(Frankfurt a. M., 10.04.2025) The integration of artificial intelligence (AI) in the healthcare sector opens up numerous new possibilities. However, strict requirements apply to such applications - especially in Europe. In order to develop and approve AI-supported medical devices more quickly, the experts from VDE Health and the AI Quality & Testing Hub (AIQ) have worked closely with the company Prof. Valmed - with great success: Prof. Valmed is the first* AI-supported tool for clinical decision support be certified as a medical device in risk class IIb for the European Union by a Notified Body.
The system, which is based on AI and the use of a large language model (LLM), supports medical professionals by providing trustworthy recommendations for the diagnosis and treatment of illnesses. It processes natural language input and questions in the form of a prompt and generates precise and evidence-based answers in natural language, which are displayed in a user-friendly way with corresponding references.
Regulatory and technical requirements at a glance
The VDE has supported the development and approval process from the very beginning. The starting point was the development of a CE roadmap as the strategic basis for the entire approval process. Based on the "compliance by design" principle, according to which all approval-relevant requirements are already taken into account during development, the experts from VDE Health provided comprehensive support in creating a quality management system and technical documentation.
As a technology partner, AIQ - founded by the VDE and the state of Hesse - provided the corresponding AI development and testing tools and developed the software in compliance with all regulatory requirements on behalf of Prof. Valmed. One focus of this work was to ensure the quality of the results, which the experts achieved by implementing a special AI quality stack. Through rigorous testing, they ensured security as well as transparency and reliability. The AIQ develops and tests AI applications based on the principles of retrieval augmented generation (RAG). This means that the Large Language Model does not simply draw on general or freely available knowledge but searches specifically validated medical specialist sources within a database and provides relevant information in a reliable and up-to-date manner.
Thanks to this structured approach, the partners were able to bring the project to certification in a very short period of time.
Dr. Beate Mand, Deputy Chairwoman of the VDE Executive Board: "Artificial intelligence is one of the dominant technologies of our time. Especially when it comes to AI applications in the healthcare sector, innovation and safety must go hand in hand. The name VDE stands for safety and the highest quality. We are working with our experts to pave the way for the trustworthy and compliant use of AI systems in medicine and to create transparent markets."
As one of the largest technology organizations in Europe, the VDE has stood for innovation and technological progress for more than 130 years. Future topics such as artificial intelligence are a particular focus. The resulting benefits support the high level of safety. The VDE is also the only organization in the world that combines science, standardization, testing, certification and application consulting under one roof.
"The use of AI in healthcare enables innovation and a noticeable improvement in quality. With our AI Quality & Testing Hub, we are promoting the quality of AI systems so that users gain confidence in AI applications. 'AI made in Hessen' should become a trademark of our state that stands for innovation, responsibility and the future," emphasized Hesse's Digital Minister Prof. Dr. Kristina Sinemus.
Fully exploiting the potential of AI solutions in the healthcare sector
"The successful certification of Prof. Valmed as a Class IIb medical device impressively demonstrates that innovative AI-based solutions can meet the demanding regulatory requirements in the healthcare sector. As the first AI-supported tool for clinical decision support, Prof. Valmed confirms both the high technical quality of our product and the efficiency and precision of our regulatory support," say Dr. Vera Rödel and Prof. Dr. Heinz Wiendl from Prof. Valmed.
"Innovative approaches in medicine based on artificial intelligence or machine learning and LLM in particular offer great potential in the healthcare sector," emphasizes Dipl.-Ing. Hans C. Wenner, Senior Manager Regulatory Affairs at VDE. "The structured approach through the CE roadmap and the consistent application of the 'compliance by design' principle were key success factors in efficiently meeting the complex regulatory requirements."
