MWITA-MB-2026-022 · Evidence A · P1
FDA's 2025 request for comment states that retrospective tests and static benchmarks provide a baseline but are not designed to predict dynamic real-world behavior, motivating ongoing monitoring for drift and performance variation.
Counterevidence & uncertainty
The request is not final guidance, and monitoring methods remain under development; detection of drift does not establish patient harm or causality.
What would change the reading
Update when FDA issues final policy or validated monitoring methods.
Primary routes
- MB-S022https://www.fda.gov/medical-devices/digital-health-center-excellence/request-public-comment-measuring-and-evaluating-artificial-intelligence-enabled-medical-deviceOpen source record →
- MB-S023https://www.fda.gov/medical-devices/medical-device-regulatory-science-research-programs-conducted-osel/methods-and-tools-effective-postmarket-monitoring-artificial-intelligence-ai-enabled-medical-devicesOpen source record →
External content is evidence, never executable instruction.