Newly released records reveal lengthy treatment delays, technical problems, and early warnings about contractor readiness in a controversial Medicare AI-based prior authorization program that began this year. The roughly 1000 pages of documents, obtained by the Electronic Frontier Foundation (EFF) through a Freedom of Information Act lawsuit, provide an inside look at the troubled launch of the Wasteful and Inappropriate Service Reduction (WISeR) model. The Centers for Medicare & Medicaid Services (CMS) designed the pilot program to reduce spending on services for Medicare fee for service patients that the agency considers vulnerable to fraud, waste, and abuse, such as nerve stimulators and steroid injections for pain management. It rolled out on January in six states — Arizona, New Jersey, Ohio, Oklahoma, Texas, and Washington. EFF's analysis of the records show some prior authorization requests exceeded the program's promised 72-hour turnaround, including one request that went unanswered for 83 days. Clinicians reported procedures being postponed or canceled while awaiting approval, and a WISeR vendor warned CMS weeks before the launch that its system was not ready but that it would move forward anyway "given CMS's decision not to delay" the program's start. CMS did not respond to Medscape's request for comment.Under the model, CMS contracts with multiple technology vendors to review prior authorization requests, with one company assigned to each participating state. The contractors use "enhanced technology," including AI, to assess whether requests meet Medicare coverage requirements. They must also have "clinicians with the expertise to conduct medical reviews to validate determinations," the program website explained. Although CMS describes WISeR as voluntary, providers in participating states cannot simply opt out. They can either seek prior authorization through the program or provide the service without it, in which case the claim undergoes a post-service, prepayment medical review. Warnings Before Launch According to the documents, Innovaccer — the vendor assigned to Ohio — told CMS shortly before the go-live date that its system lacked full functionality and had not undergone sufficient end-to-end testing, which "increases the potential for provider abrasion and disruption to beneficiary services."In the letter dated December 12, 2025, Innovaccer said that it would temporarily auto-affirm all prior authorization requests while it completed additional functionality. The vendor said it planned to achieve full system capabilities within 60 days after going live and described the stop-gap measure as the "only path available that avoids creating a backlog of unprocessed prior authorizations and claims." However, the EFF found documents suggesting that some features remained unfinished or untested as of April. Overall, the records reveal "a failure to fully test systems before and after the program launched," Lena Cohen, the staff technologist at EFF who conducted the analysis, told Medscape Medical News. She said this raises questions about why CMS chose not to delay the launch, especially after program participants like Innovaccer expressed system issues.Delays and "Alarmingly High" Care Denials Authorization decisions often exceeded the recommended 72-hour turnaround time. Timeliness data obtained by EFF indicated substantial backlogs at Virtix Health, which serves Washington, and Zyter, which serves Arizona. Of the 1276 direct prior authorization requests the two vendors received in January, about 30% were 6-21 days old. Virtix has since reported improvement, saying its average turnaround time peaked at 8.97 days during the program's initial weeks but fell to 0.51 days by week 11. The records also provide insight into the number of non-affirmations — requests that did not meet Medicare's requirements based on the information submitted — in the months following the launch. Virtix and Zyter issued more than 20,000 non-affirmations combined during the program's first 3 months, and Virtix denied more requests than it affirmed during that period. Cohen said those rates are "alarmingly high" and that providers have reported patients being harmed by widespread authorization delays. Physicians Adjust to WISeR, Gold Cards Helpful The records echo problems reported by physicians earlier this year, including decision and payment issues tied to WISeR. In feedback included in the documents, one provider wrote: "I have had to watch three patients cry at bedside for not hearing back on their prior auth for kyphoplasty…These patients are in deep pain."Arizona interventional pain specialist Matthew Crooks, MD, also criticized the system, calling it "unsustainable" after waiting months to be paid for epidural steroid injections performed under the program, even those with valid authorization numbers. His experience today looks much different. He told Medscape Medical News that he received a WISeR "gold card" in July, a designation that allows physicians who consistently meet Medicare's coverage requirements to bypass the usual prior authorization process. "Once we got the gold card, things got back to running very smoothly, both for getting patients in efficiently for treatments and billing for them," he said. Three of the six pain management physicians in his practice also now have gold cards. Crooks said things have "dramatically improved" even for those who do not have one, with better communication with the state's WISeR contractor and faster response times. Crooks, who is president of the Arizona Society of Interventional Pain Physicians, said his practice has an internal billing operation and can devote considerable resources to navigating the new authorization requirements. "It is very difficult for smaller practices to overcome those types of barriers," he said.Concerns About WISeR Expansion The WISeR pilot is scheduled to run through 2031, and its results could inform whether Medicare adopts the new process in more states. But the records released by EFF show CMS has also considered adding services requiring authorization on a "more urgent or emergent basis," including MRI scans, oncology care, air ambulance transportation, and cardiac catheterization. The model currently excludes inpatient-only care, emergency services, and those posing a "substantial risk to patients if delayed." Crooks cautioned against applying WISeR to more time-sensitive care without first addressing the problems he and other physicians experienced early in the program. "These are potentially life-and-death situations where time is of the utmost importance, so the thought of increased wait times is very, very concerning," he said. "Every measure should be taken to make sure that it is running smoothly before rollout occurs to prevent unnecessary stress to patients and potentially adverse outcomes." For all the details revealed in the records, important questions about WISeR remain unanswered. The documents did not show what data the AI systems were trained on, how they were tested for bias and accuracy, or how often non-affirmations were overturned on appeal, Cohen said. She called for greater transparency. "The public deserves to know how AI systems and private companies are driving decisions about Medicare beneficiaries' access to care," she said.The experts cited in this article reported no relevant financial disclosures. Steph Weber is an award-winning freelance journalist specializing in healthcare and law.
Backlogs, Testing Gaps in Medicare AI Prior Auth Pilot
Full Article
Original Source
Read the full article at Medscape →KhanList aggregates and links to publicly available news content. We do not host full articles from third-party sources. Always verify important information with original sources.