Rimsys recently held a panel discussion, Taking SaMDs to market in the US. During it, Prabhu Raghavan, Principal at MDQR Solutions, and Rimsys Chief Solutions Officer, Brad Ryba, shared an overview of SaMDs and provided their insights about getting and maintaining market clearance for them in the United States. Topics ranged from FDA risk classifications and submissions, cybersecurity best practices, and machine learning algorithms, which brought about an important question: How is the FDA currently regulating adaptive machine learning algorithms in SaMDS?
Adaptive machine learning algorithms use post-market data in real time and evolve their models based on the data they're consuming. As such, every patient utilizing a device with adaptive machine learning algorithms may have a new model compared to the previous patient. While the FDA doesn't have any formal guidance on the subject just yet, manufacturers can work with the FDA to get a plan in place for maintaining a state of validation post market.
Watch the snippet from the webinar to learn about taking a staged approach with the FDA to get a proper validation plan in place.
To watch all discussion topics, download the webinar replay here.