Will Your Health Insurance Pay Out? AI Will Decide

In the wake of the boom of artificial intelligence (AI), more healthcare insurance companies are using the technology to speed up the process of evaluating patient medical claims, making it quite likely AI played a hand in determining the outcome of your payout.

Newsweek spoke to experts about AI in the health care industry.

Why It Matters

Controlling the use of AI in the healthcare industry has been on the legislative agenda in a number of states in recent months, as critics argue that the technology increases inaccuracies and bias in medical decision-making, warning that patient health care is not being fairly reviewed.

It comes as part of the wave of scrutiny against health insurance companies following the assassination of United Healthcare CEO Brian Thompson, as some subsequently accused the health providers of treating customers unfairly.

What To Know

Explaining specifically how AI can process a medical claim, University of Pennsylvania Professor Hamsa Bastani told Newsweek that "when a claim comes in, an algorithm can review details like medical codes, patient history, and patterns of past claims, to see whether the claim is valid, consistent with policy coverage."

She said that the payout "may be automated if the claim looks normal," otherwise, it may get passed on for manual review by a human.

"It's highly likely that AI is used in some part of the process, but hard to say if the decision is fully automated or still involves some element of human review," Professor Bastani said.

AI, when working accurately, can find quicker resolutions to claims, while also having huge cost-saving impacts for health insurance companies, according to a report by law firm McKinsey & Company.

The law firm estimated that for every $10 billion in revenues, health insurers could save $150 million to $300 million in administrative costs and $380 million to $970 million in medical costs, and additionally that these technologies could help generate between $260 million to $1,240 million in additional revenue.

So, the use of AI can be used as a tool to help speed up the administrative process and cost-save for health insurers; however, using the technology in claims processing requires insurance claim evaluators to have a confident understanding of the system, Georgetown University Professor Will Fleisher told Newsweek.

He added that the AI used in healthcare can also be black box systems "that no one fully understands." An artificial intelligence system is referred to as a black box if its internal workings are not visible or understandable to users.

This could "make it harder for medical evaluators to integrate the information in a reliable way," Professor Fleisher said, "and they might have a harder time explaining why the decisions are justified, if they don't fully understand how the system operates."

This could pose problems for companies as even when they seek to help patients understand why a claim has been denied, they may struggle to answer.

If the AI system is a black box, it also "gives a company cover for denying coverage," Professor Fleisher told Newsweek, as "they can explain the denial by blaming it on the computer."

Concern about the use of AI in claims processing has been growing as while the use of the technology has increased, so has the number of claim denials in the U.S.

According to a survey cited by the statistics website advertisement.com last month, around 30 percent of doctors said that they had seen an increase in denied claims over the past year.

Major health insurance companies including UnitedHealth, Humana and Cigna are also facing class-action lawsuits alleging they relied on algorithms to deny lifesaving care.

One of the lawsuits alleges that Cigna denied over 300,000 claims in two months, meaning physicians spent an average of 1.2 seconds on each case.

What People Are Saying

Georgetown University Professor Will Fleisher told Newsweek: "Various kinds of AI could be used to make decisions about who should receive what payouts. They might be used to replace in-house doctors to determine whether a treatment is necessary based on a set of criteria. Or they might use AI to fill out forms automatically to speed up the work process."

University of Pennsylvania Professor Hamsa Bastani told Newsweek: "While the efficiency gains in claim review are obvious, there are many concerns that legitimate claims can be denied, slowing access to necessary care. Worse, the algorithm may systematically make errors and deny claims for vulnerable populations or for rare but serious medical conditions. This can cause significant human harm."

What's Next

A number of states are continuing to push for AI to have tighter controls when used in the healthcare industry. Earlier this month, Texas Representative David Spiller introduced a bill that would prohibit health insurance companies from using AI to delay, deny, or modify claims, legislation similar to what other states have brought forward.