eLLM-oH – Enabling Large Language Models in the Organization of Hospital

The eLLM-oH project utilizes Large Language Models (LLMs) to optimize emergency department operations in acute healthcare. Its goal is to automate administrative tasks, reduce staff workload, and improve patient and nurse support. The aim is to decrease waiting times, enhance admission accuracy, rersource optimization, and provide improved assistance to patients with disabilities or language barriers. The target audience is healthcare staff and patients, with hospitals being the primary customers. The key features of this system include speech interaction, natural language processing for symptom gathering, self-service registration, and multiple language options. These features are aimed at minimizing patient-staff contact and improving the emergency care experience.

The minimum viable product (MVP) focuses on voice input analysis for creating admission forms, integrating basic speech recognition, multilingual capabilities, and GDPR compliance, without supplanting medical expertise. The MVP is scheduled for testing with healthcare professionals. Its goal is to improve documentation efficiency and accuracy, with a focus on emergency rooms initially, but with potential for broader healthcare applications.

A project at the St. Pölten University of Applied Sciences
Master Program Digital Healthcare
Project Coaches: FH-Prof. Andreas Jakl, M.Sc., Dipl.-Ing. David Feilacher, B.Sc.