A featured contribution from Leadership Perspectives: a curated forum reserved for leaders nominated by our subscribers and vetted by the Construction Tech Review Advisory Board.
Director Quality Assurance Elekta
Artificial Intelligence regulations and its impact on medical devices
Leo Hovestadt
Introduction
Artificial intelligence (AI) has huge potential to improve healthcare using vast amounts of clinical data. AI software algorithms can use real-world data to support better healthcare decisions or enhance the performance and safety of medical devices. Manufacturers have embedded AI technology into various applications such as imaging, laboratory testing, patient monitoring, personalized apps and robotics.
AI also presents unique challenges due to its complexity and the iterative and data-driven nature of its development. This triggered the development of the European Union AI Act and the development in the US, Canada and the UK of the Good Machine Learning Practices.
European Union AI Act
The European Union AI Act (EU-AIA) is aimed at covering AI in general. The Medical Device Regulation (MDR) structure was used as a blueprint for the EU-AIA. This is very pleasant since it makes the EU-AIA easy to understand. The MDR includes requirements on risk management, conformity assessment by notified bodies, post-market surveillance and a vigilance system. Those elements can also be found in the EU-AIA.
The EU-AIA, however, has substantial changes when compared to MDR. The MDR focuses on treating individual patients with safe and performing medical devices. The EU-AIA focuses on protecting the population from AI solutions. The goals are incompatible and maybe even conflicting, so the substantial changes are logical.
The incompatible requirements are a significant barrier for high-quality AI-based medical devices being placed on the EU market. The EU-AIA changed concepts for risk management and did not define risk. EU-AIA risks for fundamental rights conflict with MDR patient safety risks, potentially making medical devices unsafe. The EU-AIA also defined provider, user, importer and putting into service differently than the MDR.
Medical device manufacturers also often lack direct access to training data because of patients' rights, but authorities in the EU-AIA need to be granted direct access to patient training data. In addition, datasets for the EU-AIA need to be error-free and complete; however, for medical devices, testing data is used to check the quality of AI clinical data.
The list of inconsistencies and duplications is very long, creating legal uncertainty, increasing costs and potentially blocking market access.
Most software medical devices are within the scope of the EU-AIA, not only the AI-based medical devices. The EU-AIA will have an adverse effect on the availability of (AI-based) software medical devices, which cannot be the purpose of the act.
Availability in 2022 of medical devices is already endangered since less than 20% of the devices have been transferred from the AIMDD and MDD to the MDR. The administrative burden and unclarity of requirements of the MDR play a major role. The proposed EU-AIA can make an unacceptable situation worse for the European patient. Currently, most AI-based medical devices are placed first in the USA market, and it is questionable if they will reach the EU market.
When the MDR is removed from the EU-AIA annex II section A, then most inconsistencies and duplications are removed. AI-based medical devices are already in great detail regulated under the MDR. The European Commission has a scientific workgroup CORE-MD, investigating what additional requirements are needed for AI-based medical devices under the MDR.
Good Machine Learning Practices
The FDA, Health Canada, and the UK MHRA have published Good Machine Learning Practices (GMLP) that can be used to develop AI medical devices. The goal of Good Machine Learning Practices is to promote safe, effective, and high-quality medical devices that use artificial intelligence and machine learning. The ten practices identify areas where the International Medical Device Regulators Forum (IMDRF) and international standards organizations could work to advance GMLP. The guiding practices can also be used to develop MDCG guidance for the EU-MDR instead of using the EU-AIA.
The guiding practices are shown in the table. Each guiding practice has an explanation in the related publication on how to apply it. Guiding practices 1, 2, 5, 6, and 9 are straightforward, logical and good implementable advice. Guiding practice 3 makes sense; however, it might conflict with the EU-AIA. Guiding practice 4 seems logical but might need adoption for rare diseases where only small data sets are available. Guiding practices 7, 8 and 10 deserve special attention from manufacturers and users since they cause regular issues in practice.
The MDR uses additional concepts to the GMLP, such as benefits need to outweigh the risks, the medical device needs to be state of the art, and clinical evidence needs to be available. However, these concepts are already part of the MDR requirements.
Conclusion
The EU-AIA makes it difficult or impossible for good AI-based medical devices to be placed on the EU market. The EU-AIA should be fully consistent with the MDR. This is most easily achieved by not including the MDR in EU-AIA annex II section A but by including the requirements of the EU-AIA in MDR MDCG guidance.
Recently the USA has become the preferred location to place AI-based medical devices on the market because of the complexities of the MDR. The additional complexity introduced by the EU-AIA will accelerate this development. There will be less choice and delayed access to crucial digital health innovation for European patients and healthcare professionals, which cannot be the purpose of the EU-AIA.
Biography
Leo Hovestadt is involved with developing and implementing medical device regulations and guidance, starting with the Active Implantable Medical Device Directive 30 years ago. Leo is the author of the recent MDR Guide for Medical Device Software, and is a regular key note speaker on topics like medical device clinical evidence, artificial intelligence and digital health. He is the Director EU Governmental Affairs of Elekta, a radiation therapy medical device company.
