🔎︎
A new Columbia Law Review note by J.D. candidate Emma Ziegler argues that the increasing use of artificial intelligence by Medicare Advantage insurers to make medical-necessity determinations is causing real harm to beneficiaries and that the Centers for Medicare and Medicaid Services (CMS) needs to do more to address it. The note, "AI-Generated Denials: Medical Necessity in Medicare Advantage Today," appears in volume 126 of the journal and was peer reviewed.
The note opens with the case of Carol Clemens, who entered a skilled nursing facility after a life-threatening collapse in late December 2023. Ten days into her stay, her insurer denied coverage for the remainder of her physician-prescribed rehabilitation, citing that the continued stay was no longer medically necessary. At the time of denial, Clemens could not eat solid foods, walk unassisted, or speak more than a few words at a time. Her appeal was denied within a day. She returned home on January 3, 2024, and within three days fell again, sustaining a traumatic subarachnoid hemorrhage that required readmission and a restart of her rehabilitation. Clemens is one of multiple plaintiffs in a class action filed in Minnesota against UnitedHealth Group and its subsidiaries, including naviHealth. Parallel suits are pending against Humana in Kentucky and Cigna in California.
Ziegler situates these cases inside a broader pattern. A 2023 STAT investigative series by Casey Ross and Bob Herman documented widespread algorithmic denial of post-acute care for seniors. More than fifty members of Congress subsequently wrote to CMS urging stronger oversight, and a 2024 Senate Permanent Subcommittee on Investigations report examining practices at UnitedHealthcare, Humana, and CVS concluded that further intervention was warranted.
CMS issued a rule in 2023 intended to address some of these concerns about AI use in coverage determinations, but Ziegler joins other commentators in concluding that the rule leaves substantial questions unanswered, particularly around how insurers actually implement utilization management algorithms in day-to-day practice.
Her recommendations center on transparency and access. CMS should require Medicare Advantage insurers to share more information about how denials are generated and on what basis, and it should make the appeals process more genuinely accessible to individual beneficiaries, who currently bear the burden of contesting determinations made by systems they cannot see and may not even know are involved. The note frames these reforms as a first step rather than a complete fix, noting that broader regulation of coverage algorithms remains on the table.
For managed care leaders, the note signals that plans should expect rising pressure to defend determinations not only on outcomes but on process, with documentation, algorithmic transparency, and beneficiary appeals likely to become central compliance points in the years ahead.
