Good Machine Learning Practice for medical devices

2 minute read  22.07.2024 Vanessa Mellis and Elizabeth-Erin Valassakis

The International Medical Device Regulators Forum is consulting on draft Guiding principles on good machine learning practice for medical device development.

Key takeouts

  • AI technologies, including machine learning, have the potential to transform healthcare. Consultation on the draft Guiding Principles is designed to shape a common set of principles to promote the development of safe, effective and high-quality medical devices that incorporate AI.
  • This is a call to action to help further advance good machine learning practice (GMLP). Consensus on the principles is desirable to help inform international standards and regulatory policies and guidelines across the globe.
  • The consultation period closes on 30 August 2024.

On 1 July 2024, the International Medical Device Regulators Forum (IMDRF) published its draft guiding principles on Good Machine Learning Practice (GMLP) for medical device development for public consultation. The IMDRF includes regulatory agencies such as the Therapeutic Goods Administration (TGA) (Australia), Health Canada, US Food and Drug Administration, Medicines and Healthcare products Regulatory Agency (United Kingdom).

The consultation aims to establish a common set of principles to promote the development of safe, effective, and high-quality medical devices that incorporate artificial intelligence (AI). In particular, the IMDRF seeks to use these guiding principles to develop consensus standards that can be adopted by regulatory agencies and research bodies to establish an international set of standards. The IMDRF has also signalled that guiding principles from other non-healthcare related sectors may be considered and incorporated in the ultimate set of principles.

10 guiding principles are outlined in some detail and distilled to:

  1. The device's intended use/intended purpose is well understood, and multi-disciplinary expertise is leveraged throughout the total product life cycle.
  2. Good software engineering, medical device design, and security practices are implemented.
  3. Clinical study participants and datasets are representative of the intended patient population.
  4. Training datasets are independent of test sets.
  5. Selected reference standards are fit-for-purpose.
  6. Model choice and design are tailored to the available data and the intended use/ intended purpose of the device.
  7. Performance is assessed with a focus on the human-AI team in the intended use environment.
  8. Testing demonstrates device performance during clinically relevant conditions.
  9. Users are provided clear, essential information.
  10. Deployed models are monitored for performance and re-training risks are managed.

These draft guiding principles are largely reflective of the guiding principles that were jointly identified by the U.S. Food and Drug Administration, Health Canada and United Kingdom's Medicines and Healthcare products Regulatory Agency, published in October 2021.

In Australia, the TGA monitors the outcomes resulting from IMDRF consultations and uses these insights to inform its regulatory approach.

As such, healthcare, biotechnology, research and AI companies should take this chance to provide their response on the draft guiding principles and contribute to the development of the guiding principles. The consultation will close on 30 August 2024 and the consultation response template can be found on the IMDRF website.


If you need assistance, whether it's engaging with the IMDRF regarding the consultation, learning how the regulatory guidance may impact you, or you simply want to know more about the regulation of medical devices and AI, please contact one of our experts today.

 

Contact

Tags

eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJuYW1laWQiOiJlZDlmMDZjYS02MmYyLTQ1MTEtYWIxYy04ZDk0YTQwMWMwZDIiLCJyb2xlIjoiQXBpVXNlciIsIm5iZiI6MTczNDA4MjgxNywiZXhwIjoxNzM0MDg0MDE3LCJpYXQiOjE3MzQwODI4MTcsImlzcyI6Imh0dHBzOi8vd3d3Lm1pbnRlcmVsbGlzb24uY29tL2FydGljbGVzL2dvb2QtbWFjaGluZS1sZWFybmluZy1wcmFjdGljZS1mb3ItbWVkaWNhbC1kZXZpY2VzIiwiYXVkIjoiaHR0cHM6Ly93d3cubWludGVyZWxsaXNvbi5jb20vYXJ0aWNsZXMvZ29vZC1tYWNoaW5lLWxlYXJuaW5nLXByYWN0aWNlLWZvci1tZWRpY2FsLWRldmljZXMifQ.J3AZjaHyj4zODz8MK2VIEeydjRF54YiQbR50QnZQDe0
https://www.minterellison.com/articles/good-machine-learning-practice-for-medical-devices