CEO’s Murder Sparks Outcry Over UHC’s Coverage Denials

Law enforcement are looking for potential motives in the murder of Brian Thompson, CEO of UnitedHealthCare, and bullet shell casings reportedly found at the crime scene point to a possible cause – anger over UnitedHealthCare’s alleged practices for denying some coverage of medical care.

See Also: Maximizing data utility in mission delivery, citizen services, and education

The insurer has been under fire for its use of artificial intelligence and machine learning technology in claims decisions, and the New York Times and several other news outlets on Thursday reported that investigators are looking into the words “delay” and “deny” that appear to have been inscribed on casings collected after Thompson was shot outside a Hilton hotel where UnitedHealthCare was hosting an investors meeting on Wednesday.

Frustration – if not deep hostility – over the company’s alleged coverage decision practices – and the healthcare insurance industry overall – boiled over in dozens – if not hundreds – of the comments penned by current or past members on social media and on the feedback sections of news stories about Thompson’s death.

On top of all that, UnitedHealthcare is also facing proposed federal class action litigation alleging its use of AI and ML is driving the company to “systemically” deny claims.

One such lawsuit complaint filed in November 2023 and amended in April by the estates of eight deceased elderly members of UnitedHealthCare’s Medical Advantage Plans, claims that UHC has “illegally deployed AI in place of real medical professionals to wrongfully deny elderly patients care owed to them under Medicare Advantage Plans by overriding their treating physicians’ determinations as to medically necessary care based on an AI model that UHC know[s] has a 90% error rate” (see: Lawsuit: Health Insurer’s AI Tool Illegally Denies Claims).

The lawsuit alleges UHC’s use of AI is a “fraudulent scheme” giving the company “a clear financial windfall in the form of policy premiums and federal funding without having to pay for promised care, while the elderly are prematurely kicked out of care facilities nationwide or forced to deplete family savings to continue receiving necessary medical care, all because an AI model ‘disagrees’ with their real live doctors’ determinations.”

UHC did not immediately respond to Information Security Media Group’s request for comment on the lawsuit or on other allegations regarding the company’s coverage decision making practices.

‘Ethical Dilemma’

David N. Hoffman, an attorney and assistant professor of bioethics at Columbia University, said the death of the UnitedHealthcare’s CEO brings society back to the “core ethical dilemma” involving AI systems, and work that’s needed for them to mature, and gain acceptance and trust.

“We are seeing, time and time again, that systems that were designed and implemented to provide decision support are being morphed, either intentionally or unintentionally into the decision substitution tools,” he said.

Human oversight of the AI output can fail for a number of reasons, he said. “First among them is the time pressure put on those humans to generate work product. Second, is the lack of human insight into the foundational knowledge used by the AI system to generate its own work product,” he said.

As an example, Hoffman points to a commercial currently being used by Apple to promote its Apple Intelligence product.

“The scene opens with an employee at a meeting, being asked to summarize a prospectus that he clearly had failed to review in advance. He is seen rolling his chair back, clicking on the AI software to generate a summary, and then, without ever comparing the summary to the actual perspective document, he delivers the summary as his report,” he said.

“This is a glaring example of a software vendor openly, but unknowingly, acknowledging the risk of AI decision support becoming decision substitution.”

The litigation UnitedHealthcare faces from the estates of the elderly patients spotlights similar concerns – about “decision substitution and lack of transparency,” Hoffman said.

Federal regulators are also scrutinizing the use of AI by health plans to potentially deny claims.

Last week, the Department of Health and Human Services issued a proposed rule for “guardrails” to help ensure that the use of artificial intelligence for Medicare Advantage insurance plans does not result in inequitable access to healthcare-related services (see: Feds Proposed Guardrails for Medicare Advantage Plans).

UnitedHealthcare’s parent company, UnitedHealth Group, is also facing dozens of other lawsuits filed in federal court, many of them related to the cyberattack on UHG’s Change Healthcare IT services unit in February, which disrupted the healthcare ecosystem for months.