Hualien Tzu Chi Hospital x iBodhi x FedGPT: AI Senior Supports Nurses for On-Time Shift Endings
2025/04/25
“Friends from overseas Tzu Chi chapters are amazed—Taiwan is truly ahead in this area!”— Dr. Hsin-Jung Lin | Superintendent, Hualien Tzu Chi Hospital
“I commend Hualien Tzu Chi’s AI integration. Indigenous-language health education can truly benefit Indigenous communities.” — Bi-khim Hsiao | Vice President of Taiwan
Background
Customer: Hualien Tzu Chi Hospital | Industry: Medical Service
Challenges
As the only medical center in Eastern Taiwan, Hualien Tzu Chi Hospital handles over 800,000 outpatient visits annually and has around 1,000 nurses. Among their core responsibilities, nursing handovers are particularly challenging—especially for newcomers.
Traditionally, the handover process followed an apprenticeship model: a senior nurse orally conveyed the status of 6–7 patients, while the junior nurse took notes. This method not only tested the new nurse’s ability to organize complex information but also left room for manual errors and omissions, ultimately affecting care quality and extending nurses’ working hours.
Solutions
- Product: FedGPT & ASR (On-premises)
The Most Medically Literate GPT: FedGPT Surpasses Large Models in Medical Knowledge
Hualien Tzu Chi Hospital has adopted FedGPT, developed by Yating Intelligence. FedGPT is the world’s first generative AI to seamlessly integrate ASR (Automatic Speech Recognition), TTS (Text-to-Speech), virtual avatars, and a federated learning platform. Tailored for medical settings, the team trained FedGPT on over 15 billion tokens of biomedical corpora, equipping it with deep expertise in healthcare.
With just a tap to record conversations, nurses can rely on FedGPT to transcribe doctor-patient dialogues, summarize patient conditions, generate handover notes, and list key reminders—such as which patient needs a medication change or an exam. Like a “senior AI nurse,” FedGPT helps juniors complete handover documentation while gently reminding them not to miss important tasks.
FedGPT is also resource-efficient: despite using less computational power, it outperformed peer models with similar 14B parameters in medical comprehension based on the Open Medical-LLM benchmark. Its performance is on par with GPT-4 and even exceeds larger 22B and 34B models—making FedGPT a powerful and energy-saving solution aligned with ESG sustainability goals.
Benefits & Results
Support Nurses for a Better Future! AI Enhances Nursing Handover and Care Quality
Director of Nursing Hui-Chun Chung expressed that as a senior nurse, she hopes AI not only lightens the burden of her team but also empowers new nurses, enhances patient interaction, and improves health literacy—fostering a positive healthcare cycle.
At Hualien Tzu Chi, FedGPT has achieved over 80% accuracy in nursing handovers and admission record generation, with a satisfaction score of 4.7 out of 5 among nurses. This innovation has shaped a new standard: “Effortless Handover, On-Time Shift Completion.” The hospital was awarded Eastern Taiwan’s first Gold Medal in the SNQ National Quality Awards, becoming a benchmark for smart healthcare. As one report noted: “AI Senior Nurses capture every detail during handover, boosting both accuracy and efficiency.” By integrating AI in clinical documentation and nursing workflows, Hualien Tzu Chi Hospital is leading the way in sustainable, high-quality care.