The Trump administration's WISeR pilot program, launched in January, uses AI to make prior-authorization decisions for certain Medicare services in six states. The program was intended to reduce fraud, waste, and abuse, but early reports and federal documents obtained by the Electronic Frontier Foundation paint a different picture: technical failures, delayed decisions, and denials that have left patients in pain and providers frustrated.
One vendor, Innovaccer, was so unprepared that it asked for a delay and temporarily set its system to auto-approve all requests. Another vendor, Zyter, had data discrepancies for months because it confused Medicare Part A and Part B. A third, Virtix, denied 53 percent of the 6,096 requests it reviewed in a weekly report from March 30, and at least one request was still pending after 83 days.
The Government Accountability Office determined in May that Trump officials did not follow proper procedure in setting up the program, calling its legality into question. Despite that, the program appears to be moving ahead, with plans to expand. Last week, Republicans voted down a Democratic effort to force the administration to release more documents about WISeR, and Rep. Suzan DelBene said the administration is trying to conceal information because public outrage would grow if more were known.
Virtix, which was put on a corrective action plan for missing the 72-hour decision window, says it has since improved its turnaround times and that the plan ended on August 14. The company also noted that it uses CMS-established coverage determinations to evaluate requests. But for many providers, the experience has been deeply damaging; one called the program "a disgrace to the human race" and reported seeing patients crying in pain while waiting for care.