Being able to carry out meaningful incident investigations is fundamental to the continual improvement of any industry, with the value of the learning gained from an investigation being largely dependent on how the investigation has been conducted. In addition, the significance of the economic and commercial demands and culture inherent within the construction industry should not be underestimated.
Accident causation models have an important part to play in investigations as they provide a structure for the investigator in which to investigate and, thereby, influence the investigation itself due to the assumptions they make about accident causation. There are over a dozen models in circulation in various industries with each model characterised by varying underlying factors telling the investigator what to interrogate. In addition, with most accident causation models having been developed by academics, they are not readily known or easily understood by those in industry. This makes the challenge for organisations within industry that much more difficult to understand the various underlying factors characteristic of each model and to understand the importance of selecting a model that is sympathetic to the customs, practices, processes and demands that make up their respective industry accident causation risk profile.
Most people will be familiar with Heinrich’s (1931 and 1959) “Domino Theory” model where the role of the person, immediate surroundings and management are interrogated as the “cause” of accidents. Unfortunately, such simple models ask investigators to interrogate little beyond the immediate causes and assume that an accident is the end result of a series of logical sequential events much like a row of dominos, where the elimination of one of the barriers will be enough to prevent the accident. However, accident causation is often not a simple cause-effect (also known as a simple linear causation model) relationship but, instead, is almost always a complex construct of both active and latent failures or conditions that often interplay simultaneously.
On the other hand, complex linear models do extend beyond simple cause-effect relationships at the “sharp end” as they also interrogate the more latent conditions much higher up the safety system, including economic influences, as contributory factors. Complex linear models analyse the combined effect of how these various contributing factors interact that ultimately result in the incident. For this reason, complex linear models are also often referred to as systems-based or epidemiological models and are one of the more prevalent types of accident causation models used throughout industry. The problem is that these contributing factors are constantly changing due to the dynamic effect of the system itself, creating the complex interplay relationship of how the latent failures (organisational influences, unsafe supervision and the preconditions for unsafe acts) within the system influence the active failures/unsafe acts at the “sharp end”.
Probably the most recognisable of all complex linear models is the “Swiss Cheese” model developed by Reason (1997) that, although considered to be to suitable to interrogate around 80% of all accidents, there are industry specific causation models that, where developed, are better suited to the customs, practices, processes and demands that make up the risk profile of that particular industry.
One such industry specific model is the Loughborough Construction Accident Causation (LCAC) model (Haslam et al., 2005) that was developed by a team of researchers from Loughborough University and the University of Manchester Institute of Science and Technology (UMIST) and has been found to be capable of identifying a far greater number of both proximal (immediate) and, particularly, distal (latent) factors regardless of the severity/fatality of the adverse event.
In fact, using the LCAC model to evaluate 258 fatal incidents, Cook and Lingard (2011) found that shaping factors and originating influences were seldom identified in investigation reports, resulting in, unfortunately, a heavy bias towards immediate causes declared within investigation reports. Similar empirical research results have been found across various incident types and severity, including near misses within the construction industry using the LCAC model. The LCAC model is illustrated in Fig.1 below that, working from the centre outwards, shows the relationship between the “Immediate Accident Circumstances” (worker factors, site factors, materials/equipment factors), the “Shaping Factors” that shaped the “Immediate Accident Circumstances” that were, in turn, ultimately shaped by the “Originating Influences” on the peripheral of the LCAC model illustration.
Figure 1: Loughborough Construction Accident Causation model (Haslam et al., 2005). Reprinted from Applied Ergonomics, Vol. 36 (4), R.A. Haslam, S.A. Hide, A.G.F. Gibb, D.E. Gyi, T. Pavitt, S. Atkinson, A.R. Duff, Contributing factors in construction accidents, page 15, Copyright (2005), with permission from Elsevier.
Complex non-linear causation models are perhaps the least known and utilised type of accident causation model employed within industry, yet, it is argued, offer a far more holistic and thorough approach to investigating incidents. Rasmussen’s (1997) socio-technical model is an example of a complex non-linear causation model where, unlike simple and complex linear models, the emphasis is not so much on the actions and decisions of the operators being the cause of adverse events as it is about understanding the factors that shape these behaviours that start much higher up, and propagate throughout the safety system that can unwittingly contribute to accidents.
Rasmussen (1997) and Hollnagel (2004) argue that it is inevitable that humans will occasionally deviate from the established rules and procedures; often motivated by adapting to work load and time demands. These deviations either, “trigger(ing) an accidental flow of events or divert a normal flow” (Rasmussen, 1997) with a diversion being more likely in familiar working environments where workers know the consequences of any deviated actions, whereas, in unfamiliar environments, operators are likely to revert to more of a skill and rule-based approach. Dekker (2014) argues that because of this ability, people should be seen as, “the solution” as opposed to, “a problem to control” within the safety system.
However, and of note, Rasmussen (1997) argues that the “limits of acceptability” of these deviations being an error, only, “if a standard of judgement exists and whether or not an act is judged as an error depends on the perspective and reference for judgement chosen”. Unfortunately, this can make it appealing to management to blame the person(s) operating towards the immediate causation end of the accident sequence when it suits (usually workload and time pressures), rather than having to identify, acknowledge and correct the more latent systemic conditions or even the failures of those in control of the safety system itself i.e. the originating influences. Few other industrial catastrophic disasters demonstrates this better than the Deepwater Horizon disaster of 2010 where, before the disaster, a health and safety audit “declared the project site to be performing excellently, at the same time as there were significant failings in the process safety elements which then led to disastrous consequences with multiple deaths and very significant environmental impacts” (Bly, 2010 as cited by: Gibb et al., 2014).
The highly competitive economic and commercial climate in which construction companies operate is well known, but the problem is that these same economic and commercial constraints very often “obscure” the investigation process, declared findings and remedial actions and, thereby, reducing the learning outcomes.
This type of top-down commercially driven organisational culture, influencing accident causation from a corporate perspective, was found by the courts to be major factors in the Bhopal, Flixborough, Zeebrugge and Chernobyl industrial disasters. In fact, it was the catastrophic Three Mile Island nuclear disaster that had a huge part to play in the development of complex non-linear models. In addition, Rasmussen (1997) argues that those in power will sometimes even use safety initiatives and safety campaigns to increase safety awareness under the “safety culture” banner to offset the commercial challenges! Although, those in power are unlikely to be willing to acknowledge this or, possibly, even be aware of this flawed perception if they do not fully understand the complex interplay of how the many variables interact dynamically throughout the safety system in accident causation.
This, combined with the ease of identifying the immediate cause(s), could also explain why it is so often the case that those in control of the system often think that accidents are caused by individuals unsafe behaviours rather than considering failures and/or latent conditions within the system itself to the point of ending the investigation when someone has been found to blame and where prevention strategies are often naively focussed around.
It is argued that the industry specific LCAC model has adopted many of Rasmussen’s (1997) socio-technical “behaviour shaping factor” characteristics and, therefore, qualifying as both a complex linear and complex non-linear accident causation model and highlights the complexity of the construction industry.
Although there is evidence-based research supporting industry specific accident causation models (such as the LCAC model) and the benefits of including a socio-technical approach to support the investigation process; there is a fundamental need for more targeted education and training within industry around the latest academic research findings with regards to accident causation, investigation methodologies and how to meaningfully address the shaping and originating influences identified during the investigation process to maximise the learning outcomes and mitigation measures implemented.
However, the significance of the economic and commercial demands and culture inherent within industry are often so dominant that they impact on the investigation process, findings and outcomes to the extent that it is suggested that industry is merely “masking” the originating influences.
Finally, until industry learns to better manage the very real influence that the economic and commercial factors has on all other downstream accident causation influencing factors as identified within the LCAC model (for example), the benefits of the industry specific accident causation models, including socio-technical models, are unlikely to be realised.
REFERENCES
Haslam, R.A., Hide, S.A., Gibb, A.G.F., Gyi, D.E., Pavitt, T., Atkinson, S. and Duff, A.R. (2005) ‘Contributing factors in construction accidents’, Applied Ergonomics, 36(4), pp. 401–415.
Cooke, T. and Lingard, H., 2011. A retrospective analysis of work-related deaths in the Australian construction industry. In ARCOM Twenty-seventh Annual Conference (pp. 279-288). Association of Researchers in Construction Management (ARCOM).
Heinrich HW (1931). Industrial accident prevention: a scientific approach. McGraw-Hill.
Heinrich HW (1959). Industrial accident prevention: a scientific approach (4th ed.). McGraw-Hill.
Rasmussen, J. (1997) ‘Risk management in a dynamic society: A modelling problem’, Safety Science, 27(2-3), pp. 183–213.
Reason, J. (1997) Managing the risks of organizational accidents. Aldershot: Ashgate Publishing.
Hollnagel, E. (2004) Barriers and accident prevention: [or how to improve safety by understanding the nature of accidents rather than finding their causes]. ALDERSHOT: Ashgate Publishing.
Gibb, A., Lingard, H., Behm, M. and Cooke, T. (2014) ‘Construction accident causality: Learning from different countries and differing consequences’, Construction Management and Economics, 32(5), pp. 446–459.
Horn, C (2017) A critical analysis into the factors that influence the accident investigation process in a construction contractor organisation. Unpublished evidence-based research project. Loughborough University.
Dekker, S. (2014) Safety differently: Human factors for a new era, Second edition. CRCnetBASE (CRC Press, part of the Taylor and Francis Group).
The Safety Conversation with SHP (previously the Safety and Health Podcast) aims to bring you the latest news, insights and legislation updates in the form of interviews, discussions and panel debates from leading figures within the profession.
Find us on Apple Podcasts, Spotify and Google Podcasts, subscribe and join the conversation today!
Really interesting article. As with all aspects of health and safety, lessons are never learned until money is the main contributing factor. Many employers just investigate, not to make things better but to cover themselves from future action. This needs to change! Investigations should lead to better workplace choices, no matter how small.
[…] Although the UK construction industry health and safety performance has improved over the years, it is statistically, still one of the largest contributors of workplace incidents, suggesting that lessons learned are not being adequately addressed. In this article, Clinton Horn (CFIOSH) considers some of the reasons why. Read more – SHP Online […]
I would hazard a guess the author has provided a precis of his disseration. Interesting albeit complex issues, especially the influence the client has on projects and accident causation. Clients rarely accept any responsibility or even assit, indeed in some cases they prefer to conduct the investigation rather than allowing the PC who is legally responsible for site safety.
The problem for the average practitioner is the time and resources needed to conduct a thorough investigation. There is a good argument, at least for serious incidents, that a specialsit in accident investigation should conduct the investigation to ensure impartiality inter-alia.