NL
The project ‘Testing, adapting and implementing Epic Generative AI (GenAI) tools to the Radboudumc context’ was recently launched. Generative AI (GenAI) offers potential opportunities to reduce the workload of healthcare professionals by using various applications in existing work processes within the electronic patient record system Epic. In recent months, several GenAI applications have been made available to healthcare professionals in a pilot setting, under the guidance of the Radboudumc (Gen)AI Epic working group. The project focuses on two GenAI applications: an application that can automatically generate answers to patient questions via Epic's In Basket messaging system, and a second application that summarises recent medical notes. With this research, we identify what is needed for a structured validation and implementation process of GenAI applications. In this, we combine research on the applications' effects on the work process with implementation research.
This project consists of both quantitative and qualitative research into the effectiveness, user experiences, and bottlenecks and promoting factors in the implementation of these applications in everyday healthcare practice. In addition, work is being done in cooperation with the Institute for Computing and Information Sciences (iCIS) at Radboud University to set up a testing infrastructure. This will allow, for instance, the underlying prompts to be analysed and optimised.
The ultimate goal of the project is to develop a blueprint for validation and implementation for these and future GenAI tools. This blueprint should contribute to secure and efficient deployment of GenAI in Epic.
The project is part of the Data-driven Innovation sector plan and will run until the end of May 2026. The research i s a collaboration between the departments IQ Health (Dr Julie Swillens, Dr Koen Verburg, Kim Bladder MSc), Information Management (IM) (René Willemsen), Intensive Care (Dr Tim Frenzel) of Radboudumc and Radboud University (Prof Arjen de Vries, Institute for Computing and Information Science).