A system-based framework for human factors risk analysis in maritime transportation: integrating STPA, HFACS, and Bayesian networks

Understanding human error in maritime accidents is challenging due to the complex interactions between humans and technology. This study proposes a comprehensive, system-based approach to assess human error risk in maritime transportation. The System-Theoretic Process Analysis (STPA) is employed for holistic hazard identification, the Human Factors Analysis and Classification System (HFACS) structures the classification of human errors of STPA's loss scenarios, and a Bayesian Network (BN) quantifies the relationships among identified factors. Results reveal that unsafe acts are not solely attributable to seafarers’ cognitive limitations or operational mistakes but are significantly influenced by contextual factors such as organisational conditions, supervisory practices, technical systems, and environmental aspects. STPA, HFACS, and BN integration offer a structured framework to trace causal pathways leading to human error. The study also recommends risk control strategies to address these contributing factors, supporting stakeholders in enhancing maritime safety through more effective interventions.