Moreover, vendors involved in the WISeR model are embracing AI-driven pre-approvals to capture a portion of what CMS refers to as “avoided spend.” This strategy can lead to revenue generation by denying certain care requests. It raises significant discussions about long-standing concerns over commercial activities that discourage patients from obtaining medically necessary care. Several members of parliament have introduced resolutions and corrections to block funding for the WISeR model due to concerns regarding patient access.
During the Trump administration, prior approval faced scrutiny. As CMS enhances its use of AI within Medicare, the agency’s goal is to minimize and streamline prior approval processes utilized by private insurers, including Medicare Advantage Plans. CMS Administrator Mehmet Oz cautioned insurance executives that failure to reduce prior authorization burdens could prompt federal regulation, as discussed on the television news program National News Desk here.
In an effort to preempt executive action or legislative measures, health plans have recently released data indicating compliance with administration demands. Industry research shows that from June 2025 to April 2026, 11 percent of pre-approval requests were denied, although the overall rejection rate remains unclear.
In correspondence with industry organizations, a recent investigation revealed that all responding health plans agreed that “AI algorithms not reviewed by clinicians will not be used to deny prior authorization requests based on medical necessity.” Furthermore, insurance companies promised greater transparency in their clinical reasoning for prior authorization decisions.
This may alleviate some concerns about the absence of human review in AI-driven decisions but addressing critics is challenging.
Dr. Jared Daszewski, physician and founder of the media and education platform Healthcare Huddle, expressed that while AI has potential to “remove barriers, reduce administrative waste, and enhance patient interaction,” the current trajectory is flawed. He argues, “Instead of improving outcomes, we’re automating a broken system that requires overhaul.”
This article was initially published on Undark. Read the original article here.
Source: arstechnica.com