mapping an enrollee’s diagnosis codes into government-defined categories called Hierarchical Condition Categories, each carrying its own numerical weight. More or higher-weighted categories mean a higher risk score, and a higher risk score means a bigger monthly payment to the plan.
A diagnosis code that isn’t backed by a documented clinical finding still raises that risk score, and the payment that follows it, even when no matching care was actually delivered. That is the mechanism the inspector general’s office tests when it periodically samples an insurer’s diagnosis codes against the medical records behind them, focusing on categories with a documented history of being submitted without adequate support. Across the whole Medicare Advantage program, the stakes are large: CMS paid MA organizations $494 billion in 2024, or 44 percent of all Medicare spending that year.
The Eleven Diagnosis Categories Under Review
Auditors did not review HumanaChoice’s records at random. Using data-mining techniques and input from medical professionals, the inspector general’s office had already grouped specific diagnoses into eleven categories considered prone to miscoding: acute stroke, acute myocardial infarction, embolism, five separate cancer types (lung, breast, colon, prostate and ovarian), sepsis, severe pressure ulcers and a category the report calls potentially mis-keyed diagnosis codes, where a data-entry error such as a transposed digit can turn one condition into an entirely different one on paper.
In most of those categories, the test for support was concrete: a stroke or heart-attack diagnosis logged on a single physician visit needed a matching hospital claim; a cancer diagnosis needed evidence of surgery, radiation or chemotherapy within six months; a sepsis diagnosis needed an inpatient admission. Where that corroborating record was missing, the diagnosis code should typically have been logged as a history of the condition instead, a distinction that does not carry the same payment weight.
Why the Finding Doesn’t Change What Enrollees Owe or Where They’re Covered
Nothing in the audit asks individual HumanaChoice members to repay money or changes what they paid in premiums or copays during 2020 and 2021. The overpayment the inspector general identified ran from CMS to Humana, Inc., based on diagnosis codes the insurer submitted on enrollees’ behalf, not on anything an enrollee filed or claimed, and the recommended fix runs the same direction: money moving from the insurer back to the federal government. Contract H5216 also covers only a portion of Humana’s total Medicare Advantage enrollment; Humana administers other MA contracts nationwide, each with its own CMS contract number and its own separate compliance history, so this finding does not by itself establish anything about coding accuracy on those other contracts.
What the Inspector General Wants Humana to Do Next
The audit closes with four recommendations. Humana, Inc. is asked to refund the federal government the $130.9 million in estimated overpayments; to review 212 additional enrollee-years in one high-risk diagnosis group that auditors did not personally examine and refund any overpayments those records fail to support; to check for similar unsupported diagnoses in years after the audit period and refund those as well; and to keep strengthening the internal procedures meant to catch unsupported codes before they reach CMS.
According to the inspector general’s published report highlights, Humana, Inc. disagreed with some of the findings and with all four recommendations. The audit record stands as issued regardless of that disagreement, and it now becomes part of the public compliance history CMS and future auditors can draw on when they next examine how HumanaChoice’s diagnosis coding holds up against the medical records behind it.
The Only Window to Compare Plans Runs October 15 Through December 7
The audit above turns on records CMS and Humana argue about years after the fact; it says nothing about whether a current HumanaChoice or other Medicare Advantage enrollee’s own plan still matches their prescriptions, doctors and budget going into next year. That comparison has its own fixed calendar, one rarely printed next to the coding rules that determine what a plan gets paid. Once the annual window to switch plans closes, the choice locks in for another full year regardless of what a later audit finds.
The 2027 Medicare Open Enrollment Decision Kit includes a cost calculator spreadsheet that compares plans on cost, drugs and doctors and a prescription-by-plan comparison.
See the cost calculator spreadsheet that compares plans on cost, drugs and doctors in The 2027 Medicare Open Enrollment Decision Kit.
AI-assisted research and drafting; reviewed by an editor.
