By: Christopher Parrella, Esq., CPC, CHC, CPCO
Parrella Health Law, Boston, MA.
A Health Care Provider Defense and Compliance Firm.

The legal temperature around payer use of artificial intelligence just went up significantly. A federal court in Minnesota has ordered UnitedHealth Group to produce internal documents about its nH Predict algorithm, the AI tool allegedly used to determine coverage decisions for Medicare Advantage patients. While the ruling does not resolve the underlying claims it represents a major moment in the rapidly escalating battle over algorithm-driven claim denials.

The case, Estate of Lokken v. UnitedHealth Group, involves allegations that UnitedHealth used AI to improperly deny coverage for medical care for beneficiaries enrolled in its Medicare Advantage plans. Plaintiffs claim the system effectively replaced physician judgment with algorithmic decision-making and was used to prematurely terminate or deny coverage for medically necessary services. The March 9 ruling allows plaintiffs to obtain discovery about how the algorithm works its development goals, its intended benefits and whether it was designed to replace physician decision-making. The court also ordered UnitedHealth to disclose the identities of individuals involved in the design development implementation and approval of the AI tool. In other words the curtain is beginning to open on how a major payer’s algorithm was conceived and deployed.

For providers, this case matters far beyond UnitedHealth. Hospitals, physicians, post-acute providers and behavioral health organizations have increasingly complained that algorithm-driven utilization management tools are functioning as hidden gatekeepers for coverage decisions. The concern is not simply that algorithms exist but that they may be used to override clinical judgment or impose automated coverage limits that are inconsistent with the actual patient record.

Courts are beginning to take these allegations seriously. This is where things begin to get uncomfortable for payers. Internal communications design, decisions, performance metrics and implementation strategies can reveal whether an algorithm was truly intended to assist clinicians or whether it was primarily designed to control utilization and reduce claims payments. Once those documents surface in discovery, they often become powerful evidence not just in civil litigation but in regulatory investigations.
Here is the call to action for providers. If your organization routinely faces denials that appear automated algorithmic or disconnected from the clinical record, you should begin documenting those patterns carefully. Track denial timing language and appeal outcomes. Preserve correspondence with payers when automated determinations appear to override physician recommendations. Strengthen documentation supporting medical necessity particularly in areas like post-acute care behavioral health and complex chronic care where algorithms are most frequently applied.

The broader lesson is clear. AI-driven utilization management is rapidly becoming the next major litigation and regulatory battleground in healthcare. As courts begin to force transparency into these systems the heat will rise for payers that relied heavily on opaque algorithms to make coverage decisions. If you have any questions or comments about the subject of this blog or want assistance challenging payer denials that appear driven by automated decision tools, please contact Parrella Health Law at 857.328.0382 or Chris directly at cparrella@parrellahealthlaw.com.