{"slug":"cliniques-universitaires-saint-luc-deploys-an-on-premises-qwen3-235b-assistant-inside-its-epic-ehr","url":"https://findausecase.com/use-cases/cliniques-universitaires-saint-luc-deploys-an-on-premises-qwen3-235b-assistant-inside-its-epic-ehr","title":"Cliniques universitaires Saint-Luc deploys an on-premises Qwen3-235B assistant inside its Epic EHR","description":"A Brussels university hospital hosts a GDPR-compliant LLM assistant on-site on a server with 8 NVIDIA H200 GPUs, using Qwen3-235B, vLLM, LiteLLM, Open WebUI and retrieval over EHR documents and literature. After a one-month pilot with 28 physicians, it was deployed hospital-wide and adopted by 1,028 users who generated 14,910 conversations over five months.","company":"Cliniques universitaires Saint-Luc","industry":"Healthcare","country":"Belgium","aiCapabilities":["Generative AI","Conversational AI"],"technology":["Qwen3-235B","vLLM","LiteLLM","Open WebUI","Epic","NVIDIA H200","Apache Solr","Docling"],"deployment":"On-Premise","problemStatement":"Electronic Health Records have improved access to patient information but substantially increased clinicians' documentation workload, and physicians must also process a large volume of information when preparing a consultation.","solutionApproach":"The hospital implemented a secure, GDPR-compliant, on-premises LLM assistant integrated into its Epic EHR. It uses Qwen3-235B (FP8) served with vLLM, with LiteLLM managing access control and request distribution, and retrieval augmented generation over structured patient data, internal and regional clinical documents, internal literature and medical literature. The interface is built on Open WebUI and embedded in Epic via the SMART on FHIR protocol, and every answer lists the sources consulted. The system runs on an on-site server with 8 NVIDIA H200 GPUs.","businessValue":"After a one-month pilot with 28 physicians across nine specialties, 64% of whom used the assistant daily, the system was deployed hospital-wide and adopted by 1,028 users who generated 14,910 conversations over five months, with more than half of clinicians using it at least weekly. Usage concentrated on information access and documentation support. The authors state they deliberately did not measure clinical outcomes or time savings.","evidence":{"band":"high"},"sourceUrl":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12716761/","dates":{"publishedAt":"2026-10-22T05:49:09.579Z","publishedAtSource":"pipeline","updatedAt":"2026-10-22T05:49:09.579Z"},"license":"Open for reading and citing with a link to https://findausecase.com/use-cases/cliniques-universitaires-saint-luc-deploys-an-on-premises-qwen3-235b-assistant-inside-its-epic-ehr. Bulk republication requires permission."}