On December 4, 2024, Brian Thompson, CEO of medical insurance giant UnitedHealthcare was shot dead in New York. After the incident, online discourse celebrating Thompson’s death underlined the resentment that many feel towards health insurance companies, who are arbiters of healthcare services in an already imperfect system.
An Ethnic Media Services briefing on December 20 shed light on another controversial feature of the American health insurance industry: the increasing use of AI algorithms to review and evaluate insurance claims. In such cases, medical practitioners are not a part of the decision-making loop, and the rejections of claims can lead to detrimental outcomes for policyholders, and intensify inequities within the healthcare system. Throughout the briefing, the panel of experts elaborated on the wider context in which insurance companies turn to AI, the ripple effects these systems have on certain sections of society, and a new law that aims to check this practice in California.
The Stories Big Media Won’t Tell
38 years of authentic voices, fresh perspectives daily. Subscribe now for free and never miss a story that everyone’s talking about! 📰
An industry fraught with mistrust
According to Katherine Hempstead, Senior Policy Advisor at the Robert Wood Johnson Foundation, insurance has always been fraught with mistrust between insurers and policyholders.
“Insurance companies are gatekeepers in some sense,” she said. “Health insurance is worse because people’s lives … are at stake…”
The opaqueness of insurance legalese, geographical variations in coverage, administrative hurdles in accessing healthcare, and increasing rate of insurance claims denial further breeds dissatisfaction among policyholders.
Hempstead explained that the number of uninsured people is steadily decreasing, which is a good sign as it means more people have some form of healthcare coverage. However, she also mentioned that there are many new treatments and drugs flooding the market, many of which insurance companies do not cover. An example is the popularity of GLP-1 drugs like Ozempic, which many insurers are no longer covering.
“We spend one out of every $6 in the US on health care, and in a lot of ways, we don’t have the best bang for our buck,” said Dr. Miranda Yaber, an Assistant Professor of Health Policy and Management at the University of Pittsburgh. She explained that the problem of overprescription and the general perception that more care is better care – which is not always the case – also leads to denial of claims.
“So what we end up doing is having these insurance companies imposing barriers on what our physicians can prescribe us,” she said. “Sometimes that’s not a bad thing… but sometimes a lot of these denials end up actually being wrongful and really destabilizing.”
The Stories Big Media Won’t Tell
38 years of authentic voices, fresh perspectives daily. Subscribe now for free and never miss a story that everyone’s talking about! 📰
AI and its ripple effects
In such a situation, insurance companies are turning to AI systems to evaluate and decide claims, in a bid to save costs. Strict adherence to these algorithms and the dilution of medical experts’ role in the decision-making process lead to wrongful denial of essential services. As a result, Insurance giants United Healthcare and Cigna have had lawsuits brought against them. The lawsuits argued that 90% of the denials were reversed when appealed, which suggests that these were wrongful denials, to begin with. This is not surprising given that the algorithms do not have as much information to go by, as a trained doctor or healthcare professional.
“When people appeal, and the press is involved, the decision changes, which can drive a lot of cynicism, because [that shows] that insurance companies are trying to get away with something,” said Hempstead.
A coverage denial can be especially detrimental to policyholders who belong to lower-income and lower-educated communities like immigrants. Professor Yaver is currently working on a book about this subject, titled Coverage Denied: How Health Insurers Drive Inequality in the United States. Of the people she surveyed, 36% reported being denied medical coverage, and 60% of that group reported facing multiple denials. She found that while everyone is vulnerable to coverage denials they cause more disruption among marginalized groups, which deepens the inequities in the American healthcare system.
For instance, she pointed out that while the average American adult reads at about an eighth-grade reading level, written healthcare materials are typically at an 11th or 12th-grade reading level.
“The learning costs of figuring out your eligibility to appeal a denial, your compliance costs associated with figuring out how an appeal works and how to do it successfully, the psychological cost, the emotional burden, all of these things just get amplified for people who are scoring lower on the health literacy scale,” she said.
Yaver argued that while efficiency gains of AI can help in the healthcare industry, but the gains needs to be weighed against the devastating impact that wrongful denials have on policyholders, especially those from marginalized backgrounds. A new California law that came into effect this year hopes to eliminate this issue.
“Physicians Make Decisions“
In September 2024, Governor Gavin Newsom signed California State Senator Josh Becker’s bill SB 1120 into law. This bill – also called the Physicians Make Decisions Act – prohibits insurance companies from making decisions on claims solely based on algorithms. Licensed healthcare providers need to oversee decisions made by AI algorithms and health insurance processes such as claims processing and prior authorization requests. The California Medical Association, comprising 50,000 physicians across the state, sponsored the bill and helped develop it.
“We need the human element, and this bill responds to concerns by ensuring that health care decisions prioritize patient well-being over automated processes,” said Senator Becker.
To illustrate the scale of the issue, he quoted some sobering numbers that came out of the lawsuit against Cigna. According to corporate documents and interviews with senior officials at the firm, senior doctors denied over 300,000 requests for payment over two months. This translates to an average of 1.2 seconds spent on each claim, which clearly suggests that the companies are relying on algorithms to make the decisions.
But if insurance companies roll back the use of AI algorithms and spend more money on paying doctors to evaluate each and every claim, will this move force the companies to raise premiums to protect their bottom line? Senator Becker thinks not; in 2023, the top seven insurance companies earned a profit of around $70 billion.
“It seems from the $70 billion in profits, they’re doing a pretty good job of cost containment right now,” he said. “We’re saying that we trust physicians, and so let’s give them more say!”