Large Language Model for Dutch Healthcare
In the Netherlands, general practitioners (GPs) are essential for healthcare. However, GPs face overwhelming workloads, leading to stress, staff shortages, and reduced access to care. Our project unites experts from Erasmus MC, TU Delft, and AI startup Juvoly to develop a Dutch medical large language model (LLM) specifically designed to reduce administrative burdens for GPs. This innovation will improve GP well-being, enhance patient care, and set a new benchmark for AI in healthcare.
GP workloads are at a critical level, with 68% reporting unsustainable pressures and 18% reporting very high levels of stress. Over half spend more than a day per week on administrative tasks, limiting time for patient care. These challenges, driven by an aging population and increasing healthcare demands, harm both doctors and patients. Current AI models are not tailored to Dutch healthcare, creating a need for a solution that is local, efficient, and effective.
Our Dutch medical LLM will handle tasks such as writing referral letters, summarizing patient records, and finding medical guidelines. Developed in close collaboration with GPs and patients, it will seamlessly integrate into clinical workflows. By focusing on Dutch healthcare, our model will be smaller, faster, and more sustainable than existing alternatives, reducing costs and environmental impact.
Once completed, this project will deliver a validated, high-fidelity prototype ready for implementation in general practices. By reducing workload, the LLM will improve accessibility of healthcare for all Dutch residents and enhance the well-being of GPs, reducing burnout and attrition rates. Economically, the LLM could save up to one day per week per GP, enabling more focus on patient care, improving health outcomes, and lowering healthcare costs. Beyond its immediate impact, the LLM will generate valuable insights into AI design, paving the way for broader adoption of inclusive, effective AI in healthcare.
