UnitedHealth is currently at the center of a significant legal battle as a class-action lawsuit alleges that the company misused AI to deny specific insurance claims. This raises crucial questions about how health care companies utilize AI and its impact on the care received by patients.

Uncovering the Truth Behind AI in Health Insurance Claims

Part I: The Origins of the Issue

Christine Huberty, an attorney who used to work in Wisconsin providing free legal assistance to Medicare beneficiaries, noticed strange denials for care at skilled nursing facilities. One particular patient, who suffered a major stroke and was enrolled in Medicare Advantage through UnitedHealthcare, was put in a difficult situation. After just two weeks, he had to decide between going home when it wasn't safe or staying and paying out of pocket. This led Huberty to start collecting more denied claims and notice a pattern with UnitedHealthcare.The comment she made before the deadline for public comment on the Medicare Advantage system found an audience with Casey Ross and Bob Herman from STAT News. Their subsequent reporting uncovered the use of an algorithm called nH Predict, developed by Senior Metrics and later acquired by naviHealth and eventually UnitedHealthcare.

Part II: How the Algorithm Works

nH Predict uses a variety of data about patients, including their living situation, hospitalized condition, other illnesses, and demographic details, to make predictions about their length of stay. However, many aspects of the data and how it's used remain unclear as UnitedHealth and its subsidiaries were reluctant to provide detailed information.Amber Lynch, a former case manager with naviHealth, expressed concerns that comorbidities weren't factored into the algorithm's analysis. Caregivers on the front line often felt pressured to follow the algorithm's recommendations, even if it didn't align with what they saw in person with the patients.Over 90% of coverage denials were reversed or are reversed through UnitedHealth's internal appeals process, suggesting that the algorithm may be incorrectly denying coverage. In 2022, naviHealth told its case managers to keep patient stays within 1% of the algorithm's prediction.

Part III: The Federal Government's Involvement

Federal inspectors found that private insurers, including UnitedHealth, were straying beyond Medicare's coverage rules in using internally developed criteria to delay or deny care. The Office of Inspector General within Health and Human Services has been active on this issue.In October, a report from the U.S. Senate Permanent Subcommittee on Investigations found that Medicare Advantage insurers were intentionally using prior authorization to boost profits by targeting costly stays in post-acute care facilities.Ziad Obermeyer, an associate professor of health policy and management, emphasizes the importance of human oversight in claims decisions. UnitedHealthcare's use of AI is flawed as case managers are given instructions to adhere to the algorithm instead of using their own expertise.However, there is potential for an ethical use of AI in health insurance, such as using it to determine when additional care is needed. But major insurers like UnitedHealth insist on keeping proprietary data private, hindering the development of better tools.Casey Ross reflects on how reporting these stories has changed his view of health insurance in America. Insurance decision-making through algorithms feels opaque and can lead to unjust decisions that impact patients' lives.Christine Huberty suggests potential improvements, such as denials only being triggered by a change in the patient's condition and having a treating medical professional override for denials.For individuals facing denials, the only thing they can do is know their rights, ask questions, and appeal the decision. Gathering records and submitting an appeal directly to the insurance company can lead to a high overturn rate on appeal.