Development and Validation of an AI-Based Hybrid Simulation Program for the Initial Response to Clinical Deterioration in High-Risk Pediatric Patients

다국어 초록 (Multilingual Abstract)

This methodological study developed and evaluated an AI-based hybrid simulation program to improve nursing students’ competency in initially responding to clinical deterioration in high-risk pediatric patients. Guided by the ADDIE model, the 220-minute program was developed from relevant literature, pediatric nursing practicum objectives, international simulation standards, and AI ethics guidelines. It comprised pre-learning, prebriefing, AI-based hybrid simulation, debriefing, and reflective learning, integrating an AI tutor, AI virtual caregiver, and AI-assisted debriefing tool. Content validity and AI-response validity were evaluated by 10 multidisciplinary experts, and feasibility was assessed with eight senior nursing students. For transparency, the LLM-based environment, model-use documentation, and prompt safeguards were specified. Validity was high: content S-CVI/Ave=.96 (domains .93–.98), AI-response S-CVI/Ave=.95 (components .94–.96), and feasibility=4.48±0.23/5. These findings support further quasi-experimental testing of educational effectiveness.