A Medicare pilot program employing artificial intelligence to support prior authorization reviews has led to notable delays in patient care and concerns over testing adequacy. These insights come from government documents released due to a Freedom of Information Act (FOIA) lawsuit by the Electronic Frontier Foundation (EFF). The documents provide a detailed view of the federal government’s Wasteful and Inappropriate Service Reduction (WISeR) model. Initiated in January, this Medicare project uses AI to review certain medical procedure requests.
EFF’s lawsuit targets the Centers for Medicare & Medicaid Services (CMS), demanding transparency about AI’s role in Medicare. As finance expert Michael Ryan notes, the problem is not solely with the algorithm. He argues that integrating AI into the prior authorization process before comprehensive testing might have created a bottleneck, delaying patient treatment.
The result is a faster way to create a bottleneck. When a request that is supposed to take 72 hours can sit unanswered for weeks, that is no longer administrative efficiency. For the beneficiary, it is delayed treatment.
The WISeR program is a first in using AI in Medicare administration and operates in six states: Texas, Arizona, New Jersey, Ohio, Oklahoma, and Washington. It covers 13 medical services identified by CMS as prone to fraud, waste, or misuse. Providers use tech platforms, managed by contractors employing AI, to secure treatment authorization. EFF-disclosed documents reveal widespread operational issues since the pilot’s launch, including prolonged response times and technical glitches.
EFF highlighted “widespread delays” in authorization responses, with some requests going unanswered for up to 83 days, and reports of inappropriate denials. These issues were attributed to inadequate pre-launch testing, according to Kevin Thompson, CEO of 9i Capital Group.
The systems were not thoroughly tested prior to launch, which is now causing lengthy delays beyond the 72-hour prior authorization turnaround.
EFF emphasizes the significance of safeguarding AI-driven decisions to prevent discriminatory delays or denials of healthcare. Despite CMS requiring human clinician review of denials, AI recommendations can influence human judgment. CMS previously assured updates to the WISeR model according to feedback from medical professionals and patients.
CMS is firmly committed to ensuring timeliness, accuracy, and transparency under the WISeR model, and will take corrective actions as appropriate to ensure these goals are met.
Records reveal echoes of earlier concerns raised by doctors and patients. A June KFF Health News investigation reported patient confusion and additional requirements due to WISeR’s authorization rules. Medical professionals have described the rollout as challenging, leading to forced rapid adaptation to new procedures.
One prominent concern involves delayed treatment. Longer approval times, according to Thompson, may hinder timely patient care. He predicts CMS will persist in promoting the WISeR model as a cost-saving measure despite the challenges.
Among EFF’s central issues is the adequacy of CMS’s technology testing pre-launch. The original FOIA request pursued detailed records related to testing for accuracy, potential bias, and audits regarding the AI systems used by private companies in Medicare decisions.
Records released thus far echo concerns that providers have raised since WISeR launched, including long delays, financial incentives to deny care, and technical problems.
Ryan suggests that the lawsuit doesn’t signal an end to AI in Medicare, noting WISeR’s continuance until 2031. However, he stresses the need for proven safety and efficiency before expanding AI’s role in care authorization.