Following the Quality Management System (QMS) ensures that software as a medical device (SaMD) and AI applications in the hospital are safe and effective, and that they comply with applicable laws and regulations.
The (in-vitro) medical devices regulation (MDR/IVDR) stipulates that a hospital may develop medical devices for in-house use, provided that it has an appropriate Quality Management System (QMS) in place. In addition, the AI Act also requires the use of a quality system when developing high-risk AI applications. For this reason, the Digital Health department has implemented a QMS in accordance with the ISO 13485 standard (quality management for medical devices) for the development of SaMD and AI software for use within UMC Utrecht. In addition to our QMS, similar systems exist at MTKF and the Department of Radiotherapy.
The QMS describes the processes that must be followed to develop a safe and effective device. Essential processes include the software development and maintenance process (in accordance with IEC 62304), risk management (in accordance with ISO 14971), and clinical evaluation. It is also supplemented with AI best practices, such as mapping to the leidraad AI and Impact Assessment Mensenrechten en Algoritmes (IAMA). By following this QMS, we ensure the quality of our (AI) software, meet regulatory requirements such as MDR/IVDR, and prepare for full implementation of the AI Act. Furthermore, well-structured processes contribute to a faster innovation cycle.
Not only does the AI for Health team use the QMS, but it is also available hospital-wide, allowing different departments within UMC Utrecht to develop AI software for clinical use with the appropriate guidance. An overview of all QMS-related documents is available for UMC Utrecht staff on Zenya.
Current State
The QMS has now been fully adopted within the AI for Health team, and the first projects have been completed under the new system. ISO 13485 Certification has been obtained for this QMS..
Developers: Jacqueline Pronk-van der Noort en Richard Bartels (AI for Health)
Commissioned by: AI for Health (Digital Health)
Users: Hospital-wide