Machine Learning-Enabled Medical Devices (MLMD) represent a revolution in healthcare. These Medical Devices, based on artificial intelligence and machine learning, assist healthcare professionals in multiple tasks, including diagnosis, monitoring and treatment of patients. With the ability to analyse huge amounts of data and to ‘learn’ from previous results, MLMDs constantly improve their performance. Among the best-known examples of use are imaging tools, devices for continuously monitoring patients even remotely, and clinical decision support systems.
However, they present tough challenges related to security, reliability, transparency and need targeted regulation.
Transparency for Machine Learning-Enabled Medical Devices: Guiding Principles
In 2021, the major international regulatory authorities – the US FDA, Health Canada and the UK’s MHRA – outlined ten guiding principles for Good Machine Learning Practice (GMLP), to ensure the safe development of MLMDs.
Based on the same good practices, the FDA, Health Canada and MHRA published a new guideline in June 2024: ‘Transparency for Machine Learning-Enabled Medical Devices: Guiding Principles‘, to promote international harmonisation and emphasise the importance of transparency throughout the life cycle of Medical Devices. Transparency, according to the new guidelines, means providing complete and clear information about the use of the device, including its limitations, benefits and associated risks. This is crucial to improve doctors’ and patients’ confidence in the reliability of such devices and to ensure informed clinical decisions.
The Role of Transparency in Machine Learning-Enabled Medical Devices
Transparency is one of the key concepts to promote trust and security in the use of MLMD. But what exactly does transparency mean in this context? It is not only a matter of providing information on how the device works but also of explaining how a certain result is achieved. This principle, known as ‘explainability’, is essential to ensure that doctors and patients understand the decision logic of the device, i.e. how the results may influence treatment.
The new guidelines place particular emphasis on the need for users, including healthcare professionals, patients and other stakeholders, to have access to clear and detailed information at all stages of the device lifecycle.
The Six Principles of Transparency for MLMDs
To ensure adequate transparency, the guidelines outline six basic principles that the manufacturer must follow.
- Who (Relevant Audience): transparency concerns different categories of users, including health professionals, patients and clinical decision-makers. Each group needs access to clear and customised information on the device.
- Why (Rationale): it is crucial to explain the rationale for using the device to promote safety and efficacy, as well as to identify potential errors or biases in learning models.
- What (Relevant Information): Information should include the purpose, functionality and limitations of the device, with a clear description of how it fits into the clinical workflow.
- Where (Positioning of Information): information must be easily accessible through the user interface of the device. This must be designed to be responsive and adaptable to users’ needs, using different channels such as visual, audio or text notifications.
- When (Timing of Communication): transparency must be maintained at all stages of the life cycle of the device. It is important to provide timely updates when the device changes or new findings occur, especially at critical moments in the workflow.
- How (Methods to Support Transparency): Human-centred design is essential to ensure that information is useful and understandable. Iterative design, continuous validation and customised communication are key tools to support transparency in MLMDs.
Duty to Inform About Benefits, Risks and Limitations
The guidelines require manufacturers to describe in detail the benefits and risks associated with the use of MLMDs. Manufacturers must clarify the clinical limitations of the device, such as information gaps or contraindications, to avoid misunderstandings or inappropriate use.
Manufacturers must also be transparent about the possible errors and biases that the device might generate and how these risks are managed and reduced.
Conclusions and Implications for Manufacturers
In summary, manufacturers must ensure that key information on MLMDs is communicated in a clear and accessible manner to all users involved. They must provide transparent explanations of the device’s capabilities, limitations and operation.
The new guidelines offer medical device manufacturers an important opportunity to strengthen user confidence and competitiveness in the global market. However, it is crucial to act early and with a strategic approach, given the complexity of the regulatory challenges associated with the sale of MLMD devices based on machine learning systems.
>>> With strategic-regulatory consulting services, Thema and US-based Thema Corp can support you at all stages of the artificial intelligence and machine learning-enabled medical device lifecycle to achieve and demonstrate full regulatory compliance.
To learn more about Medical Devices using artificial intelligence, request our Whitepaper.
