CLINICAL AI

A federal judge's recent decision to deny UnitedHealth Group's bid to narrow discovery proceedings in a high-profile artificial intelligence coverage denial lawsuit marks a pivotal moment in the intersection of healthcare technology and patient rights. The September 8 ruling by Minnesota District Judge John Tunheim rejected the insurance giant's attempt to limit the scope of evidence-gathering, instead allowing plaintiffs to pursue comprehensive discovery regarding UnitedHealth's alleged use of the nH Predict algorithm to systematically deny Medicare Advantage claims for post-acute care.
The class action lawsuit, filed by families of deceased Medicare Advantage members, alleges that UnitedHealth deployed NaviHealth's nH Predict algorithm to override physician determinations and prematurely terminate coverage for medically necessary services. Central to the plaintiffs' case is the claim that this AI system operates with a staggering 90% error rate, meaning nine out of ten appealed denials are ultimately reversed upon review. The algorithm allegedly applies rigid criteria that prioritize cost containment over individualized patient assessment, forcing elderly patients to either pay out-of-pocket for continuing care or forgo treatment entirely.
Judge Tunheim's decision to allow full discovery represents more than a procedural victory for plaintiffs—it signals a judicial willingness to scrutinize the black box of healthcare AI decision-making. The court found that UnitedHealth's proposed bifurcated discovery approach would create "unnecessary delays, duplicative litigation and disputes," while noting the company had already resisted discovery efforts in earlier proceedings. This ruling enables attorneys to examine internal documents, communications, and algorithmic processes that could reveal how deeply AI influences coverage determinations across UnitedHealth's Medicare Advantage operations.
The legal challenge occurs against a backdrop of intensifying regulatory scrutiny from the Centers for Medicare and Medicaid Services. In February 2024, CMS issued explicit guidance clarifying that while algorithms can assist in coverage determinations, they cannot solely dictate denial decisions. The agency emphasized that Medicare Advantage organizations must ensure their AI tools comply with all applicable coverage determination rules and cannot substitute algorithmic predictions for individualized medical necessity assessments. CMS specifically prohibited using algorithms as the sole basis for terminating post-acute care services, requiring consideration of each patient's unique circumstances and present condition.
The UnitedHealth case exemplifies a broader industry pattern that has drawn congressional attention. A Senate Permanent Subcommittee on Investigations report revealed that the three largest Medicare Advantage insurers—UnitedHealth, Humana, and CVS—dramatically increased post-acute care denials after implementing predictive technologies, with denial rates reaching approximately 25% of all requests by 2022. These findings have prompted calls for enhanced CMS oversight, including targeted audits of prior authorization data and expanded regulations governing predictive technology use in coverage decisions.
The implications extend beyond UnitedHealth, as similar lawsuits targeting Humana and other major insurers suggest systemic issues with AI deployment in healthcare coverage determinations. The American Medical Association has reported that 61% of physicians fear unregulated AI is increasing prior authorization denials, with 29% of doctors witnessing serious adverse events—including hospitalization, permanent injury, or death—resulting from coverage delays. These statistics underscore the clinical significance of ensuring AI systems enhance rather than undermine physician-patient decision-making.
As discovery proceeds in the UnitedHealth case, the healthcare industry faces a reckoning over the appropriate role of artificial intelligence in coverage determinations. The court's decision to permit comprehensive evidence-gathering sends a clear message that algorithmic opacity will not shield healthcare organizations from accountability. For Medicare Advantage beneficiaries and their providers, this legal precedent may herald a new era of transparency and oversight that prioritizes patient care over algorithmic efficiency, ultimately reshaping how AI technologies are implemented and regulated across the healthcare ecosystem.