Registered Nurse (RN)
KMU Hospital is successfully implementation of our Knowledge-Oriented Nursing Informatics Architecture. By integrating advanced Generative AI and our proprietary KMUGPT platform, we have effectively solved the long-standing challenge of excessive documentation time for our nursing staff.
Through Retrieval-Augmented Generation (RAG) technology, our system delivers personalized nursing records that adhere strictly to medical standards. highlights include:
One-click generation of descriptive records and Focus Charting (DART) notes, reducing complex documentation time from 1 hour to just 3–10 minutes.
In 2025, the system generated nearly 270,000 records, significantly improving medical record completeness and staff satisfaction.
Our AI patient education features are available on our KMUHome app, bringing professional guidance directly to the community people.
By alleviating administrative burdens, we are empowering our nurses to focus on what matters most: patient care.
Core Technologies: Hospital-built/independently developed LLM Platform – KMUGPT.
Highlight: Generative record model training based on the nursing process framework integrated with a specialized nursing knowledge base.
Features: Integrates structured NIS (Nursing Information System) data and multi-source HIS (Hospital Information System) data within specific timeframes to ensure personalized and accurate patient records. It enables one-click automatic generation of nursing notes and medical handover summaries.
Innovation: Flexible recording modes that can be adjusted to specific needs, such as SOAPIE or DART formats. The application is further extended to the KMUHome APP (Public Version) for patient health education services.
Generative Nursing Records: The usage volume increased from nearly 260,000 entries in 2024 to nearly 270,000 entries in 2025, with an adoption rate: 80%
One-Click Integration of Generative Nursing Records into Medical Handovers: The usage volume rose from 5,000 entries in 2024 to 14, 230 entries in 2025.
Nursing record completeness: 83.4% before using the generative record system, improved to 100% after import.
Nursing process record accuracy: No significant difference before and after import; the quality of the generative record content is reliable.
Nursing record completeness: 83.4% before using the generative record system, improved to 100% after import.
Staff Satisfaction: Satisfaction was higher among nurses with less than 1 year of service (100%) and those with more than 20 years of service (88.2%). This is significant for nurse retention and simplifying work processes.
Top three satisfaction items in user feedback: accurate content that meets clinical needs, ease of use and integration with workflows, and contribution to improved record-keeping efficiency.
No international achievements yet.
The database used for AI-assisted nursing record generation covers care standards, technical guidelines, health education materials, and ISO quality management documents, possessing high localization and departmental characteristics. This ensures that medical record data complies with regulations and is accurate.
The prerequisite for generating nursing records is the provision of accurate data, allowing nurses sufficient time to perform patient assessments, provide care, and input data. AI retrieves guidance from the database that matches the patient's condition to assist nurses in completing the record.
