WISeR AI rollout marred by technical failures and limited communication, CMS documents reveal

As part of its ongoing lawsuit against the Centers for Medicare & Medicaid Services (CMS), a major digital rights nonprofit has obtained new records pertaining to an artificial intelligence pilot program the government authorized to effectively add prior authorizations to traditional Medicare.

The Electronic Frontier Foundation (EFF), which filed its lawsuit back in March seeking more details on CMS’ Wasteful and Inappropriate Service Reduction (WISeR) model, said it has evidence showing the AI has resulted in patient care delays, causing chaos by denying care and creating friction with providers.

EFF released those records related to the impact of WISeR back in September. This week, it released 800 more pages, writing in a blog that the documents provide insights into how the AI pilot works.

Details include applications from AI tech vendors seeking to be involved with the pilot—including the six selected companies—along with “task orders and contracts” meant to guide the deployment of the program, including ongoing monitoring and success metrics.

Related to the previous release about patient care troubles, EFF also obtained requests for corrective action sent by CMS to four of the six WISeR participants, along with their responses.

In summarizing the findings, EFF said the documents show that the rollout of WISeR was “rushed and rocky,” exemplifying the “dangers that arise” when government agencies place more faith in AI applications than evidence suggests they deserve.

“The documents also highlight the need for more transparency about safeguards against inappropriate, AI-driven [care reimbursement] denials,” the nonprofit wrote. “The public still lacks the information needed to assess whether clinician review is providing meaningful oversight of AI systems and whether appropriate data is being used to evaluate prior authorization requests.”

Public remains in the dark

That lack of information is central to EFF’s legal complaint, which is still in the discovery stages. The group said CMS failed to respond to its Freedom of Information Act (FOIA) request for more details on WISeR, including how it selected participating vendors, how patient data was being protected, and what the agency was doing to police issues related to algorithmic bias and wrongful care denials.

EFF raised the question of how these AI companies can deploy their technologies for public use if the underlying code is not revealed to the public. According to the nonprofit, the companies participating in the WISeR pilot are compensated based on the “associated savings” they generate from denying medical claims, giving them an incentive to do exactly that.

WISeR rolled out in six states—New Jersey, Ohio, Oklahoma, Texas, Arizona and Washington—beginning in January. The vendors selected for the pilot include Cohere Health, Genzeon Corporation, Humata Health, Innovaccer, Virtix Health and Zyter.

“By outsourcing decision-making to private AI companies without sufficient transparency, testing, or legal limits, WISeR jeopardizes seniors’ access to medical care,” EFF wrote in its blog. “Congress should step in to place greater legal limits on the use of automated decision-making in Medicare.”

An attempt to do exactly that died in the U.S. Senate in July. An attempt by a group of Democrats to halt the program ultimately failed in a 46-50 vote, falling along party lines. There hasn’t been any revival of a plan to rein in CMS from deploying WISeR since.