A Doctor Billed $152,594. He Got $7,879.

Introduction

A doctor bills $152,594. UnitedHealthcare runs the claim through MultiPlan's algorithm, the algorithm recommends paying $7,879, and UnitedHealthcare pays it. Then it charges the employer a $50,650 fee for the service. So the employer paid $58,529 in combined fees and payment, more than seven times what the doctor got for actually treating the patient. That example comes from the New York Times investigation, built on more than 50,000 pages of documents two federal judges unsealed. And if the doctor had refused to swallow $7,879 as payment in full, the $144,715 gap would have landed on the patient as a balance bill.

MultiPlan earned a cut of every dollar in that gap.

How a Repricing Fee Becomes a Toll Booth

MultiPlan, which rebranded as Claritev in February 2025, sits between insurers and out-of-network doctors. When you see a doctor outside your plan's network, the insurer routes the bill to MultiPlan's tool, Data iSight. The tool spits back a recommended payment, and that recommendation is designed to land low. MultiPlan's own product page says Data iSight targets "savings of 61%-81% off billed charges."

Here is the part that makes it a toll booth rather than a cost-control service. MultiPlan gets paid a percentage of the "savings," meaning the gap between what the doctor billed and what the insurer actually pays. The AMA put the incentive in one line: "MultiPlan gets paid more as physicians get paid less." There is no version of this where the company earns more by recommending a higher payment to your doctor. The bigger the underpayment, the bigger the fee, and that inversion is the whole business model. The algorithm doesn't process the gap, it manufactures the gap, and everyone in the chain except the doctor and the patient earns a percentage of it.

The figures are not small. MultiPlan's repricing revenue grew from $23 million in 2012 to $709 million in 2021, per the AMA's lawsuit, roughly a 3,000% climb in nine years. In 2024 Claritev says it identified $24.7 billion in "potential medical cost savings," which is both the fee base and, by definition, money that never reached the people who delivered the care.

Nine Insurers, One Machine

The detail that turns this from a billing scandal into an antitrust case is who else was feeding the same algorithm. The Arizona Attorney General's complaint, filed June 1, 2026, names eight of the largest health insurers in the country as defendants: Aetna, Cigna, UnitedHealthcare, Humana, Elevance, HCSC, Molina, and Centene. Blue Cross Blue Shield entities are named separately in the federal MDL, bringing the total to nine major insurer groups alleged to have priced through the same machine. The complaint says MultiPlan's clients cover 81% of the commercial out-of-network buyer-side market, and that the company processes 360,000 claims a day off a warehouse it touted at "approximately 15 petabytes of data."

What makes that a problem is sworn testimony from MultiPlan's own VP of health economics, Sean Crandell. He testified that Data iSight's base methodology "cannot even factor in [who] the client [is]," reported by the Boondoggle newsletter from the court record. Asked whether the rates change depending on which insurer is using the tool, he answered: "No." Competing insurers, in other words, were getting the same number for the same procedure code, generated from a shared pool of each other's confidential pricing data. (Insurers can layer their own downward "overrides" on top, a defense point the company leans on, though plaintiffs argue that an insurer setting a ceiling below the algorithm's output is evidence of suppression, not innocence.)

There is also a surveillance layer. The Arizona complaint describes a MultiPlan analytics product that let any insurer client see, in real time, whether competitors were reimbursing at higher or lower rates. A system where competitors price through a shared algorithm and can watch each other's rates is the structure courts call a hub-and-spoke conspiracy, MultiPlan at the hub, the insurers as spokes.

They Did This Once Already

If the scheme feels familiar, that is because the same companies ran a version of it before. The Arizona complaint walks through Ingenix, a UnitedHealth-owned database that calculated "usual and customary" rates 10% to 28% lower than the real benchmarks. An AMA class action over Ingenix settled in 2009; UnitedHealth paid $350 million to class members, contributed $50 million to create the independent nonprofit FAIR Health, and Aetna paid $20 million while Cigna and Elevance's predecessor paid $10 million each.

The most damning receipt in the complaint is a 2015 email from a Cigna executive, quoted verbatim. Explaining why Cigna needed an outside vendor to set out-of-network rates rather than building them in-house, the executive wrote: "We cannot develop these charges internally (think of when Ingenix was sued for creating out-of-network reimbursements). We need someone (external to Cigna) to develop acceptable rates."

That's a company citing the last antitrust case as the design spec for the next one: outsource the number-setting to a shared third party so no single insurer is the one "creating" the rates. Cigna began using Data iSight in April 2015, the complaint says, the earliest of the insurers. UnitedHealth followed in 2016.

The Wall of Lawsuits

The legal pressure right now is unusual for how broad it is. A federal MDL in the Northern District of Illinois has consolidated more than 100 provider lawsuits, and in June 2025 Judge Matthew Kennelly denied the defendants' motions to dismiss, letting nearly all the antitrust and consumer-protection claims proceed to discovery. His ruling included a line that should worry every defendant: "An agreement to fix prices within a below-market range through use of an algorithm is no different for antitrust purposes than an agreement to fix prices to a single point."

The MDL runs on a hub-and-spoke theory. Arizona AG Kris Mayes and the DOJ frame it harder, as a per se illegal buyers' cartel among competing insurers, a legally distinct argument that hasn't been tested on the merits yet. The DOJ filed a 14-page Statement of Interest in March 2025 backing the providers on two legal points: that a common pricing algorithm can be concerted action under the Sherman Act, and that competitors swapping sensitive data through an intermediary can violate the law. That filing came under the Trump administration's antitrust division, which means the scrutiny isn't a one-party project.

Then there's the criminal track. In its May 18, 2026 8-K, Claritev disclosed it had received a grand jury subpoena from the DOJ Antitrust Division back in August 2024, saying it has "been cooperating fully" and is not a named target. The stock dropped 15% the day the probe was reported, then recovered. The first bellwether trial is set for December 7, 2027.

Who Benefits

Three parties earn off the gap. MultiPlan/Claritev takes the first cut, a percentage of every dollar of "savings" it generates. With a roughly 62% adjusted EBITDA margin in 2024, the algorithm is its profit engine, and it only makes money when your doctor gets paid less.

The insurers take a second cut. Most large employers self-fund their health plans, paying claims directly while the insurer administers them, and the insurer charges those employers a "shared savings" fee on top of MultiPlan's take. That is how an employer ends up paying a $50,650 fee on a single cut claim. The Times found UnitedHealthcare alone reaped about $1 billion in fees from out-of-network savings programs. When UnitedHealth built an in-house replacement called Naviguard that would have cut its fees in one Texas county from $263,000 a year to $30,000, the Arizona complaint alleges MultiPlan pressured it to kill the project before launch — which the complaint frames as the alleged cartel protecting its fee base.

Hellman & Friedman, the private equity firm that has owned MultiPlan since 2016, is the third beneficiary. H&F bought the company for about $4.4 billion in 2014 and took it public via SPAC in 2020 at an enterprise value near $11 billion, remaining the largest shareholder. The value creation came from a near-monopoly on out-of-network claims processing, the kind of position where switching costs lock everyone in once an insurer wires Data iSight into its claims workflow.

The losers are the doctors who eat below-cost payments, the employers paying fees that run higher than the care itself, and the patients who get balance-billed for coverage they already bought.

The Coverage You Paid For

The reason this keeps running is partly a statutory hole. The No Surprises Act, effective in 2022, banned surprise balance bills for emergency care and certain other services. But self-funded employer plans, which cover most people with job-based insurance, are governed by federal ERISA rules that leave state insurance regulators without jurisdiction. The algorithm keeps processing those claims. Senators Ron Wyden and Bernie Sanders wrote to MultiPlan in May 2024 warning that its fee structure creates "an improper conflict of interest" with a plan's duty to deliver promised benefits under ERISA.

Strip away the antitrust framing and the consumer harm is simpler than a cartel theory. You pay a higher premium for a PPO specifically because it covers out-of-network care. The premium is the promise. When the claim comes back paid at a fraction of the bill and the rest shows up as a balance bill, you paid for coverage and received a fee-extraction system instead. The newest case shows the mechanism still grinding in real time: Lifepoint's June 2026 complaint alleges that for one insurer in a single quarter of 2025, Data iSight processed $44.6 million in changes and generated $36.3 million in underpayments.

Still Pricing the Next Claim

The clean version of the story is that an algorithm found waste and cut it. The version in the court records is that nine competing insurers and a private-equity-owned vendor priced your out-of-network care through one machine that earns more the less your doctor gets paid, then recycled those low payments back in as the new baseline so the next number comes in lower still. Judge Kennelly already ruled the providers can take that theory to discovery, and the DOJ's own filing says the legal framework holds.

The trial that decides whether this gets called price-fixing in front of a jury is set for December 2027. Until then, the algorithm that produced the $50,650 fee is still the one pricing the next claim. The comparable healthcare antitrust case, the Blue Cross Blue Shield MDL, took more than a decade and settled for $2.8 billion without anyone admitting the law was broken. The open question is whether the people who paid the premiums ever see a dollar of the money the complaints say was taken from them, or whether this settles the way these always seem to, with a number large enough to make headlines and small enough to keep the machine on.