The primary concerns with the safe and effective use of AI-enabled medical devices are their trustworthiness, fairness, and performance. The FDA proposes a lifecycle management (LCM) approach to better address these concerns. This week, a series of blog posts by the FDA’s office leading review of software medical devices (SaMDs) describes an AI life cycle (AILC) approach listing expectations from each phase of development and post-market real world evidence with AI-based SaMDs.
AI-based medical devices hold great promise to increase operational efficiency in most aspects of healthcare and treatments. But like any new technology, AI poses potential risks and challenges primarily due to the nascent nature of the technology despite its exponential growth in a very short time. The AILC approach aims to consider potential challenges from the design phase to launch and post-market surveillance. The key new information is a flowchart created by the FDA listing a phase-wise approach to risk assessment, mitigation, and management of new AI SaMDs. The graphic presents 7 distinct phases in the life of an AI medical device with a total of 66 areas of information to be considered. The list is comprehensive although the information provided is brief.
This is just the first presentation on the concept of the lifecycle approach for AI- devices. But it contains sufficient information to get started.
