UHC Billed for Daily Nurses 9 in 10 Never Got
Massachusetts checked 88,696 assessments. The billing records didn't match the paperwork.
Introduction
Massachusetts pulled 88,696 health assessments that UnitedHealthcare filed for its sickest elderly patients. In 99.3% of them, UHC told the state the patient was getting a visiting nurse seven days a week. Then the state checked the billing records. For 89.47% of those same patients, there was no nursing visit at all in the week before the assessment was signed, not a reduced schedule, none.
That gap, claimed on paper versus what the billing data shows, is the entire case. On May 29, Attorney General Andrea Joy Campbell filed a 66-page complaint in Suffolk Superior Court accusing UnitedHealthcare of inventing the conditions on those charts so it could bill the Commonwealth at a higher rate.
What a Diagnosis Is Worth
The pricing is what turns this from sloppy paperwork into a business model.
Massachusetts runs a program called Senior Care Options for people who qualify for both Medicare and Medicaid, the dual-eligible seniors who are usually the poorest and the sickest in the system. The state doesn't pay insurers per visit. It pays a flat monthly amount per enrolled member, and that amount goes up depending on how sick the state believes the member is. There are three tiers. Per the payment rates in the complaint, Level 1 paid roughly $1,300 a month in 2025. Level 2, reserved for members with a behavioral health or substance use diagnosis, paid about $1,800. Level 3, for members who need skilled nursing or daily care, paid about $4,300.
Now run the incentive. Move one patient from Level 1 to Level 2 and the state pays you roughly $500 more every month, about $6,000 a year, for the cost of typing in a depression code. Move that patient to Level 3 and you collect about $3,000 more a month, $36,000 a year, for writing that they need a daily nurse. You never have to send the nurse or treat the depression. The payment is keyed to the diagnosis, not to any care that follows it. UnitedHealthcare insures about 25,000 of the roughly 75,000 people in the program. The state says this ran from January 2015 through December 2025 and pulled at least $100 million it shouldn't have.
This flips the UHC story you already know. Every prior version, the AI denial algorithms, the nursing-home claims that got rejected, was about an insurer refusing to pay for care people actually needed. Here Massachusetts says UHC ran it backwards, manufacturing sickness on paper to collect more money. Same company, and the patients whose charts got rewritten were never told it happened.
The Diagnoses Nobody Ordered
The complaint describes a few different ways the charts got inflated.
The behavioral health one is the cleanest. Behavioral Health Business reported that UHC classified members at Level 2, the tier reserved for behavioral health and substance use disorders, despite those members lacking "a corresponding diagnosis or treatments associated with the condition." So the depression code is sitting on the chart, generating the higher payment, but there's no therapist and no prescription behind it, none of the things a real diagnosis produces. Just the code and the bill. Reuters reported that former nurses described being encouraged to upcode, in one case by diagnosing occasional headaches as migraines.
The skilled-nursing one is the 89.47% figure from the top. To bill Level 3 you have to attest the member needs skilled nursing daily or several times a week. UHC's assessments said yes almost every time, 99.3% claiming a nurse seven days a week, and the billing records say the nurse wasn't there. That's the data laid out by Law Commentary, drawn straight from the complaint's review of the assessments.
The individual case the AG put in the filing strips out the abstraction. One patient had Type 2 diabetes, hypertension, and arthritis. She managed all of it on over-the-counter painkillers and, in the complaint's words, "demonstrated complete independence." UnitedHealthcare classified her at a higher care level anyway and billed Massachusetts for five years. The overcharge on that one person came to $133,000. No doctor ordered the classification, and she never knew it existed. The only people who got anything out of it were the ones writing the bill.
The 2018 Audit They Kept Quiet
Coding mistakes happen. What makes this a fraud claim rather than a coding problem is what UHC did once it found out.
Back in 2018, MassHealth ran its own audit. It pulled 30 of UHC's Level 3 assessments and found that 16 of them, more than half, weren't supported by the rest of the member's documentation. UHC reviewed those members internally and quietly downgraded their classifications, moving people back down to the tier the records actually supported. So the company had seen the problem and fixed it on its own books.
What it didn't do was tell MassHealth or give back the money it had already collected on the inflated classifications. Reuters noted the same pattern: after internal reviews found improper diagnoses, UHC failed to reimburse the state for the overpayments. Under the Massachusetts False Claims Act, knowingly keeping money you know you weren't owed is its own violation, which is why the state can ask for treble damages. Triple the $100 million is $300 million.
Who Benefits
The beneficiary is UnitedHealth Group, and the internal documents in the complaint show people inside the company understood exactly what the Massachusetts plan was producing.
UHC's own paperwork described the Senior Care Options plan as "a strong financially performing program" carrying profit margins as high as 8%. Over the period in question the state paid UHC more than $5 billion for this single plan. A $100 million overcharge is about 2% of that total, but overpayments behave differently from regular revenue. There's no care cost attached to a diagnosis nobody treated, so that money lands almost entirely on the bottom line, where 2% of revenue turns into a far bigger share of profit.
The pressure to produce it traces straight up the org chart. Bernadette Di Re, who ran UHC's Massachusetts SCO operation, testified she resigned because of pressure from corporate management to grow membership, cut costs, "cut staff," "[g]et more numbers," and "[g]et more money from the state." Heather Cianfrocco, who oversaw UHC's Medicaid business, "publicly lambasted" Di Re for missing financial targets. Nurses were told to assess patients based on their "worst days" rather than their current condition. The state's word for the whole arrangement, quoting the internal directive, was "growth at all costs."
The formula does the rest of the work on its own. None of this is possible without a payment system that converts a diagnosis directly into a dollar amount and trusts the insurer to fill in the diagnosis. Massachusetts built its tiers that way, and the federal Medicare Advantage program runs on the same logic through the CMS risk-adjustment model, where each diagnosis code raises the monthly payment. The insurer controls the assessment, profits when the assessment says the patient is sicker, and the government usually doesn't check until years later. So the system pays a premium for exactly the diagnoses the AG says UHC fabricated.
This Is the Whole Industry, Not One State
If you think Massachusetts caught a one-off, look at what UnitedHealthcare paid Kaiser to settle five months earlier.
In January 2026, the Justice Department announced that Kaiser Permanente affiliates would pay $556 million to resolve False Claims Act allegations, the largest Medicare Advantage settlement of its kind. The DOJ's description reads like a template for the Massachusetts complaint. Kaiser mined patients' past records to find diagnoses it hadn't yet billed, then sent "queries" pushing physicians to add those diagnoses through "addenda," sometimes more than a year after the visit. It set aggressive diagnosis-adding targets and linked physician bonuses to hitting them. Its own doctors raised concerns that this was fraud, and the company kept going.
The federal numbers say this is structural. HHS's Office of Inspector General estimated that diagnoses appearing only on health risk assessments, with no other medical record to back them, drove $7.5 billion in Medicare Advantage payments for 2023 alone, touching 1.7 million enrollees. CMS agreed with one of the OIG's three recommendations to curb the practice and rejected the other two. A Senate Judiciary Committee investigation built on 50,000 pages of internal documents found UnitedHealth captured more diagnosis codes, and more federal reimbursement, than any other Medicare Advantage insurer, and the DOJ opened civil and criminal investigations into its Medicare billing in July 2025.
UnitedHealthcare calls the Massachusetts suit "meritless" and says it "doesn't accurately describe our Senior Care Options program." The company's defense across these cases tends to rest on a real distinction: documenting a patient's genuine conditions, even untreated ones, is legal, while inventing conditions that have no clinical basis is not. The whole trial comes down to which side of that line the records fall on, and the 89.47% figure is the state's bet that it's the wrong one.
The Bottom Line
Underneath all of it is one question: who controls the paperwork. The same company gets to assess how sick you are and gets paid more when the answer is "very." Massachusetts is the rare case where someone compared what the insurer wrote down against what the billing records show, found a gap of nearly nine in ten on the nursing claims, and decided it was worth $300 million to litigate. The federal government had a decade of the same signals and a much larger version of the same formula, and it took a state attorney general to put the comparison in front of a judge.
CMS says it's expanding its Medicare Advantage audits now, lifting the cap from 35 records per plan to as many as 200 and moving to review every plan instead of 60. The open question is whether that catches anything, because a federal court struck down the rule in September 2025 that would have let CMS take what it finds in a sample and apply it across an insurer's whole population. Audit a hundred charts, find the fabrications, and you may only be allowed to recover on those hundred. For a plan covering tens of thousands of people, a penalty that small stops being a deterrent and becomes a line item you budget for.