UnitedHealth used faulty AI to deny medically necessary insurance to elderly patients, lawsuit claims

The families of two now-deceased former UnitedHealth beneficiaries have filed a lawsuit against the health care giant, alleging it knowingly used a flawed artificial intelligence algorithm to deny elderly patients the extended care coverage their doctors deemed necessary.

The lawsuit, filed last Tuesday in Minnesota federal court, alleges that UnitedHealth illegally denied “elderly patients care owed to them by Medicare Advantage Plans” by using an AI model known by the company to have a 90% error rate in overestimating the decision. : to the patients’ doctors that the expenses were medically necessary.

“Seniors across the country are being pushed out of care facilities prematurely or forced to use up family savings to continue receiving the medical care they need, all because: [UnitedHealth’s] The AI ​​model “does not agree” with the decisions of their real live doctors, the complaint states.

Medicare Advantage plans, administered by private health insurers such as UnitedHealth, are Medicare-approved insurance plans available to seniors as an alternative to traditional federal health insurance plans, according to the U.S. Centers for Medicare and Medicaid Services.

Use of an allegedly flawed AI model developed by NaviHealth, called “nH Predict,” allowed the insurance company to “prematurely and unfairly stop paying” its elderly beneficiaries, causing them medical or financial hardship, the lawsuit alleges.

Using AI to determine health coverage

NaviHealth spokesman Aaron Albright told CBS MoneyWatch that the AI-powered tool is not used to make coverage decisions, but as a “help guide. [UnitedHealth] inform providers … of what type of help and care the patient may need.”

Coverage decisions are ultimately “based on CMS coverage criteria and the terms of the member’s plan,” Albright said, adding that the lawsuit “has no merit.”

In their complaint, however, the families accuse UnitedHealth of using erroneous AI to deny claims as part of a financial scheme to collect premiums without paying coverage for elderly beneficiaries who believe they lack the knowledge and resources to “appeal the erroneous AI.” working with intelligence. decisions.”

UnitedHealth continues to “systematically deny claims using their flawed AI model because they know that only a tiny minority of policyholders (roughly 0.2%)1 will appeal denied claims, and the vast majority will either pay out of pocket or I will give up the remaining amount. about their assigned post-acute care.”

The family’s attorneys want to represent “All persons who purchased Medicare Advantage Plan health insurance from the defendants in the United States during the four years prior to the filing of the complaint until now.”

The utility of AI in the health insurance industry

The implementation of artificial intelligence algorithms can help health insurance companies automate 50% to 75% of the manual work involved in approving insurance requests, such as gathering medical information and validating dates in conjunction with patient records, resulting in a faster turnaround, which can benefit beneficiaries, counseling. The McKinsey company said last year.

However, some doctors have advised health insurers to temper their expectations about the usefulness of AI in health insurance.

In June, the American Medical Association (AMA) praised the use of artificial intelligence to “speed up the prior authorization process,” but called on health insurers to require a human review of patient records before denying care to their beneficiaries.

“AI is not a silver bullet,” AMA board member Dr. Marilyn Heine said in a statement.

Cigna health insurer doctors rejected more than 300,000 claims in a two-month review process using artificial intelligence, according to a ProPublica review.