Quantifying resilient safety culture using complex network theory

Abstract

Safety is defined as the absence of accidents where accident is an event which lead to unacceptable loss. Previously, most systems employed conventional risk management systems to deal with risks which was based on knowledge of previous experiences, failure reporting and risk assessments by computing historic data. But today, these are traced to organizational factors, functional performance variability and unexpected outcomes or it can be pointed towards systems thinking. Resilience engineering is recognized as other alternative to traditional approaches in safety management. The idea behind resilience engineering is that an organization must continually manage risks and create an anticipating, monitoring, responding and learning culture. This is resilient safety culture. Resilient safety culture is a new concept which has been proposed in order to cover the weaknesses of traditional approaches of safety culture. It is a safety culture with resilience, learning, continuous improvements and cost effectiveness. This resilience takes into consideration the dynamic aspect of the safety culture which makes it resilient to any risks which a safety system faces. The main drawback is the dynamic aspect of the culture is not taken into consideration which is the interaction between people, technology and administration. These interactions are quite complex in nature and difficult to understand and quantify. That is why this study investigates the understanding of these interactions using complex network theory. Once these interactions are understood to some extent, the prediction and prevention of incidents can be done to some extent. There are four different kinds of indicators in the system. Two are system performance indicators, leading and lagging and the other two are the risk indicators that as well leading and lagging. The system performance indicators are indicators which show how the system is performing either in current state which is leading and the system performance indicator which is lagging is gauged by efficiency of the system after a time such as injury rate. Risk indicators leading is found by understanding the various risks which are prevalent in the system and lagging risk indicators are the indicators which led to an accident in previous time frame. Since the system is dynamic, it needs to be understood that these indicators have a time value attached to it. If there is an accident which happened due to some lagging risk indicator, that is in previous time frame, that may have already changed by the time the accidents happened so safety-1 concept which looked at just lagging indicators to dictate the future evaluations of the organization need to be modified and thus resilient safety culture methodology is getting evolved using resilience engineering. Using fault tree analysis, the interactions of various components in a safety system can be understood. Resilient safety culture is treated as a system and it has three sub systems. The sub system further has factors which are important relationships to understand the whole system. These relations between the factors and subsystem are used to measure the resilience of the whole system. This is an innovative quantifying way in which we can improve the resilience in safety culture of an organization. In this study, the qualitative variables defined using the literature are correlated using qualitative as well as quantitative approaches. In the qualitative approach, Leximancer tool is used which model the variables using the literature data. Next, the resilient safety culture model is generated and then fault tree analysis is used to decipher the complex interactions which can help understand which relationships can lead to incident. This study would generate a tool which would help organizations look at the weak links and nodes in their organization to better equip and enhance resources to make the organization more resilient against any safety risks. Multiple case studies are done to validate this model and to show how the whole process is done to understand a way to reduce and mitigate risks. Resilience index is generated which helps in finding which constructs are lagging or weak in giving that index number and the index can be used to compare to companies or organizations irrespective of the number of respondents or the type of indicators which are used. It also helps in reducing the linguistic bias. The findings of this study show that in resilient safety culture model, which components should be focussed first and how the components of resilient safety culture model are related with each other. This helps in optimization of the components or subcomponents to get the maximum resilience in an organization. It is also found that weak areas in an organization can be successfully deciphered using the fault tree analysis approach along with visualization of failure paths. This resilience safety culture model generated along with the methodology adopted in this study can help the industry to making right decisions in enhancing the resilience of the organizations with minimum intervention. It can help the industry find the weak areas where the intervention is needed. It can also give leading indicators which can cause future incidents.