7 Possible Concerns About Medicare’s AI Prior-Authorization Pilot

The federal government is testing whether artificial intelligence can help Medicare spot unnecessary care, speed up prior authorization, and lower healthcare spending. But newly released records are prompting doubts about whether the experiment is ready to deliver those benefits.

The documents were released after the Electronic Frontier Foundation (EFF) won access through a lawsuit against the Centers for Medicare & Medicaid Services (CMS). EFF, a San Francisco-based nonprofit, focuses on digital privacy and civil liberties in the use of technology.

The records describe early trouble with the federal Wasteful and Inappropriate Services Reduction, or WISeR, model. The program brought in six private technology companies to help review prior authorization requests for certain Medicare services. It is set to run in six states — New Jersey, Ohio, Oklahoma, Texas, Arizona, and Washington — through late 2031. CMS says the model is meant to protect taxpayers by using advanced technology and human clinical review to ensure Medicare payments are timely and appropriate.

According to EFF, the model requires prior authorization for more than a dozen procedures, including steroid injections for pain management, nerve stimulators and treatments, arthroscopic knee surgery, skin substitutes, sleep apnea treatment, and incontinence devices.

EFF’s review, along with the documents themselves, points to several possible problems with CMS’ AI experiment.

A launch that appears to have been rushed

EFF says CMS moved forward with WISeR before vendors were fully prepared. Innovaccer, one of the six participating companies, warned CMS that its provider portal might not be ready in time. The records suggest the readiness problems were broader than a single vendor.

CMS questioned denial rates

In a letter to another vendor, Virtix Health, CMS raised concerns about how often requests were being denied. Virtix rejected 53% of more than 6,000 prior authorization requests it handled over three months, which the company blamed on difficulties processing a surge in claims. That figure does not prove the denials were incorrect, but it does raise questions about whether coverage rules were being applied properly, especially under heavy volume.

Delays could interfere with care

CMS also flagged vendors for missing turnaround targets. One prior authorization request sent to Genzeon reportedly went unanswered for 83 days. Delays like that can leave patients waiting for care that may relieve pain or prevent a condition from getting worse.

Some technology was not fully ready

The records show CMS was worried that Innovaccer did not have a fully working provider portal for WISeR providers and suppliers. Even after the company warned that parts of its software might not be ready, CMS kept the launch schedule in place. That raises concerns about the technology supporting a program that affects access to medical treatment.

Providers had trouble getting answers

In a letter to Zyter, CMS said the company may not have had adequate procedures for responding to provider concerns. If providers cannot get clear answers about authorization requests, the process can become even more burdensome.

The payment structure raises incentive questions

WISeR vendors can receive a share of spending tied to care that was judged wasteful and then blocked through their reviews. CMS says pay is adjusted based on accuracy, but the arrangement still raises concerns that vendors could be encouraged to deny care inappropriately.

“Human in the loop” may not be enough

CMS requires a licensed clinician to review every denial recommendation. Still, more transparency about how that safeguard works could help determine whether it truly protects patients. EFF said the records offer limited insight into whether human reviewers have real power to change AI-assisted recommendations.

WISeR responds to a real issue: unnecessary procedures can expose patients to complications while driving up Medicare costs. CMS had hoped AI could streamline reviews and reduce those expenses. But the early records suggest the program may have been rolled out too quickly.

EFF concluded:

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