European and North American regulators are increasingly demonstrating a willingness to collaborate on safety guidelines, especially in the medical device industry. In 2021, the U.S. Food and Drug Administration (FDA), UK Medicines and Healthcare products Regulatory Agency (MHRA), and Health Canada jointly released the Good Machine Learning Practice for Medical Device Development: Guiding Principles, which outlines ten guiding principles intended to lay the foundation for developing Good Machine Learning Practice (GMLP) in medical devices.
Latest updates
In June 2024, the three regulators released additional guiding principles related to transparency for machine learning-enabled medical devices (MLMDs), which build on two of the existing GMLP principles: Principle 7, which states that “focus is placed on the performance of the human-AI team,” and principle 9, which states that “users are provided clear, essential information.” For the purpose of these principles, transparency describes “the degree to which appropriate information about a MLMD is clearly communicated to relevant audiences.”
According to the FDA, effective transparency ensures that information that could impact risks and patient outcomes is communicated; considers the context for the information; delivers the information using the optimal format, timing and strategies; and employs a “holistic understanding” of factors such as users, environments and workflows.
The latest guiding principles for transparency of MLMDs advocate for an approach that considers who (relevant audiences), why (motivation), what (relevant information), where (placement of information), when (timing of communication), and how (methods used to support transparency). They also place a heavy focus on human-centered design.
Given the critical nature of medical devices to a user’s health and the growing complexity of devices equipped with machine learning or AI, it is more important than ever that users and healthcare providers understand the intended use, development, and performance of MLMDs and the related risks.
Big picture
The guiding principles from the FDA, MHRA, and Health Canada are not binding regulations; however, they are intended to lay the foundation for GMLP principles from the International Medical Device Regulators Forum (IMDRF) and other international standards organizations. The IMDRF convened the Artificial Intelligence/Machine Learning-Enabled (AI/ML) Working Group in summer 2023. The group is tasked with developing a new document on GMLP principles “that looks to generate safer AI/ML-enabled medical devices and global alignment across the total product lifecycle.” The working group is co-chaired by representatives from the FDA and MHRA, and has established as one of its goals to “build upon the US FDA/Health Canada/UK MHRA joint document on GMLP Guiding Principles.”
Bearing this in mind, medical device manufacturers whose devices are AI/ML-enabled should closely review the collaborative 2021 Guiding Principles and familiarize themselves with any supplemental materials including the new transparency principles. While the IMDRF has yet to establish a timeline for its international standards, manufacturers should take steps now to get ahead of preparations. This should include updating recall and crisis plans for AI/ML-enabled devices and reviewing existing transparency practices for communicating about MLMDs and potential risks with relevant parties.
Trusted by the world’s leading brands, Sedgwick brand protection has managed more than 7,000 of the most time-critical and sensitive product recalls in 100+ countries and 50+ languages, over 30 years. To find out more about our product recall and remediation solutions, visit our website here.
The post Health regulators team up to create transparency guidance for machine learning in medical devices appeared first on Sedgwick.