Medicare's technology-assisted WISeR prior-authorization pilot is facing renewed scrutiny after federal records documented long waits, technical failures and high denial counts during its first year. Ars Technica reports that provider complaints and internal status reports describe patients waiting for decisions while vendors struggled to get their systems working.
The Electronic Frontier Foundation obtained roughly 1,000 pages of CMS contracts, status reports and provider feedback through a public-records lawsuit. EFF says one request remained unanswered for 83 days and two vendors denied 5,944 requests in the program's first three months. Provider feedback in the release described delayed surgeries and patients waiting in pain.
Denial data and a rocky rollout
One March 30 report showed Virtix, the vendor assigned to Washington state, denied 3,233 of 6,096 prior-authorization requests, or 53%. Ars says CMS later placed Virtix on a corrective action plan for missing the program's 72-hour standard. Virtix told the publication that the plan ended Aug. 14 and that its average turnaround had fallen to about 1.18 days.
The records also show that Innovaccer warned before launch that its full rules-based system was not ready. It initially auto-affirmed requests to avoid building a backlog while development and testing continued. Another vendor, Zyter, had months of data discrepancies; its CEO told Ars the system is now fully functional across Medicare Parts A and B.
CMS pairs AI assistance with clinician review
CMS describes WISeR as a six-year test running through 2031 in Arizona, New Jersey, Ohio, Oklahoma, Texas and Washington. It applies to selected Original Medicare services with existing coverage criteria, including skin and tissue substitutes, electrical nerve stimulators and knee arthroscopy for osteoarthritis. Emergency and inpatient-only services are excluded.
CMS says licensed clinicians make all recommendations for non-payment. The agency also pays participating companies a share of expenditures associated with care that their reviews avert, adjusted by measures that include decision quality and provider experience. EFF argues that this structure creates pressure to deny requests; CMS says its quality adjustments and review rules are meant to discourage inappropriate denials. A denied request that is overturned on appeal does not earn a vendor payment.
That distinction matters. The program uses AI and machine learning to support the review process, but the official design does not authorize software alone to issue a final non-payment recommendation. The documented delays, denial patterns and vendor incentives are the issue under examination, not proof that every affected request was medically necessary or improperly rejected.
