I’ve recently published an article on the application of systems thinking concepts in accident analysis. The article is published in Safety Science available via ScienceDirect. The paper includes a discussion on the Bellevue fire in Western Australia.
Anyone interested in the article can contact me via y.goh @ curtin.edu.au (remove the spaces adjacent to “@”)
Rasmussen (1997) mentioned that one of the key things in accident prevention is to identify the “behaviour shaping mechanisms”. What causes persons at the front end to do what they do to cause accident? Why was it logical to do things that they do?
To me the idea is related to safety culture. One of the most useful definitions of culture is the way we do things around here. If that is the definition of [safety] culture, the way that things are done will affect how individuals do things, be it safely or unsafely.
Another related concept is Senge’s (2006, 1990) concept of how systemic structures affect patterns of behaviour and how patterns of behaviour influence events. As highlighted above patterns of behaviour is the observable portion of safety culture. Then what is systemic structure in the context of safety culture. What shapes safety culture? I believe it consist of a wide range of factors due to internal and external stakeholders. Internal holders will be mainly employees and management, it will also be influenced by the company structure and procedures. External stakeholders will include customers, suppliers, competitors, substitutes, regulators, NGOs, trade associations, unions… The relationship and interactions of these stakeholders across time will have an impact on the safety culture of the company… extremely complex. This is what Senge calls dynamic complexity… something systems thinking is supposed to help tackle…
If you have worked in a large organisation before you might have had the feeling that the organisation is frequently implementing rules, buying new safety equipment, and getting staff to go through certain safety training to help improve the safety performance of the organisation. However, how many of such “fixes” actually improve safety performance? Take a look at this website for a system archetype called “Fixes That Fail”.
These fixes or interventions introduced by the organisation are often not tackling the fundamental problems that are producing the safety-related issues that the fixes attempt to solve (usually quickly). The problem can be the unsafe behaviour of sticking air hoses into overalls to cool the body while doing hotwork (possibility of sticking the wrong hose, e.g. oxygen hose!). However, instead of tackling the more fundamental problem of poor ventilation, the organisation choose to reprimand workers and introduce fines to prevent such unsafe behaviours (quick fix). That being said, in safety and health, where lives are at stake, a lot of time both quick fix and fundamental solutions need to be implemented simultaneously. We just need to ensure that we don’t get lost in the quick fix.
Behaviour based safety (BBS) is widely accepted as a methodology to improve occupational health and safety of an organisation. It applies behavioural psychology to measure behavioural trends and design intervantions. Some organisations treat it as an advance tool that should only be used when the safety culture of the organisation is mature. This is because successful BBS programmes must involve employees.
However, there had been concerns that the BBS approach might be “blaming” workers as the reason for accidents. Click here for an example of such concerns.
My view is that any programme can be twisted to shift blame to workers if the organisation is not sincere in improving safety. Hence, the problem is not BBS per se. When we use BBS, we have to apply one of the basic principles of human factors, which accepts humans as imperfect beings and it is how we design equipment and tasks to suit the humans to improve safety. Measuring behaviour gives us a handle on the situation so that interventions can be designed appropriately.
The measurement of behaviour can also be used to plot behaviour over time (BOT) plots, which is one of the basic tools in systems thinking to interpret patterns of events. With the BOT, organisations can then surface possible systemic issues that promotes unsafe behaviours. When we are looking at systemic or structural issues it is no longer about any individual or party, it is how the different components of the system interact to produce the safety performance of the organisation. There is no blame involved.
I developed a series of steps for causal analysis during accident and incident investigation. The main steps are listed below: