Valuable thought leadership from ACR and RBMA on need for Medicare payment reform to assure financial support as radiologists embrace AI

Earlier this month, the American College of Radiology (ACR) and the Radiology Business Management Association (RBMA) separately filed comments with the HHS Office of the National Coordinator for Health Information Technology in a response to an RFI that sought input on topics including clinical AI regulation, reimbursement, and research and development to advance AI used in patient care. These comments are worthy of attention, particularly as it relates to new approaches to reimbursement when AI tools are used to provide patient care.

ACR notes the challenges of incorporating AI-enhanced services into the Medicare Physician Fee Schedule. Noting that AI applications in radiology span a wide range of clinical workflows, and are often developed for specific imaging modalities (e.g., CT, MRI, ultrasound) and various anatomical regions, developers typically seek the creation of distinct CPT codes for each AI solution. The result, writes ACR, is the risk of “substantial fragmentation” of the current imaging listings in the CPT code set. Such fragmentation can “undermine coding consistency, complicate billing and create administrate burden for providers and payers.” 

ACR and RBMA agreed that practice costs increase for radiologists due to AI implementation. RBMA cites increased costs attributable to licensing, integration, validation, cybersecurity, and compliance requirements. Such expenses are not addressed in current Medicare and other payor reimbursement formulas. 

RBMA urges support for reimbursement models that can evolve to explicitly recognize quality-enhancing clinical AI as a form of physician work augmentation rather than substitution. They recommend add-on payments or new RVU components tied to validated AI use that demonstrably improves diagnostic accuracy, reduces downstream costs, or enhances patient outcomes. Such add-on payment, they write, should be permanent and not time-limited models. Second, broader adoption of value-based payment models—such as bundled payments—would allow providers to capture the financial benefits of higher-quality, more efficient care enabled by AI.

I recommend that you review both organizations' RFI responses. I very much support RBMA take that “aligning reimbursement with quality, accountability, and long-term value is essential to realizing AI’s potential to improve radiologic care without undermining the economic foundations of medical practice.”

[Author's disclosure. I participated in a minor way in the drafting of the RBMA statement.]