AI/Machine Learning: Regulation, Development, and Real-World Performance Evaluation

On March 22, 2022 from 12:30 – 4:00 pm (EDT) the Duke-Margolis Center of Health Policy, with participation from the U.S. Food and Drug Administration (FDA) speakers and panelists, will hold a public webinar to spotlight updates and progress made since the release of FDA’s Artificial Intelligence/Machine Learning Action Plan in early 2021.

There will be a fireside chat between Jeff Shuren, FDA’s director of the Center for Devices and Radiological Health (CRDH), and Mark McClellan, director of Duke-Margolis, followed by three panel discussions focused on the following topics:

1. overarching frameworks for regulating and evaluating AI
2. good machine learning practices (GMLP)
3. post-market evaluation of AI/ML SaMD 

The webinar will end with a business roundtable discussion on how developers and health care systems are considering these issues.

This webinar is partially supported by The Gordon and Betty Moore Foundation.

See Related Work

As the FDA considers how manufacturers can use real-world data (RWD) to monitor the performance of software tools built with artificial intelligence (AI) in health care settings, a new report from Duke-Margolis considers both the potential and the hurdles, given how important is it is to be able to efficiently test how generalizable these tools are and if their accuracy holds up over time. The report, with support from the Gordon and Betty Moore Foundation, provides recommendations for how the evolving regulatory model can facilitate the use of RWD to evaluate software-based medical devices, specifically addressing what type of data would be needed, data access and privacy, and data sharing and security.