Healthcare Lawyers, Bioethicists Call for Tighter AI Regulation Under Prior Authorization

AI promises significant efficiency benefits, but its use in the insurance prior authorization process requires broader and more comprehensive oversight by state and federal regulators, three experts argued this week.

Long-term care providers and consumer advocates are increasingly raising alarm about insurers using artificial intelligence to determine whether patients can receive care. Medicare Advantage prior authorization denials and alleged premature terminations of coverage have been the subject of scrutiny and even class action lawsuits over the past year.

Writing an article for Health matters On Monday, legal and ethics experts from the sector joined the call for more government oversight, applauding efforts by some insurers to self-regulate but arguing that unified oversight would be needed to ensure fair use of AI at all levels.

“While partnerships with industry are critical to these efforts, self-regulation is not enough to protect patients, and regulators must develop clearer rules that encourage innovation while also providing safeguards against abuse,” wrote co-authors Carmel Shachar, assistant professor of clinical law at Harvard Law School; Amy Killelea, senior associate in the Health Policy and Management Program at Johns Hopkins University; and Sara Gerke, assistant professor of law at the University of Illinois at Illinois.

‘Dirt in, dirt out’

The authors noted that AI could eliminate up to 75% of the tasks needed to process pre-authorization requests. This significant improvement in efficiency could help both payers and providers save time and money, they explained.

But because AI results are limited by the quality of the data they input, errors or bias in that data can have significant negative consequences. Some algorithms have been shown to discriminate against marginalized communities—for example, by giving sick Black people less additional care than similarly sick white people.

Such biased algorithm results could expose insurers to lawsuits under the Affordable Care Act’s antidiscrimination provisions, the researchers noted. Especially if the AI ​​directly makes patient insurance decisions.

The authors argue that AI-based cost savings could also have an additional side effect of “review proliferation.”

“Because AI software can process claims quickly and is inexpensive to use per claim,” they wrote, “insurers can expand PAs (prior authorizations) and review coverage to lower-cost procedures and treatments. This suggests we may be seeing ‘review creep’ as insurers increasingly use their new tools across an ever-expanding catalog of services and treatments.”

In short, AI’s improved performance could give insurers the ability to process — and reject — more coverage appeals.

Common road map

The researchers noted that previous advocacy pressure and policymaker concerns on the issue have already led to some regulatory action. However, these actions are inconsistent and still leave many open gray areas.

“There remain many regulatory gaps and opportunities for variability across insurance products and states, especially when it comes to regulating more substantive elements of PA,” the authors explained, later adding that “more proactive federal and state oversight of AI used by payers is needed, particularly the use of AI for utilization management and other coverage arrangements.”

The authors point to Colorado as an example that other states could follow. The state has already regulated the use of AI in life insurance and is currently considering how to apply the rules to health insurance.

The California Senate also passed legislation that would require doctors to review AI use as part of prior authorizations.

In April, five federal agencies committed to reviewing AI’s potential civil rights implications—including, most relevant to providers, the Department of Health and Human Services. This joint approach by regulators is an important first step in addressing a complex problem, the authors noted—one that promises to continue to impact senior care as the benefits and pitfalls of AI continue to unfold.