The European Commission (EC), the European authority in the sphere of medical devices, has proposed a new framework for the use of Artificial Intelligence technology in medical devices. The document is intended to introduce harmonized rules for AI-based medical devices allowed to be marketed and used in the European Union.
Regulatory Background
By virtue of the document, the EC will implement a new regulation that will address all matters related to AI-based products. The authority acknowledges the increasing importance of this technology and also the need for a clear and efficient regulatory framework to be introduced. Further development of AI-based medical devices could provide additional benefits for patients. At the same time, it is important to ensure such devices meet safety and performance requirements, which require certain improvements due to the unique features of AI-based products resulting from the specifics of the technology itself. Thus, a balanced approach is required. The document reflects the principles initially announced in the White Paper on AI published by the EC earlier in February 2020. The aforementioned document describes an approach that balances the creation of a favorable regulatory framework encouraging further development of AI-based products with the need to ensure that associated risks are duly mitigated.
The new regulation is intended to introduce a harmonized framework for AI-based medical devices in order to ensure their safety and effectiveness, together with data protection, when applicable. In particular, the European Commission outlines the following goals of the new regulation:
– Ensuring the safety of AI-based products,
– Improving legal clarity of the regulatory framework,
– Making existing regulations more effective,
– Encouraging and facilitating further development of AI technology.
For this purpose, the EC intends to introduce a balanced regulatory approach reducing the regulatory burden to the lowest extent possible regarding the safety-related aspects to simplify the procedures to be performed to place AI-based medical devices on the market. The authority expects the suggested risk-based regulatory approach to be flexible enough to reflect the current needs of this emerging sphere. In particular, it provides the definition of artificial intelligence and also distinguishes its permitted and prohibited use. Once the new regulation becomes effective, any and all medical devices based on the use of AI technology should comply with the regulatory requirements set forth therein in order to be allowed for marketing and use in the European Union.
The authority additionally emphasizes that the obligations of the medical device manufacturers should be clear and predictable – this will encourage further development of medical devices employing novel technologies. It is also stated that additional steps will take place, namely:
– Establishment of a European Artificial Intelligence Board,
– Regulatory sandboxes for AI-based products,
– Special support programs for start-ups engaged in the development of AI-based medical devices.
The document also describes in detail the way the new regulation would correspond to existing legislation. It is important to mention that the actual scope of application of the proposed regulation exceeds the healthcare sphere and covers any use of AI-based products.
The proposed regulatory framework encourages moving the focus from national to international rules and requirements. As stated in the document, regulation by means of national legislation will only create additional complications, while international cooperation of all Member States will improve the effectiveness of the mechanisms necessary to ensure the proper balance between safety and availability of AI-based products. This becomes especially important in terms of international trade and open markets.
As was mentioned before, the new regulatory framework employs a balanced risk-based approach intended to improve transparency and remove the unnecessary regulatory burden medical device manufacturers can face nowadays. However, the regulatory requirements to be applied for high-risk AI-based products would remain quite strict in order to ensure the safety of such products. Due to the specific features of AI-based products, additional attention should be paid to the accuracy of data used to improve the operations of the products. The authority additionally emphasizes the important role to be played by harmonized standards and additional guidelines dedicate to compliance matters – they should become a valuable source of information to be used by the manufacturers for ensuring compliance with the applicable regulatory requirements.
AI-based Products: General Approach
The EC further describes the way the new regulatory framework has been developed. According to the document, the Commission has carried out consultations to receive feedback and suggestions from the interested parties. It is stated that the appropriate consultation commenced on February 19, 2020, when the White Paper on AI-based solutions mentioned hereabove has been published. The feedback and suggestions provided by stakeholders during these consultations have been considered by the authority when developing a new regulatory framework described herein. In the course of consultations, the Commission received feedback from industry representatives, individuals, research institutions, and public authorities. According to the responses provided by the respondents, the following aspects are the most important:
- There are certain regulatory gaps, so legal clarity is definitely needed.
- Overregulation is something to be avoided.
- The new framework should be neutral and balanced.
- A risk-based approach is optimal and acceptable for all the parties involved.
- Clear definitions of the most important terms, as well as detailed clarifications on the applicable risk-based classification rules in the context of AI systems, are required.
- Special regulatory sandboxes would allow to encourage further development of novel technologies and medical devices based thereon.
Apart from consultations with the stakeholders, the Commission has already requested input from experts. In particular, a High-Level Expert Group on AI (HLEG) has been established earlier in 2018.
As the next step, the EC conducted an impact assessment in order to determine the impact potentially to be caused by the suggested regulation. This allowed the authority to identify the optimal way the initial goals could be achieved.
Additional Aspects
The new framework for high-risk AI-based products would apply to any and all manufacturers and users of such products. For instance, it will introduce clear and balanced rules the manufacturers shall follow in order to be allowed to place their products on the EU market, while the detailed rules will also prescribe the way the AI-based products should be used in order to ensure their safety and effectiveness. As mentioned, the document is also intended to introduced special regulatory measures aimed at the facilitation of further technology development, such as regulatory sandboxes and other mechanisms creating favorable conditions for the parties engaged in the development and manufacturing of innovative products.
The authority also mentions that in order to evaluate the actual effectiveness of the new regulation, it will also perform ongoing monitoring. The appropriate system will be established to register stand-alone high-risk AI applications in order to make information about them easily accessible on an EU-wide level.
In summary, the new regulatory framework for AI-based products developed in the EU is intended to fill existing gaps in legislation and create favorable regulatory conditions for the parties engaged in the development of novel AI systems. The document employs a risk-based approach and announces special measures to be introduced to facilitate further development of artificial intelligence technology, such as regulatory sandboxes.
How Can RegDesk Help?
RegDesk is a next-generation web-based software for medical device and IVD companies. Our cutting-edge platform uses machine learning to provide regulatory intelligence, application preparation, submission, and approvals management globally. Our clients also have access to our network of over 4000 compliance experts worldwide to obtain verification on critical questions. Applications that normally take 6 months to prepare can now be prepared within 6 days using RegDesk Dash(TM). Global expansion has never been this simple.
