In the world of commissioning, qualification, and validation for life science companies our world its pretty simple – you make a regulated product, you qualify the facility, the equipment, you validate the manufacturing process, and do change controls as necessary for equipment, facility, or process changes with updates to FDA depending on the magnitude of the change.
But what do you do when the device (software as a Medical Device) has the ability to learn and adapt?
FDA recently published the a discussion paper on a proposed regulatory framework.
As noted in the paper: The traditional paradigm of medical device regulation was not designed for adaptive AI/ML technologies, which have the potential to adapt and optimize device performance in real-time to continuously improve healthcare for patients. The highly iterative, autonomous, and adaptive nature of these tools requires a new, total product lifecycle (TPLC) regulatory approach that facilitates a rapid cycle of product improvement and allows these devices to continually improve while providing effective safeguards.
This type of innovation will make a significant difference in some aspects of health care as we know it. Several years ago (2012) there was a competition to build a functional tricorder, this was an X-prize contest with a 10M award. The purpose was to spur innovation and take the market where it may not have gone on its own. It is exciting to see where AI/ML medical device technology will take us in the future and part of this question is how will these companies validate algorithms that can self-learn and modify their programs.