2

Home

Bo Vanhoof edited this page 2026-02-02 14:20:24 +01:00

Welcome to FRAIT's Wiki.

The main objectives of the FRAIT project are the following:

• Develop a Human-Centered Framework for Prompt Creation: Create and validate a structured, replicable process for generating individualized prompts that guide large language models (LLMs) in producing discharge summaries tailored to the specific needs of healthcare providers.
• Enhance Clinical Decision Support: Provide clear, structured, and trustworthy summaries of patient discharge letters to reduce complexity, minimize errors, and support better clinical decision-making.
• Co-Creation with Healthcare Professionals: Engage clinicians through workshops and questionnaires to define ideal summary formats, ensuring usability and alignment with real-world workflows.
• Evaluate LLM Performance in Healthcare Context: Systematically assess the accuracy, reliability, and interpretability of AI-generated summaries using a dedicated evaluation framework and expert consensus.
• Ensure Compliance and Ethical Standards: Implement GDPR-compliant processes, protocols for the use of synthetic data (such as synthetic discharge letters), and secure infrastructure to ensure the responsible and safe application of AI in clinical environments.

The project was coordinated by Mieke Deschepper (mieke.deschepper@uzgent.be).