Patient Outcomes
1. A1C <7% achievement rate reached 58.3%, exceeding the national average of 51.5% and the medical center average of 52.3%.
2. Poor control group (A1C >9%) improved by an average reduction of 2.74%.
3. 49.7% of patients achieved at least a 1% A1C reduction, and 30.8% achieved a 5% reduction.
Clinical Nursing Outcomes
1. Nursing documentation saved 11,235 minutes of work time per month, reducing labor costs by NT$39,323.
2. Documentation quality improved, cutting documentation time by a factor of 2–3.
3. GPT/AI-assisted nurse–patient teaching enabled new staff to identify care priorities more effectively, lowering shift handover stress scores from 9 to 6.
4. Reduced case managers’ time spent searching across systems, ensuring uninterrupted continuity of care.
Medication Safety and Efficiency Outcomes
1. Digital insulin management eliminated paper use, saving 1.2 trees annually and reducing medication preparation time by 148 hours per year.
2. The insulin body-map system aligned with patients’ injection habits and strengthened nurse–patient collaboration on injection-site selection.
Integration of AI Health Education into Nursing Processes
1. Provided Indigenous elders with diabetes education in the Taroko language, improving accessibility of clinical information.
2. Use of mother-tongue health communication reinforced human-centered care and respect for life.