Hot off the presses! The FDA has published a Draft Guidance on the Change Control Plan for AI/ML-enabled Medical Devices. Ultimately, the plan must include "a detailed Description of Modifications, a Modification Protocol, and an Impact Assessment . . . ." Since this is the first official guidance on AI/ML, it covers more than just defining what is in the PCCP. For example, the guidance defines important terms (particularly regarding data sets and testing) and gives a glimpse into the FDA's expectations for AI/ML-enabled devices for data management, algorithm training/re-training, and performance evaluation. The guidance includes detailed examples that should help Digital Health companies better understand the expectations. Keep in mind that this is not just about new content for a Regulatory Submission but impacts your Quality System and Risk Management activities.

There's a lot of detail to consider in this guidance. Don't hesitate to reach out if you want to discuss. As this is a draft guidance, I encourage everyone to submit comments to the public docket (FDA-2022-D-2628). As we saw with the CDS Guidance, you can influence the policy but only if you submit comments. Brooke & Associates can help!