クリティカルケアユニットにおける避難計画強化のためのデジタル技術：定量的・実験的シミュレーション

This study aims to evaluate evacuation performance in a high-acuity intensive care unit (ICU) by integrating building information modeling (BIM) and computational simulation and examine how different parametrizations affect the fidelity of predicted evacuation dynamics when compared with data from a real evacuation drill. A detailed BIM model of a tertiary hospital ICU was developed and imported into Pathfinder to simulate 3 protocol-based evacuation scenarios and 1 scenario calibrated with preparation times obtained from a 2024 evacuation drill. Simulations assessed occupant flow, bottlenecks, movement trajectories, and total evacuation times. Architectural constraints, behavioral rules, and patient dependency profiles were incorporated to reflect real operational conditions, including explicit geometric incompatibilities between door widths and ICU bed dimensions. Evacuation times produced with generic parameters showed a gap of up to 7 minutes 40 seconds less than the real drill duration, whereas the empirically calibrated simulation narrowed this difference to 2 minutes 30 seconds. Across all scenarios, bottlenecks consistently emerged at the main critical patient unit exit, primarily driven by architectural constraints such as insufficient door widths relative to bed size, which hindered bed maneuverability and intensified staff circulation conflicts. Flow rates peaked between 0.40 and 0.55 persons/s, with higher occupancy producing more sustained congestion. Preparatory actions for clinically complex patients significantly shaped overall evacuation performance. The integration of BIM and empirically informed simulation enhances the accuracy and operational relevance of evacuation analyses in complex clinical environments. Findings highlight the need for calibration based on real behavioral data and demonstrate how specific architectural mismatches, particularly between door geometry and bed dimensions, act as critical drivers of evacuation bottlenecks. This approach supports more robust risk assessment and emergency planning within critical socio-technical infrastructures.