IMDRF publishes AI/ML guiding principles echoing US, UK, Canadian regulators

The International Medical Device Regulators Forum (IMDRF) has published guiding principles for artificial intelligence/good machine learning (AI/ML) practices in medical device development that are identical to a 2021 document published by US, UK, and Canadian regulators.

On 1 July, IMDRF published its 10 guiding principles for stakeholders developing ML medical products developed by its Artificial Intelligence/Machine Learning-enabled Working Group. These principles include ensuring the device's intended use and purpose are well understood and that a range of experts are involved during the product's lifespan to evaluate it, all of which are part of the guiding principles developed by US, UK, and Canadian regulators. (RELATED: Regulators release 10 principles for good machine learning practice, Regulatory Focus 27 October 2021)

“The 10 guiding principles for Good Machine Learning Practice (GMLP) presented in this document are a call to action to international standards organizations, international regulators, and other collaborative bodies to further advance GMLP,” said IMDRF. “Areas of collaboration include research, creating educational tools and resources, international harmonization, and consensus standards, to inform regulatory policies and regulatory guidelines.”

“These guiding principles may be used to adopt practices from other sectors, tailor them to the medical technology and healthcare, and to develop novel practices for this domain,” the group added.

Other principles include ensuring the device is well-designed and takes into consideration the best cybersecurity practices at the time, as well as ensuring the reference standards used to develop the product are fit for purpose.

“Accepted methods for developing a fit-for-purpose reference standard ensure that clinically relevant and well characterized data are collected and that the limitations of the reference are understood,” IMDRF said. “This includes documentation of the rationale of the choice of reference standard and assessment of its suitability to address the intended use environment.”

IMDRF added that available reference standards should be used in developing models and testing that ensure the product can be used across the intended patient population, and those standards should be based on a broad consensus of experts.

Another important principle in the document is that the device performance should consider how it will be used in the clinical setting, or how the human-AI team will likely use the product in its intended use environment. It adds that human factors such as normal user skills, user expertise, user understanding of the model outputs and limitations, and user error should be considered.

While the GMLP guidance echoes the US Food and Drug Administration (FDA), Health Canada, and the UK Medicines and Healthcare products Regulatory Agency (MHRA)’s guiding principles from 2021, it does not include the latest update to those guiding principles from June which incorporates certain transparency elements. It does take into consideration principles such as the performance of the human-AI team, and asks that users are provided clear, essential information about their product, but the update to the 2021 guiding principles goes further in terms of transparency and asks manufacturers to factor in a high level of detail for the intended audience and consider a human-centered design approach. (RELATED: FDA, Health Canada, MHRA release guiding principles on transparency for machine learning medical devices, Regulatory Focus 14 June 2024)

The IMDRF document does, however, state that as AI/ML products evolve, especially with generative AI, it’s important to describe the intended use/ intended purpose of the product and its regulatory status.

IMDRF guiding principles

IMDRF publishes AI/ML guiding principles echoing US, UK, Canadian regulators

Related topics

Recommended content

Welcome to the new RAPS Digital Experience