A lawsuit working its way through federal court in Kentucky is testing whether an insurer can be held liable for letting an algorithm decide when a patient’s rehabilitation care ends. In Barrows et al. v. Humana Inc., plaintiffs whose post-acute care coverage was terminated allege Humana used AI tools to improperly deny continued Medicare Advantage coverage, cutting off care the plaintiffs say they still medically needed. The case received a scheduling order on July 17, 2026, according to Georgetown Law’s litigation tracker, confirming it remains active and moving toward further proceedings rather than being dismissed or quietly settled.
What post-acute care denial actually means for patients
Post-acute care covers the stretch after a hospital stay — skilled nursing, rehabilitation therapy, home health services — when a patient is recovering but not yet able to manage independently. For Medicare Advantage enrollees, continued coverage of that care typically requires periodic reauthorization, and the Barrows plaintiffs allege Humana’s AI-assisted review process cut off that reauthorization prematurely, before patients had medically recovered enough to safely do without it, forcing some to either pay out of pocket or leave care earlier than their doctors recommended.
Part of a pattern across the industry
Barrows is one of several active lawsuits accusing major insurers of using algorithmic tools to systematically deny post-acute and rehabilitation care. UnitedHealth faces a parallel case in Minnesota over its nH Predict algorithm, where a federal judge ordered the company in March 2026 to produce internal documents after plaintiffs alleged a 90 percent error rate on appealed denials. Cigna faced a similar class action, Kisting-Leung v. Cigna, over its PxDx algorithm, though that case saw a dismissal order in August 2026 whose full resolution remains unclear from public records. Together, the cases suggest the use of AI in Medicare Advantage coverage denials has become a coordinated target for plaintiffs’ attorneys across multiple insurers rather than an isolated dispute with a single company.
Humana’s defense
Like other insurers facing similar suits, Humana has maintained that its AI tools function as decision support for clinical staff, not as automated denial engines, and that final coverage decisions rest with licensed medical reviewers who weigh the AI’s output alongside clinical judgment. Humana has not publicly detailed, in response to this specific case, exactly how much weight its reviewers give to an AI-generated recommendation before a denial letter goes out — a gap plaintiffs’ attorneys are likely to probe through discovery as the case proceeds.
The regulatory gap this case exposes
Georgetown health policy researchers have noted that the Trump administration’s first regulatory action on Medicare Advantage, the CY2026 final rule, did not adopt several previously proposed guardrails, including a rule that would have explicitly prohibited Medicare Advantage plans from using AI to discriminate against enrollees in prior authorization decisions, and did not finalize clearer requirements for internal coverage criteria disclosure. Researchers argue that legislation, not just rulemaking, may ultimately be necessary to close those gaps — meaning cases like Barrows are, for now, doing much of the regulatory work that rulemaking hasn’t finished.
What patients and families can do in the meantime
Patient advocacy groups recommend that Medicare Advantage enrollees facing care termination request the specific clinical criteria behind a denial in writing and pursue the formal appeals process aggressively, since internal appeal reversal rates — the same figures at the center of the UnitedHealth litigation — suggest a meaningful share of algorithmic denials don’t hold up once challenged.
What’s next
With the scheduling order now in place, Barrows moves toward discovery and likely toward the same kind of internal-document fight playing out in the UnitedHealth case. A ruling or settlement that establishes real transparency requirements for how Humana’s AI tool informs denial decisions could set a precedent other Medicare Advantage insurers would have to reckon with, particularly as 23 states and Washington, D.C. have already adopted the NAIC’s AI Model Bulletin aimed at governing exactly this kind of algorithmic decision-making in insurance. Elder-law attorneys who handle Medicare Advantage appeals say they are already citing the Barrows and Lokken litigation in their own client correspondence with insurers, using the pending cases as leverage to push for faster, more favorable resolution of individual coverage disputes before they ever reach a courtroom.
Photo: SHVETS production / PEXELS via Pexels
