Medicare’s AI Pilot Program Faces Delays and Criticism
A new Medicare program using artificial intelligence for reviewing prior authorization requests is causing significant delays in patient care and raising concerns about inadequate testing, as revealed by government records. The Electronic Frontier Foundation (EFF), through a Freedom of Information Act (FOIA) lawsuit, uncovered details about the Wasteful and Inappropriate Service Reduction (WISeR) model, a Medicare initiative beginning in January that leverages AI-assisted reviews for specific medical procedures.
The EFF has sued the Centers for Medicare & Medicaid Services (CMS) to compel the release of records about the use of AI in Medicare. Michael Ryan, a finance expert, noted that the problem lies not only with the algorithm but with how AI was integrated into prior authorization processes without thorough testing. This situation can turn the intended efficiency into a service bottleneck, delaying treatment significantly.
Program Details
The WISeR initiative represents one of the federal government’s major steps toward incorporating AI in Medicare administration. The pilot operates in six states: Texas, Arizona, New Jersey, Ohio, Oklahoma, and Washington. It involves 13 medical services identified as prone to fraud, waste, or improper use. Providers must gain authorization through tech platforms managed by private contractors using AI to help assess requests.
Despite CMS’s pledge for a 72-hour response time, EFF’s findings indicate widespread delays, with some requests going unanswered for extended periods, including one instance of an 83-day wait. Reports also mention inappropriate denials and cases of patient harm due to operational problems and system failures.
Ongoing Issues
Kevin Thompson, CEO of 9i Capital Group, expressed concern over the lack of rigorous testing before the program’s launch, resulting in extensive delays. The incentive structure may be contributing to increased denials and prolonged authorization timelines. EFF highlighted systemic issues where providers faced approval delays for necessary procedures and dealt with system outages.
EFF has emphasized that ineffective safeguards could lead to unjustified and discriminatory delays or denials in medical care. Little information is available about the AI systems used by WISeR vendors. Although CMS claims a qualified clinician reviews all denials, AI-generated recommendations can heavily influence these human decisions.
Responding to concerns, CMS has committed to addressing timeliness, accuracy, and transparency within the WISeR model. The agency pledges to actively correct issues to meet these objectives and to work with stakeholders to identify and resolve problems promptly.
Previous Concerns
Past reports echo the current issues, with doctors and patients expressing dissatisfaction. A June investigation by KFF Health News revealed prolonged waits, confusion, and extra travel requirements under the new rules. Medical professionals described the rollout as problematic, and organizations noted the administrative challenges imposed on healthcare providers.
The primary worry remains delayed treatment, with extended approval times affecting when patients access necessary care. Thompson suggests CMS could further expand the WISeR model, presenting it as a cost-saving measure despite the outcomes.
Continued Scrutiny
A central question by EFF is whether CMS properly evaluated the technology before initiating the program. The FOIA request seeks records concerning accuracy, bias testing, and other evaluations of participating vendors. Transparency is crucial, as AI-generated suggestions can affect human reviewers’ decisions, even when clinicians have the final say.
EFF maintains that unresolved questions remain about the AI systems in use. Despite these challenges, Michael Ryan believes the lawsuit’s outcome won’t deter Medicare’s interest in AI. WISeR is set to continue until 2031, with CMS viewing technology-assisted prior authorization as integral to future processes.