The U.S. Food and Drug Administration (FDA) has released its plan for regulating certain medical devices that utilize either artificial intelligence (AI) or machine learning (ML) in their device software.
Published in mid-January, the FDA’s “Artificial Intelligence/Machine Learning (AI/ML)-Based Software as a Medical Device (SaMD) Action Plan” details the next steps that the agency plans to take to provide the necessary oversite for these advanced medical devices. The plan reportedly takes a holistic approach based on oversight of the entire lifecycle of the total product in order to ensure patient safety.
In brief, the Action Plan outlines five specific actions that the FDA intends to take, as follows:
- Develop the proposed regulatory framework, including guidances on a predetermined change control plan (to better monitor software learning that takes place over time);
- Support the development of good machine learning practices;
- Develop additional methods to evaluate and improve machine learning algorithms;
- Advance real-world performance monitoring pilot programs; and
- Foster a patient-centered approach, including device transparency for users.
Comment on the FDA’s Action Plan for AI/ML-based medical devices can be submitted through the government portal at http://www.regulations.gov (reference Docket FDA-2019-N-1185).
