UnitedHealthcare Ruling Highlights AI's Impact on Medicare Advantage Coverage
A federal magistrate judge in Minnesota recently instructed UnitedHealthcare (UHC) to release a significant collection of documents in the Estate of Lokken v. UnitedHealth Group, Inc. class action. This lawsuit claims that UHC improperly denied post-acute care coverage to Medicare Advantage enrollees through an artificial intelligence algorithm.
The March 9 ruling allows for extensive discovery into UHC’s AI-driven claims processes, highlighting a rising focus on AI-driven coverage denials in litigation. The lawsuit was initiated by families of two deceased Medicare Advantage beneficiaries, who filed the complaint in November 2023. They argue that UHC’s use of nH Predict, an AI tool from its subsidiary naviHealth, supplanted physicians' clinical judgments, leading to premature care denials. UHC maintains that nH Predict is merely a supportive tool and that medical necessity decisions are made by qualified physicians following CMS guidelines.
The court had previously denied part of UHC's motion to dismiss, allowing claims of breach of contract and the implied covenant of good faith and fair dealing to continue, while dismissing other claims due to Medicare Advantage preemption. The central issue is whether UHC's AI reliance contravened promises that decisions would be made by clinical staff. The court opted for broad class-wide discovery given the interconnected factual and legal elements, rejecting UHC’s request to limit this to certain periods.
Under the recent order, UHC is required to provide comprehensive document sets, including company policies on post-acute care claims and analyses related to nH Predict. The ruling also mandates the disclosure of employee training materials, performance reviews, and compensation data pertinent to care coordinators and medical directors. This includes information about the internal AI review board and its members.
UHC’s attempt to limit the scope of document requests was denied, with the court referring to previously addressed issues about restricting discovery. The court dismissed UHC's argument that pre-July 2019 records were irrelevant, noting that earlier data might offer crucial context for the alleged breaches.
A US Senate report cited by the court mentioned increased denial rates for post-acute care claims after the introduction of nH Predict, indicating that earlier records could set a relevant baseline. This decision affects healthcare providers by indicating what data might become pertinent in AI-driven denial litigations. Although the documents remain private, they could offer insights into the influence of AI tools on clinical judgments and coverage decisions.
Overall, this case underscores the insurance industry’s shift towards AI technologies in claim processing and decision-making. It accentuates the importance of transparency in AI usage for coverage decisions, aligning with regulatory and legislative measures. The industry should watch this and similar cases as indicators of potential changes in coverage and reimbursement practices driven by AI integration.