Skip to main content
frontier

Provider payouts jump to $14.85 billion in 2025 amid No Surprises Act arbitration surge

By Elena Petrova

The Surge

When the No Surprises Act took effect in 2022, the Congressional Budget Office projected roughly 17,000 payment disputes a year, plus another 4,900 from air ambulances, flowing through the new Independent Dispute Resolution process. The agencies designing the system assumed most fights would settle in the mandatory 30-day negotiation window before anyone filed for arbitration. They missed by two orders of magnitude.

From the system's launch in mid-2022 through May 2025, providers filed 3.3 million disputes, Georgetown University's Center on Health Insurance Reforms found in federal public-use files. The first nine months alone produced about 190,000 filings, more than ten times the annual forecast. Air ambulance companies, expected to generate roughly 5,000 cases a year, filed nearly 10,000 in that same span.

The backlog tells the story. As of May 2025, 85 percent of filed disputes had been closed, but nearly 500,000 remained pending. Median time to determination hit 81 days in the fourth quarter of 2024, down from a peak of 96, still nearly triple the 30-day statutory deadline. Only 34 percent of single disputes and 22 percent of batched disputes met that deadline.

Volume isn't the only distortion. Providers have won the vast majority of decisions. In 2024, they prevailed on 85 percent of line-item claims, up from 81 percent in 2023. When they win, the median payment determination has climbed steadily: 327 percent of the qualifying payment amount (QPA) in 2023, Georgetown's figures put, 445 percent in 2024, according to Georgetown, and 459 percent in Q4 2024 alone, CHIR's data shows. For HaloMD, a middleman organization that grew from 1 percent to 6 percent of line-item claims between 2023 and 2024, the median award reached 934 percent of QPA, Hoadley and Watts reported. When insurers win, their median prevailing offer sits at 110 percent of QPA, barely above the benchmark.

The money follows the wins. Arbitration firms collected roughly $1.3 billion in fees last year.

That spending is concentrated. Four states — Arizona, Florida, Tennessee, and Texas, produced 63 percent of all decided line-item claims in 2023-2024 while representing just 17 percent of the U.S. population. Texas alone accounted for 37 percent. California, Illinois, and Pennsylvania each registered around 1 percent. The skew tracks the footprint of two private-equity-backed provider groups: Radiology Partners and affiliates (28 percent of line items) and Team Health (15 percent). The top five provider organizations together drove 59 percent of claims.

Batching has become a force multiplier. By Q4 2024, 30 percent of emergency and non-emergency disputes were filed as batched cases, each averaging five similar line-item claims (median three). That pushed the 2023-2024 total of separate line-item claims with payment determinations to 2.5 million, the Center found, a number that dwarfs the original projections.

About 20 percent of closed disputes are deemed ineligible, clogging the pipeline further.

Metric Period/Type Amount Source/Notes
Provider payouts 2024 $4.08 billion Wall Street Journal analysis
Provider payouts 2025 $14.85 billion Wall Street Journal analysis
Provider payouts 2025 ~$15 billion Article statement (Inside the Platform)
Total system costs (Georgetown) Through 2024 At least $5 billion Georgetown modeling
Total system costs (Georgetown) Annual $2–2.5 billion Georgetown modeling
Administrative fees collected Through 2024 $218 million Federal public-use files / system data
Entity fees collected Through 2024 $636 million Federal public-use files / system data
Air ambulance entity fees collected Through 2024 $30 million Federal public-use files / system data
Administrative fee per dispute 2024 $50–$115 CMS rule / article
IDR entity fee (single dispute) 2024 $200–$840 CMS rule / article
IDR entity fee (batched dispute) 2024 $268–$1,173 CMS rule / article
Internal stakeholder cost per dispute Estimate (original rule) $857 Original rule
Additional administrative burden (implied) Through 2024 ~$1.9 billion Derived from $857 per dispute

The No Surprises Act succeeded in its core consumer protection — patients are largely shielded from surprise bills. But the arbitration engine it created has become a high-volume, high-cost channel that few anticipated. Independent practices without the scale of Radiology Partners or Team Health face a choice: absorb underpayments, hire expensive middlemen, or automate.

Inside the Platform

Clearest Health, founded in 2023 by Nicolas Raga with three employees per its Y Combinator profile, offers an AI-enabled revenue recovery engine built specifically for the Independent Dispute Resolution process established by the No Surprises Act. The company markets the service with no upfront cost and a free claims analysis, taking a contingency fee only when revenue is recovered. Its go-to-market runs through revenue-cycle management firms, specialty billers, and physician consultants, the channel partners that already manage the claims workflow for surgical groups, anesthesia practices, neuromonitoring providers, and emergency-medicine groups.

The platform exposes an embedded feed and white-label API so those partners can run benchmarking and dispute assembly inside their existing RCM tools. That channel strategy is how a small team scales to the thousands of independent providers generating IDR-eligible claims.

In a market where they collected that amount last year, the leverage comes from turning each decision into training data for the next one. A biller that can offer IDR appeals as a line item captures a slice of that fee pool while reducing the write-offs that drag down its own collection rates.

This channel approach mirrors a broader shift in health-tech investing. Bessemer Venture Partners describes the emerging category as "AI Services-as-Software": B2B services that leverage AI to sell an outcome or deliverable, tapping existing procurement processes with large operating-expense and services budgets rather than small IT and software budgets. Abridge, SmarterDx, Qventus, and Plenful follow the same logic: embed the AI where the work already happens, charge for the result.

The embedded-feed structure is the stickiest. UST, a global technology-services company, is integrating Anthropic's Claude into CarePath, an operational platform for health insurers and providers that manages member services, care coordination, and medical claims processing. Optum, part of UnitedHealth Group, unveiled a similar partnership with Anthropic on "responsible, frontier AI." The pattern is clear: the winners in health-tech AI will be the ones that disappear into the workflow.

Payers Strike Back

Zelis fired the most visible shot in the payer-side arms race on June 24, 2026, when it unveiled Zelis NSA Claim Advantage, an AI-native platform built to help health plans manage the No Surprises Act's dispute machinery. The launch arrived from a company that processes claims for more than 750 payers, including the top five national health plans, and that booked $2.39 billion in NSA-related claim savings in 2025 while keeping only about 8 percent of those claims from reaching arbitration. That scale gives Zelis a data moat smaller entrants cannot replicate: every dispute, every arbitrator ruling, every provider pattern flows through its pipes.

The timing was forced. In late May 2026, CMS finalized a rule overhaul that cut administrative fees from $115 to $15, imposed a five-day response window for arbitrator requests, formalized a 30-business-day open-negotiation period, and mandated standardized claim codes in payer communications. Jim Bridges, Zelis's president of price optimization, said: "Rule changes by CMS and HHS have improved the efficiency and long-term sustainability of the NSA, but payers are overwhelmed by complexity and costs associated with the IDR process." The rule adds deadlines payers cannot miss without forfeiting their position. Manual workflows and fragmented tools, still the norm at many plans, cannot keep pace.

Zelis's platform targets the full IDR lifecycle: automated intake, documentation, alerts, escalation workflows, QPA comparison analytics, real-time dashboards, and full communication audit trails. Its predictive layer surfaces certified IDR entity behavior and provider filing patterns to optimize settlement strategies and challenge non-qualifying disputes. The company says the goal is to move payers from reactive case management to a "strategic, data-driven operating model." That phrasing matters. It signals a shift from compliance tooling to competitive weaponry.

In the second quarter of 2025, arbitrators selected the provider's proposed rate in roughly 87 percent of closed disputes. Since April 2022, more than 5 million disputes have entered the federal IDR portal. July 2025 alone saw 213,585 new filings — more than the entire first year of the program, while certified entities closed 263,350 cases that month, outpacing intake by 23 percent. At the start of 2025, over 600,000 disputes awaited determination, and 69 percent of open disputes were older than 30 business days. By July 2025, 96.5 percent of all disputes submitted since inception were either resolved or under 30 days old, a clearance rate driven by capacity additions (15 certified IDR entities now, up from a handful in 2022) and portal upgrades. But the inflow keeps accelerating.

Zelis is not alone. The competitive pressure is structural: any plan processing meaningful out-of-network volume must now automate or absorb escalating administrative costs and adverse arbitration outcomes. The antitrust lawsuit filed by a California medical group — naming Zelis alongside Aetna, Cigna, Elevance, Humana, and UnitedHealth Group, adds a legal dimension. The complaint alleges shared reimbursement data was used to suppress out-of-network rates via a repricing tool. Zelis denies wrongdoing and says it operates in full compliance. The case will test whether the data advantages that make AI effective also create collusion risk.

For provider-facing platforms like Clearest Health, the payer response sharpens the asymmetry. Providers already win the vast majority of arbitrated cases. Payer AI aims to chip away at that edge by catching ineligible disputes earlier, optimizing QPA calculations, and forcing settlements before arbitration. The next phase of competition will be decided on data access: who sees more disputes, more arbitrator tendencies, more provider behavior, and who can act on it within the new regulatory deadlines.

The Talent Rush

The capital intensity of the IDR market is pulling specialized engineering talent into a corner of health-tech that previously relied on spreadsheets and offshore coding shops. Clearest Health illustrates the shift. The company is actively hiring across operations, engineering, and sales.

Competitive pressure compounds the demand. Zelis's AI-native NSA Claim Advantage platform creates immediate need for engineers who understand both the No Surprises Act's procedural deadlines, 30-day payment determination, 30-day entity decision — and the model-evaluation rigor required when a wrong prediction triggers a federal compliance violation. Smaller RCM vendors and regional billing shops, faced with the same volume surge, are either building in-house or buying from the new AI-native layer, further tightening the market for engineers with healthcare data fluency.

The skill profile is distinct from generalist AI hiring. Candidates need familiarity with X12 837/835 transaction sets, CMS qualifying payment amount methodology, and the evidentiary standards IDR entities apply when choosing between payer and provider offers. They also need to operate in a zero-tolerance environment: a hallucinated benchmark citation in an arbitration brief can disqualify an entire dispute batch. That constraint pushes teams toward retrieval-heavy architectures with explicit citation tracing, not opaque chat interfaces.

Andy Slavitt, former CMS administrator and now general partner at Town Hall Ventures, said in December 2025: "We're at the cusp of something that could be as transformative as anything." His fund's Medicare pilot investments signal that institutional capital views AI-driven revenue recovery as infrastructure, not a feature. If that thesis holds, the current hiring plans at Clearest Health — and similar plans at competitors, represents the leading edge of a talent migration that will scale with every billion dollars the IDR system awards.

What Comes Next

The No Surprises Act's arbitration mechanism is caught in a three-way squeeze: courts, regulators, and market forces are all pulling in different directions. The statute's "baseball-style" arbitration — where the IDR entity picks either the provider's or the insurer's offer with no middle ground, was designed for speed and finality. Congress specified that IDR decisions are binding and not subject to judicial review except in narrow circumstances. That finality is now fracturing.

In the first quarter of 2026, two federal district courts dismissed insurer lawsuits challenging IDR awards: a California court rejected Anthem's claims against HaloMD on April 9, and a Florida court dismissed Aetna's suit against Radiology Partners on April 16. Both insurers have appealed. But the judicial picture is inconsistent. As recently as January 2026, the Supreme Court declined to hear Guardian Flight's appeal of a Fifth Circuit ruling that the air ambulance company could not sue Health Care Service Corporation over unpaid IDR awards. Yet in March 2026, a Maryland district court held that the NSA "impliedly authorizes a very narrow private right of action" to enforce IDR awards — while an April 2026 Pennsylvania decision reached the opposite conclusion, dismissing Advanced Vascular Associates' suit against Horizon Blue Cross Blue Shield. The Fifth Circuit itself is rehearing en banc a challenge to the QPA calculation methodology, with briefing through September 2025. A circuit split on judicial review looks increasingly likely, which would tee up the issue for the Supreme Court.

Meanwhile, the Trump administration is reviewing a draft final rule to overhaul the IDR process. The scope isn't public, but the Biden-era proposed rule targeted two pressure points: new restrictions on batched disputes and an eligibility review overhaul. Those changes would directly address the payer lawsuits alleging that providers and "IDR middlemen" have weaponized the system by flooding it with ineligible disputes — about 20 percent of all closed cases, to overwhelm the process and extract default awards. The Department of Justice has weighed in through amicus briefs, arguing that IDR entities themselves shouldn't be proper parties to litigation, warning that such suits could thwart Congress's goal of a "low-cost, efficient" arbitration process.

IDR entities are paid only for eligible cases, creating a perceived motive to rule marginal cases eligible. Until that feedback loop is broken — or until the Supreme Court clarifies the judicial review bar, the arbitration boom that Clearest Health and its rivals are automating will keep expanding.

Scope Note

This article traces a narrow commercial arc: how Clearest Health's AI platform automates Independent Dispute Resolution appeals for out-of-network claims, how that automation converts the No Surprises Act's arbitration surge into recurring revenue for independent practices, and how incumbents such as Zelis are responding with their own AI-native tools. The story stops at the revenue-recovery workflow. It does not extend into the clinical encounter, the patient's out-of-pocket experience, or the wider policy debate over whether the No Surprises Act itself should be rewritten.

Clinical care delivery is outside the frame. Patient billing — the direct financial interaction between a practice and a patient, is also excluded. The No Surprises Act's core consumer protection prohibits surprise bills for out-of-network emergency care and certain ancillary services at in-network facilities. But the mechanics of patient-facing collections, payment plans, bad-debt write-offs, or health literacy interventions operate on a different causal chain.

Broader reform is not debated here. The article notes that they collected that amount in 2025 and that the longer-term impact of IDR awards on premiums remains "mostly unknown," per that center. It does not weigh proposals to replace the baseball-style arbitration with a benchmark formula, to expand the Qualifying Payment Amount methodology, or to revisit the independent dispute resolution entity certification process. The California antitrust suit against Zelis, Aetna, Cigna, and UnitedHealth, greenlit in 2026, signals regulatory scrutiny of payer-vendor entanglements, but the article treats that as a competitive backdrop for Zelis's AI launch, not as a policy analysis. Workforce shortages and burnout-driven quality risks appear in adjacent health-tech research, yet they do not drive the IDR volume surge; the surge comes from the statutory framework and the economics of out-of-network contracting.

In short: this story covers a software layer that turns a federal arbitration mandate into a scalable revenue line for a specific slice of providers. It does not cover the clinical decisions that generate the claims, the patient financial experience the law was written to shield, or the legislative roadmap that might rewrite the mandate itself.


Working in frontier tech? Zero G Talent tracks the openings: see every open ASML role, browse frontier tech jobs, openings at Stripe, and the people building the field.

Ready to Start Your Space Career?

Browse frontier jobs and find your next opportunity.

View frontier Jobs