UnitedHealth Gets Extra Month to Turn Over AI Claim Denial Documents
A Minnesota judge gave UnitedHealth Group until April 29 to hand over tens of thousands of documents in a lawsuit alleging the insurer uses an AI algorithm to improperly deny rehabilitation care to seriously ill patients.
The defendants - UnitedHealth Group, UnitedHealthcare and naviHealth - requested a 30-day extension from the original March 30 deadline, citing the volume and complexity of records spanning nearly nine years.
Magistrate Judge Shannon G. Elkins granted the extension without comment. Plaintiffs did not oppose it, provided they receive a 45-day extension of their own May 15 deadline to submit expert declarations.
What the lawsuit claims
The case, filed in November 2023, centers on allegations that UnitedHealth relies on an algorithm to deny rehabilitation care while knowing the system has a 90% error rate. The plaintiffs - former patients or their heirs with UnitedHealthcare Medicare Advantage plans - claim the insurer profits by denying necessary care.
"The elderly are prematurely kicked out of care facilities nationwide or forced to deplete family savings to continue receiving necessary medical care, all because an AI model 'disagrees' with their real live doctors' determinations," the lawsuit states.
The lawsuit followed an investigation that detailed how UnitedHealth allegedly used a computer algorithm to cut off care to older patients. Internal documents cited in the case show UnitedHealth managers set a goal for clinical employees to keep patients' rehab stays within 1% of the days projected by the algorithm.
UnitedHealth's position
UnitedHealth acquired naviHealth, which developed the nH Predict AI model, for over $1 billion in May 2020. The insurer has said it uses the AI tool merely as a guide in making decisions.
In a statement, a UnitedHealth spokesperson said: "Coverage decisions are only made by medical directors - not by AI - in accordance with CMS Medicare coverage criteria." The company characterized the lawsuit as based on "unfounded allegations" and said the tool helps clinicians suggest appropriate care and plan next steps.
UnitedHealth's motion to dismiss, largely denied in February 2025, had argued that patients named in the lawsuit had not completed appeals of their denied claims.
Discovery challenges cited
UnitedHealth said it began work immediately after the court's March 9 order requiring responses to nine document requests and four interrogatories. The company has reviewed more than 40,000 documents and produced over 60,000 pages so far.
However, certain categories of data require additional time due to multiple steps: identification, collection, processing, review and production. Materials at issue include previously collected documents, new custodial data from earlier periods, records from other legal matters and targeted collections such as internal policies and training materials.
One request involving identification of hundreds of individuals and related employment records was described as particularly complex and still in progress. The effort has required coordination among in-house and outside counsel, employees across multiple departments and an external e-discovery vendor.
For insurance professionals, this case raises questions about how AI for Insurance systems are deployed in coverage decisions and what oversight mechanisms exist. The discovery process will likely reveal how UnitedHealth integrated AI into its claims operations - information relevant across the industry.
Your membership also unlocks:
