Translating Systems Thinking into Practice

A Guide to Developing Incident Reporting Systems

Natassia Goode, Paul M. Salmon, Michael Lenne, Caroline Finch

• Publisher: CRC Press
• Year: 2018
• ISBN:
  9781317172475
• Edition: 1
• Topic: Technologie et ingénierie
• Subtopic: Santé et sécurité au travail

Systems thinking tells us that human error, violations and technology failures result from poorly designed and managed work systems. To help us understand and prevent injuries and incidents, incident reporting systems must be capable of collecting data on contributory factors from across the overall work system, in addition to factors relating to the immediate context of the event (e.g. front-line workers, environment, and equipment).

This book describes how to design a practical, usable incident reporting system based on this approach. The book contains all the information needed to effectively design and implement a new incident reporting system underpinned by systems thinking. It also provides guidance on how to evaluate and improve existing incident reporting systems so they are practical for users, collect good quality data, and reflect the principles of systems thinking.

Features

• 
  
  

• Highlights the key principles of systems thinking for designing incident reporting systems
• Outlines a process for developing and testing incident reporting systems
• Describes how to evaluate incident reporting systems to ensure they are practical, usable, and collect good quality data
• Provides detailed guidance on how to analyze incident data, and translate the findings into appropriate incident prevention strategies

1Systems Thinking and Incident Causation

Practitioner Summary

The so-called systems thinking approach is now the dominant approach to understanding and preventing incidents in the safety science literature. As such, there are now various systems thinking–based models of accident causation available, and it is often difficult to determine which is the most appropriate to underpin safety management efforts. This chapter provides an overview of the systems thinking approach, and discusses prominent accident causation models from the contemporary literature. The intention is for the reader to gain an understanding of key systems thinking principles and state-of-the-art accident causation models. In closing, we articulate the key implications of the models for the design of incident reporting systems.

1.1Introduction

The term ‘systems thinking’ is used throughout this book to describe a philosophy which can be used to understand and improve performance and safety in complex sociotechnical systems. In human factors and safety science, this philosophy is now widely accepted to be the most appropriate to underpin safety management practices. Accordingly, safety management tools such as accident analysis methods and incident reporting systems should be developed based on systems thinking principles.

One of the key contributions of the systems thinking philosophy is to provide a series of accident causation models that consider incidents as a systems phenomenon. When designing rather than considering incident reporting systems, the importance of using an appropriate accident causation model cannot be understated. The underlying accident causation model determines the type of data collected, the method used to analyze the data, and the recommendations that will be proposed (Underwood & Waterson, 2013). Notably, our understanding of incident causation has evolved significantly over time, resulting in three distinct types of model: sequential, epidemiological, and systemic (Hollnagel, 2004). Sequential models, such as Heinrich’s (1931) domino model, view incidents as a linear sequence of events, with human error and mechanical failures being seen as the primary causes of incidents. Epidemiological models, such as Reason’s (1990, 1997) Swiss Cheese model, view incidents causation as similar to the spreading of disease, and emphasize how latent conditions within an organization result in unsafe acts made by operators at the so-called sharp end. Finally, systemic models, such as Rasmussen’s (1997) risk management framework, view incidents as the result of multiple decisions and actions across the overall system of work.

Whilst systemic models, or systems thinking–based models as we will call them, are now undoubtedly the most widely accepted, epidemiological models are arguably still dominant in practice (Salmon, Cornelissen, & Trotter, 2012; Underwood & Waterson, 2014). Many researchers have called for better translation of systems thinking–based models in practice. In the case of incident reporting systems, it is our view that systems thinking–based models should provide the underpinning philosophy. The aim of this chapter is therefore to provide an overview of the systems thinking approach, and discuss the most widely applied contemporary models. The intention is for the reader to gain an understanding of key systems thinking principles and state-of-the-art accident causation models. In closing, we articulate the key implications of the models for the design of incident reporting systems.

1.2Introduction to Systems Thinking

The systems approach to incident causation and analysis is a long and established philosophy that first emerged in the early 1900s (e.g. Heinrich, 1931). It has since evolved through several accident causation models and analysis methods (e.g. Hollnagel, 2012; Leveson, 2004; Perrow, 1984; Rasmussen, 1997; Reason, 1990). Building on key tenets of systems and complexity theory, the philosophy is underpinned by the notion that incidents, and indeed safety, are emergent properties arising from non-linear interactions between components across entire work and social systems (e.g. Leveson, 2004).

There are three key principles of the systems thinking approach that are common across the various accident causation models presented in the literature:

1. Multiple contributory factors spanning multiple hierarchical system levels. Incidents are created by an interacting web of contributory factors that spans all levels of the work system, from the operational frontline (worker, equipment, and environment) all the way up to, and including, regulation and government.
2. Multiple actors and a shared responsibility. The web of interacting contributory factors is created by the decisions and actions of all actors within the system, including frontline workers, supervisors and managers, chief executives, and government personnel to name only a few. Accordingly, there is a shared responsibility for incident causation and prevention that spans all levels of the work system.
3. Up and out not down and in. Incident analysis and prevention efforts should be blame free and take the overall work system as the unit of analysis, rather than the individuals working within it. This involves going ‘up and out’ rather than ‘down and in’ during incident analysis. Incident prevention strategies should focus on optimising the interactions between the components in the system, rather than focusing on individual components alone.

The systems thinking approach to safety therefore involves taking the overall system as the unit of analysis. This involves looking beyond the behavior of the individuals involved in incidents and the immediate circumstances of the event. This view also encompasses factors within the broader organizational, social, or political system in which processes or operations take place. Taking a systems thinking perspective, incidents emerge not from the decisions or actions of an individual but from interactions between humans and technology across the wider system. This means that decisions and actions made at government, regulatory, and organizational levels all play a role in incidents. This calls for a more holistic approach to safety management that considers the role of all the actors within the system, as well as the interactions between them.

It is worth noting that the systems thinking approach has only recently emerged as the dominant approach to understanding incident causation. Prior to this, a more deterministic, reductionist approach prevailed (and indeed still does in some areas). This common approach involves going ‘down and in’ to examine the performance of people and equipment, and their role in incidents. This is reflected in the common fixation on ‘human error’ and the behavior of frontline operators (e.g. pilots, drivers, control room operators) as the primary cause of incidents. For over four decades, headlines describing ‘human error’ or behavior as the primary ‘root cause’ of major incidents have been the norm.

In many organizations, safety management practices are largely driven by this limited perspective. In the case of incident reporting systems, reductionist systems use data fields focused on the behavior of those at the sharp end, not permitting users to report contributory factors related to others across the wider work system. Consequently, the resulting prevention strategies tend to focus predominantly on frontline operators and aim to improve their behavior through education, training, enforcement, and the prohibition of undesirable behaviors. Whilst in some cases parts of safety-critical systems may be improved (e.g. through training programs), little consideration is given to how these parts interact with one another, or how the system functions as a whole.

1.3Systems Thinking Applied

To demonstrate the differences between the systems thinking approach, and the reductionist human error approach, it is worth discussing both perspectives when applied to a recent high-profile incident. For this purpose, we use the tragic Air France crash of 2009, in which an Airbus A330 stalled and crashed into the Atlantic Ocean, killing all 228 people on board.

The incident occurred on the 31st May 2009 during a scheduled passenger flight from Rio de Ja...