California health insurers will be prohibited from relying solely on artificial intelligence to deny claims under new legislation that was authored by state Sen. Josh Becker and that will take effect next month.
The legislation, Senate Bill 1120, aims to take on one of the mechanisms that large insurance companies have been using to make the process of denying claims more efficient. The topic took on greater public scrutiny after Brian Thompson, CEO of the insurance giant UnitedHealthcare, was gunned down on Dec. 4 in Midtown Manhattan. Police believe that the alleged shooter, Luigi Mangione, was motivated by his hostility toward the nation’s health insurance industry. UnitedHealthcare is among the insurance giants that have been subject to class action lawsuits and media scrutiny because of their use of AI to deny claims.
Becker’s bill, which is known as the Physicians Make Decisions Act, mandates that any denial, delay or modification of care based on medical necessity be reviewed by a licensed physician. Becker said in a statement that while artificial intelligence has immense potential to enhance healthcare delivery, “it should never replace the expertise and judgment of physicians.”
“An algorithm cannot fully understand a patient’s unique medical history or needs, and its misuse can lead to devastating consequences. SB 1120 ensures that human oversight remains at the heart of healthcare decisions, safeguarding Californians’ access to the quality care they deserve,” Becker said.
The Becker bill won overwhelming support in the state Capitol over the summer and was signed into law by Gov. Gavin Newsom on Sept. 28. It is sponsored by the California Medical Association, which represents about 50,000 physicians. Dr. Anna Yap, a physician with the UC Davis Health Emergency Medicine, said during a June hearing that doctors have been using artificial intelligence for years to assist with diagnosis and testing.
AI models can, for example, detect diseases such as cancer. But while AI models that doctors use to develop treatments are evaluated by the Food and Drug Administration, the tools that insurers use do not face the same scrutiny, said Yap, who was representing the California Medical Association as a bill sponsor.
“In general, physicians are enthusiastic about the potential of AI in health care, but we also want to guarantee that it’s done in a safe, responsible and ethical manner for our patients,” Yap said.
Becker, D-Menlo Park, argued during committee hearings earlier this year that the bill is a necessary safeguard to address the growing use of artificial intelligence by the insurance industry. Health insurance companies have increasingly turned to algorithms to streamline processing of claims and prior authorization requests. Bloomberg Law reported earlier this year that Humana, Cigna and UnitedHealthcare were among the insurers that are facing class actions from consumers for using advanced technology to deny claims.
One lawsuit against Humana, which was filed in Kentucky, argued that the insurer uses a computer program called nH Predict to save money by denying claims that they would be otherwise be required to pay and eliminate the labor costs associated with hiring doctors to provide individualized reviews for claims, according to the claim in the class action lawsuit that Joanne Barrows and Susan Hagood filed against the insurer. The program was developed by naviHealth, whose parent company is UnitedHealth Group.
“Humana also utilizes the nH Predict AI Model to aggressively deny coverage because they know they will not be held accountable for wrongful denials,” the claim states.
Another class action suit, filed in Minnesota against UnitedHealthcare, referred to the insurance giant’s use of its AI model to deny claims a “fraudulent scheme” that “affords Defendants a clear financial windfall in the form of policy premiums 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.”
Lawmakers have gotten increasingly concerned over the past year about the use of algorithms to deny care. In October, Democrats in the U.S. Senate highlighted the practice in a report that criticized Medicare Advantage, an alternative to Medicare in which private companies partner with the government to oversee health plans. The practice of health insurance companies denying health care based on AI algorithms was also the subject of an investigation by STAT News, which last year reported that UnitedHealthcare executives had pressured the company’s medical staff to cut off payments for seriously ill patients based on the calculations of a computer algorithm.
The October report by the U.S. Senate Permanent Subcommittee on Investigations cites the STAT investigation and singles out NaviHealth for using artificial intelligence to “fix lengths of stay for patients at various inpatient facilities, and in some cases to determine whether those patients should be admitted at all.” It recommended that the Centers for Medicare and Medicaid Services expand regulations to ensure that predictive technologies do not have “undue influence” on human reviewers.
“Even if predictive technologies are solely used to approve requests, nurses and doctors reviewing cases may face pressure to rubber-stamp the recommendations of algorithms and artificial intelligence,” the report states.
Ensuring that physicians have a say in denial of care is also the goal of Becker’s bill. During a June hearing on SB 1120, Becker said that while artificial intelligence tools have the capability to improve access to care, they face criticisms for “inaccuracies and biases, prompting calls for oversight.”
“Lawsuits against companies allege they’ve used AI to wrongfully deny essential care to hundreds and thousands of patients,” Becker told the Assembly Standing Committee on Health.
