In 2021, the U.S. Food and Drug Administration (FDA), Health Canada, and the UK’s Medicines and Healthcare products Regulatory Agency (MHRA) collaboratively identified ten guiding principles for good machine learning practice (GMLP). These principles aim to ensure the safe, effective, and high-quality development of AI/ML technologies in medical devices.
Building on GMLP, these agencies are now developing a proposed regulatory framework for modifications to Artificial Intelligence / Machine Learning (AI/ML) – Based Software as a Medical Device (SaMD) including setting out additional guiding principles specifically for transparency in machine learning-enabled medical devices (MLMDs).
Transparency in MLMDs is crucial for several reasons. It ensures that essential information about an MLMD—its intended use, development, performance, and, when available, the logic behind its outputs—is clearly communicated to relevant audiences. This encompasses healthcare professionals, patients, caregivers, and other stakeholders involved in healthcare decisions. Effective transparency promotes patient-centered care, facilitates informed decision-making, and supports the safe and effective use of these advanced technologies.
The guiding principles the agencies propose for transparency address key aspects such as:
- Audience (who) – Communicating information to the right audience such as those who use the device (HCPs, patients, carers), those who receive healthcare with the device (patients), and additional parties (those who make decisions about the device (administrators, payers, governing bodies)
- Motivation (why) – Communicating information to identify and evaluate the device’s risks and benefits and to help ensure the device is used safely and effectively
- Content (what) – Communicating information that enhances understanding of the device and its intended use (e.g. medical purpose and function, intended users etc)
- Placement (where) – Where to display elements of the device the user interacts with (those the user sees, hears and touches). For example, training, physical controls, display elements, packaging, labelling, and alarms are part of the user interface
- Timing (when) – Communicating information needed throughout each stage of the total product lifecycle can support successful transparency. Detailed device information may be needed when considering whether to acquire or implement a device, and whether and how to use it
- Methods (how) – Communicating information about MLMDs requires a holistic understanding of users, environments and workflows. You can address these by applying human-centered design principles
For instance, transparency involves providing detailed information about an MLMD’s medical purpose, target population, performance, and potential risks. It also includes explaining the device’s integration into healthcare workflows and the logic behind its outputs to help users critically assess its reliability and safety. Human-centred design is fundamental to achieving effective transparency. This approach considers the entire user experience and involves relevant stakeholders throughout the design and development process. By employing human-centered design principles, developers can ensure that MLMDs are transparent and that users have all the necessary information to use these devices safely and effectively.
These guiding principles for transparency in MLMDs are designed to support safe, effective, and patient-centered use of AI/ML technologies in healthcare. They emphasize the importance of clear communication, timely updates, and comprehensive information, fostering trust and confidence in these innovative medical devices. Continued engagement and collaboration among stakeholders will be essential to refining and implementing these transparency practices in the rapidly evolving field of medical technology.
For the MHRA’s full announcement please see here and for the FDA’s docket on the proposals please see here; the regulators are welcoming feedback and engagement on these efforts to help inform collaborate development in this rapidly evolving field.
Fur further information, please contact Julia Gillert, Jaspreet Takhar or Elina Angeloudi.
