Diagnosing firefighter safety governance fragility through CAS-IPA in an island ICE environment: A full-census study in Lienchiang County, Taiwan

Firefighter occupational health and safety (OHS) systems operate within highly dynamic socio-technical environments. In geographically isolated regions, limited resources, spatial separation, and unpredictable conditions can greatly increase operational risks. Although safety science has shifted from a Safety-I focus on error prevention toward Safety-II and Resilience Engineering, which emphasize a system's capacity to succeed under variability and uncertainty, firefighter OHS assessments continue to rely heavily on static checklists and linear tools such as Importance–Performance Analysis (IPA). This study explores CAS-IPA as an exploratory diagnostic approach that combines conventional IPA with insights from Complex Adaptive Systems (CAS) theory. Performance gaps are viewed as potential signals of disrupted feedback processes among adaptive agents. A full census of all 46 personnel in the Lienchiang County Fire Bureau, Taiwan (100% response rate), was conducted. Data from a purpose-developed 30-item occupational safety scale covering five dimensions were analyzed through reliability testing, correlations, and a reconstructed CAS-IPA matrix. The results show strong inter-dimensional coupling within the firefighter OHS system (Cronbach's α = 0.97; r = 0.72–0.92). Three notable fragility nodes were identified: assignment of on-site safety officers (Gap = −0.61), inter-island support coordination (Gap = −0.46), and emergency evacuation routes and signage (Gap = −0.54). The M-shaped distribution of service years suggests a possible weakening of mid-career knowledge bridging. Separately, the low scores for health-anomaly reporting may represent an under-reporting risk cue potentially associated with organizational silence. This census-based analysis illustrates how the observed patterns align with the three conceptual propositions advanced in the framework. Overall, the study illustrates how a CAS-informed reading of IPA data may offer additional system-level insights in Isolated, Confined, and Extreme (ICE) environments, while clearly acknowledging the exploratory and inferential nature of the interpretations.