Shin Kong Wu Ho-Su Memorial Hospital
.png)
2024 National Healthcare Quality Award
Highlights of Workload Reduction
1. AI auto-screening replaces manual image triage, reducing workload and errors.
2. Priority alerts help radiologists focus on urgent cases more efficiently.
3. A streamlined workflow allows staff to concentrate on core duties.
4. Detection time was cut from 7 hours to 12 minutes, minimizing rework.
5. Automated monitoring reduces IT maintenance effort.
6. A no-code platform enables clinical staff to develop AI tools independently.
This project was jointly led by the Department of Radiology, Research Department, AI Center, IT Department, and Health Management Department, in collaboration with DeepQ and FarEasTone Communications, to implement an Al-powered pneumothorax early warning system.
The AI model was trained using 6,529 chest X-ray (CXR) images from our hospital and integrated into the medical imaging platform. Upon image acquisition, the system automatically analyzes the CXR. If a suspected pneu-mothorax is detected, an SMS alert is immediately sent to radiologists for priority review, greatly improving detection efficiency and patient safety.
Source: 2025 Application of Digital Healthecare