In April 2026, the Patient Refunds for Bad Denials Act was introduced in the House. It proposes fines starting at $10 million for any health insurer that denies more than 25% of claims in a year, plus $2 million for every additional percentage point above that threshold.

For healthcare providers, this is a signal that the documentation and denial management practices you build today will define your revenue cycle performance in the regulatory environment ahead.

What Is the Patient Refunds for Bad Denials Act?

The bill, introduced by Reps. Pat Ryan (NY-18) and Angie Craig (MN-02) authorize the Department of Health and Human Services to audit health insurers annually and penalize those whose claim denial percentage hits 20% or higher.

Correctly denied claims confirmed fraud or legitimate medical necessity exclusions don’t count toward the threshold. Insurers that actively work to reduce denial rates may see reduced penalties. And critically, the fines go back to enrollees, not into a government account.

One major gap: the bill doesn’t cover self-insured employer plans, which account for roughly 65% of workers with employer-sponsored coverage. That’s a significant blind spot in the current proposal.

So, where does the money go?

The legislation requires HHS to distribute penalty collections directly and proportionally to enrollees who were covered during the affected plan year. The bill also introduces a new transparency requirement: insurers must provide detailed written notices for any medical necessity denial, including specific clinical rationale, which directly affects the documentation and appeal landscape for providers.

The AI Factor in Insurance Claim Denials

AI-assisted claims review is what’s driving the legislative response,  and it’s the most urgent operational issue for providers right now.

Insurers now use algorithmic tools to process prior authorization requests and claims at scale, often without individual human clinical review. The error rate is well-documented: both Stanford University and Health Affairs published research in January 2026, finding that AI-assisted Medicare Advantage prior authorization denials are overturned on appeal approximately 82% of the time.

The most prominent case involves NH Predict, a UnitedHealth AI model facing a federal class-action lawsuit alleging it denied care based on statistical population averages, not individual patient circumstances. Active litigation as of 2026 is pushing courts to answer whether AI review meets the “individualized assessment” standard insurance contracts require.

The regulatory response to AI denials

Regulators at multiple levels are already responding. Here is where things stand as of May 2026:

• The CMS Interoperability and Prior Authorization Final Rule (CMS-0057-F) took effect January 1, 2026. It requires a 72-hour turnaround for urgent prior authorization requests and expanded data transparency reporting from payers on prior auth approval and denial rates.
• Arizona, Maryland, Texas, and Illinois each passed legislation in 2025 banning AI as the sole decision-maker for prior authorization or medical necessity denials. Human clinical review is now legally required before an adverse determination in those states.
• More than 50 major health plans, including Aetna, Cigna, Humana, Elevance Health, and UnitedHealthcare, pledged in June 2025 to approve at least 80% of prior auth requests in real time. The first public progress report under that pledge was due in spring 2026.
• KFF published updated federal and state AI regulation guidance in May 2026, documenting that AI tools must account for individual clinical circumstances and history, and algorithmic pattern-matching alone is insufficient.

For providers, the practical implication is this: when a payer’s AI system reviews your claim first, the documentation submitted with the original request carries more weight than ever. There is no human clinical reviewer weighing context at the front end. The clinical record has to make the case on its own, every time.

What Healthcare Organizations Should Do Now

Denial management is the set of processes healthcare organizations use to identify, appeal, and prevent insurance claim denials. High-performing revenue cycle operations are building these into standard practice, not as a reaction to legislation, but because the environment demands it now.

1. Anchor documentation to payer medical necessity criteria

   The most effective denial prevention happens before the claim is submitted. Clinical notes should explicitly establish why the service was medically necessary using language that maps to the specific payer’s coverage criteria, not just accurate clinical language. A trained documentation specialist, like a healthcare assistant who knows both clinical standards and payer criteria catch these gaps before submission.

2. Audit prior authorization requirements quarterly

   Payer PA requirements change more often than most providers realize. A quarterly audit across your top payer contracts surfaces requirement changes before they generate a denial wave. Build this into a standing workflow, not a reactive response.

3. Prepare for AI-first payer review environments

   Documentation for AI review needs to be more explicit than documentation written for human reviewers. Algorithmic systems look for diagnostic specificity, clear clinical justification language, and coding that precisely reflects documented care. Training your teams on what AI systems look for is now a standard part of operating in a high-denial environment.