A Review of Accident Models and Incident Analysis Techniques

This review article aims to provide an overview of accident models and incident analysis techniques in the context of radiation oncology. Accident models conceptualize the mechanisms through which accidents occur. Chain-of-event models and systemic models are two main categories of accident models and differ in how accident causation is portrayed. Chain-of-event models focus on the linear sequence of events leading up to an accident, whereas systemic models emphasize the nonlinear relationships between the components in a complex system. The article then introduces various incident analysis techniques, including root cause analysis (RCA), London Protocol, AcciMap, and Causal Analysis Based on Systems Theory (CAST), which are based on these accident models.  The techniques based on the chain-of-event model can be effective in identifying causal factors, safety interventions, and improving safety.  The other techniques based on the systemic models inherently facilitate an examination of how the influence of personal conditions, environmental conditions, and information exchange between different aspects of a system contributed to an accident.  To improve incident analysis, it is essential to translate unsafe human behavior into decision-making flaws and the underlying contextual factors. Where resources allow, it is also crucial to systematically link frontline contributions to organizational and societal aspects of the system and incorporate expertise in safety science and human factors into the analysis team.  The article also touches on related concepts such as Perrow's Normal Accident Theory (NAT), Functional Resonance Analysis Method (FRAM), and Bowtie Analysis, which are not based on specific accident models but have been used for safety improvement in radiation oncology. Overall, different incident analysis techniques have strengths and weaknesses. Taking a systems approach to incident analysis requires a shift from linear thinking to a more nuanced understanding of complex systems. However, the approach also brings unique value and can help improve safety as radiation oncology further gains complexity.