How will Medicare pay for imaging AI? Health policy experts offer 3 suggestions
It’s a question on many radiologists’ minds: How will Medicare pay for artificial intelligence? Three health policy experts attempted to take a stab at finding an answer in a new viewpoint published Tuesday in JAMA Internal Medicine [1].
The U.S. Food and Drug Administration has now approved nearly 900 AI-enabled products, with the vast majority in radiology. Yet, the Centers for Medicare & Medicaid Services has only assigned payment for about 10 of these applications, limiting access for patients.
“The existing fee-for-service payment system is not well suited for these services, and payers must balance a desire for innovation and diffusion of high-value AI-enabled clinical services with a concern about their association with spending and potential overuse,” Anna Zink, PhD, with the University of Chicago Booth School of Business, and colleagues wrote May 28.
Zink and colleagues at Harvard and the University of Minnesota discussed the current economics of AI pricing. Artificial intelligence firms are likely to charge “what the market will bear.” But as interest grows, AI firms will only increase prices. In an environment of escalating costs, CMS would potentially use “the separately payable pathway, which is the most generous to clinicians.” However, “this incentivizes AI firms to set a higher price,” the auth0rs noted. Medicare also must grapple with the conundrum of keeping AI costs low while still providing enough reimbursement to fuel innovation.
To navigate these economic challenges, Zink et al. offered three suggestions, in their own words:
- “Bundle AI-enabled services with complementary services. Since the input to many AI-enabled services is a clinical image (e.g., HeartFlow Analysis inputs a coronary computed tomography angiography), the underlying imaging service is a natural bundle. For AI-enabled services without a complementary imaging service, the diagnostic service may serve as a bundle. AI-enabled services that substantially reduce the cost of providing a service will diffuse without added payment, and AI firms can set prices that will support innovation and still be affordable to clinicians.
- When setting a separate fee or transitional add-on payment for an AI-enabled service, use reference pricing as much as possible to combat the market power of AI firms. This approach may reference reimbursement for non-AI-enabled diagnostic services or other AI-enabled services with a similar clinical application. If the AI-enabled services offer quality improvements, CMS should consider encouraging their development and diffusion with an add-on payment.
- Reevaluate CMS fees regularly to reflect changes in the cost-of-service provision, including considering how AI-enabled services are associated with the valuation of other clinical work.”
Read the rest of their analysis at the link below.
Marty Stempniak has covered healthcare since 2012, with his byline appearing in the American Hospital Association's member magazine, Modern Healthcare and McKnight's. Prior to that, he wrote about village government and local business for his hometown newspaper in Oak Park, Illinois. He won a Peter Lisagor and Gold EXCEL awards in 2017 for his coverage of the opioid epidemic.