Monitoring, reimbursement and transparency—ACR offers White House recs for AI regulation
The American College of Radiology (ACR) recently provided the White House Office of Science and Technology Policy its recommendations on how to best monitor the integration of artificial intelligence (AI) technologies in clinical settings.
In an October 30 news release, ACR shared its response to the White House’s request for information (RFI), highlighting the organization’s main areas of concern—the need for oversight, appropriate payment strategies and government transparency. ACR’s input is included as part of America’s AI Action Plan, which was created to help responsibly implement emerging AI technologies into everyday workflows.
Focusing first on oversight, ACR highlighted the role of the U.S. Food and Drug Administration (FDA) in monitoring the safety and efficacy of continuous learning systems and foundation model-based functions.
“Moving forward, the FDA will need to enhance its oversight mechanisms to address adaptive/continuously learning systems, multi-function device products, foundation model-based devices for broader tasks, and other AI with novel safety and effectiveness considerations,” the response letter states.
Oversight should include real-world monitoring and post-market performance assessments, according to the response. This would ensure model performance is not overcome by AI drift over time. This could be achieved, in part, by establishing qualified national AI registries, such as ACR’s Assess-AI, which assists practices with tracking model performance across various settings and enables streamlined continuous monitoring for both clinical sites and vendors. Registry participation also would address the concerns related to data privacy, as current data sharing requirements are “unclear,” the College noted.
In terms of reimbursement, ACR suggested that the Centers for Medicare and Medicaid Services (CMS) plays a vital role in securing payments for providers who implement clinically meaningful AI. Structured reimbursement parameters would incentivize providers to integrate tools proven to enhance patient care and outcomes.
“Many of the current and future AI tools require clinician validation and/or incorporation of findings into the patient’s care plan. Without a funding mechanism, there is a lack of incentives for widespread adoption,” ACR advised. “Because of the budget neutral structure of this fee schedule (which requires downward to reimbursements for adjustments for other critical physician services when new services are added), these payments for AI cannot come from the current Medicare Physician Fee Schedule allocation.”
The College added that these payments should account for both direct and indirect costs, such as AI governance, infrastructure, monitoring, end-user training and quality and safety initiatives.
ACR also highlighted the need for transparency across HHS agencies to build user trust, further encouraging responsible, effective AI adoption.
For more detailed information on ACR’s recommendations, view the entire RFI response here.
In addition to her background in journalism, Hannah also has patient-facing experience in clinical settings, having spent more than 12 years working as a registered rad tech. She began covering the medical imaging industry for Innovate Healthcare in 2021.