Exposing resistant problems in complex systems: a review of accident causation tenets

The plateau of incident rates in many domains, including road and rail transport, aviation and workplaces indicates that system-based ergonomics models used to understand accident causation are either underutilised or underperforming in preventing adverse events. Whether this plateau is due to an increase in system complexity or out-of-date toolkits, the problems are seemingly resistant to the current ergonomics methods that pursue them. With this knowledge it is pertinent to ask what changes, if any, are required to capture contemporary system complexity and uncover resistant problems. Are current methods being used to their full advantage or are new methods required to progress safety science? The paper presents a review of the dominant contemporary safety science methods and underpinning models that could potentially provide a solution to uncovering resistant problems in complex systems. Accident causation theorists with the greatest number of citations were identified from the safety science literature. The citation information was derived from Scopus (April 2016). These included, Nancy Leveson (3950 citations Scopus, April 2016) Jens Rasmussen (3486 citations Scopus, April 2016), Charles Perrow (2041 citations Scopus, April 2016), Sidney Dekker (789 citations Scopus, April 2016) and Erik Hollnagel (672 citations Scopus, April 2016). Over ninety published works were coded into categories of safe and unsafe system behaviours. The outcome of the review produced a list of principle tenets extracted from the literature. The review has shown that despite the diversity in approaches there is considerable agreement about the core tenets of system safety and accident causation. Using this information a research program is proposed to test the quality of the tenets extracted from the literature review and the extent to which they can be used to support a proactive approach to safety in complex systems. This will provide a test of sociotechnical systems theory when used to predict accidents