Last week’s Alert (ACRS Weekly Alert No 584) contained a section in the Australia & New Zealand News section entitled ‘The similarities and differences between the Safe System and the systems thinking approaches’. This linked to the Safe System Solutions website and an Opinion piece entitled ‘Humans make mistakes: why we can’t abandon this core Safety System principle’ by the author, Dr Amy Williamson. The author raises objections to the ACRS Policy Position Paper: A New Systems Thinking Approach to Road Safety which was developed by a volunteer group of members (including me) and adopted earlier this year. The author supports the general thrust of the Position paper, stating that ‘Systems thinking offers immense value to road safety’ but is critical of the argument that road safety focus must include designing the system for human users and thereby reducing the role of human error.
The author’s primary argument is that the Safe System’s first principle, that humans are fallible, and error is inevitable, must not be abandoned. Now, the Position paper never advocated that it should be, although this principle is certainly not compatible with systems thinking. On the other hand, I argued that the principle should be modified to remove the concept of the inevitability of error in a paper published in 2021 in our Journal of Road Safety entitled ‘Why do we make safe behaviour so hard for drivers?’. Therefore, I felt it important to write a response to this Safe System Solutions Opinion piece to expand the explanation of why we must change the way road safety thinks and deals with human error and to correct some of the misconceptions evident in this Opinion piece.
The role of human error in the road system
The Safe System approach incorporates three basic principles. While there are multiple variants of the first principle, they all relate to human fallibility and that people, by nature, will make mistakes. The next two principles follow on, stating that humans are also physically vulnerable with respect to crash forces and that the road system must therefore be forgiving of human mistakes and ensure the consequences of crashes do not include death and serious injury. These Safe System principles are now extremely well-known and incorporated in almost all road safety documents of note.
Road safety action based on these principles invariably leads to strategies that forgive or mitigate the consequences of crashes. This is helpful but should not be all we do. Following the Safe System principles means we fail to acknowledge that crash prevention should also be a road safety target. The author of the Opinion piece goes further and argues that prevention of human error and crashes ‘is a monumental task’ because error is inevitable, so prevention is neither appropriate nor practical for road safety.
Unfortunately, this view suggests a lack of awareness of current research and thinking on human error and how it is incorporated into safety management in related transport sectors. In the next section, my objective is to summarise the most important points that are missing in the Opinion piece and explain why we need to put aside the Safe System principle that the human road user is fallible and error inevitable.
Why we must not ignore the role of human error in the road system
- There is considerable discussion about what is human error in other sectors of safety, but road safety sees error broadly as a human behaviour that leads to harm including crashes, serious injury or death. For example, the behaviour of overtaking a vehicle on one of our divided higher speed roads can be performed successfully and safely in most instances. It is a straightforward behaviour, you (the driver) make the appropriate checks of the adjacent lane, you signal intent, move into the next lane when it is safe and after passing the vehicle return to your original lane, again when it is safe to do so. But the context in which you do the overtaking behaviour is not always the same. What if you don’t do a head check when you make the manoeuvre? Or in doing the mirror check, you are not aware, or forget, the vehicle has convex side mirrors which make the approaching vehicle appear further away than they actually are? What if the overtaking manoeuvre is in response to the vehicle in front braking unexpectedly and you only have time to do quick checks? What if, in merging back into the original lane, you use only information from your sidemirrors and fail to check your rearview mirror?
In any of these contexts, behaviour might be judged an error if it results in a crash involving an injury to you or someone else. In each case, however the behaviour was influenced by the context, ranging from poor or forgotten driver training (failing to do head checks), misleading information from vehicle features (convex mirrors) and/or unfamiliarity with those features, and an unexpected change in the traffic movement. The context defines whether the behaviour is an error. This simple example illustrates how the context of performance is crucial to understanding behaviour and especially to understanding whether the behaviour should or should not be judged to be an error. There is a multitude of similar examples in the road system.
- Human error is a reality in the road system, but it is not as common or inevitable as the first Safe System principle leads us to think. In fact, humans have the capacity to process information to high levels and generate appropriate and safe performance even when circumstances are challenging. This is how we manage to have millions of safe trips on our roads every day with few errors and comparatively few crashes. Error while using the road system is certainly not inevitable.
- The problem is that circumstances in our road system, through their design, can make it difficult for drivers and other road users to always achieve safe behaviour. These are circumstances that promote error through confusing, distracting or overloading the road user so increasing the likelihood of error. The paper I wrote for the Journal of Road Safety, mentioned above, summarised some of the poorly designed roads, vehicles and road rules and enforcement which increase the likelihood that error occurs. For example, where drivers can’t see a speed sign due to its location or to visual clutter around it or can’t read a navigation sign quickly enough due to the complexity of the information it contains or its placement too close to a decision-point, they will not respond as road designers expect. In these circumstances, the crash that occurs when the driver was driving above the speed limit because the speed sign was too difficult to locate, or who make an unsafe manoeuvre like a quick turn, so they do not miss their turnoff, should not be attributed to driver error. Rather these crashes are due to poorly designed signage. Fix the signs and error will reduce.
- Secondly, when we treat human error as inevitable, we also deny the existence of system-generated error and, not surprisingly, take little or no action to prevent it. Thus, these types of errors persist and even though they are the product of the way we do things in the road system, we treat all error as if they are road user generated. We must base our safety action on understanding how to design the road system to maximise safe behaviour by users of the road system. By doing so, we will significantly reduce the risk of human error.
- Further, if we assume that nothing can be done about error or that it is just too hard, we never seek out the reasons for error. This takes us into a dangerous spiral where assuming that errors will always occur, and that nothing can be reasonably done about them leads us to do nothing to prevent them and as they become sufficiently common, we see them as inevitable. This spiral must be stopped.
Our current approach to analysis of the causes of crashes has also led us down this path. Crash analysis invariably cites just a single cause, usually the road user behaviour just before the crash, which is interpreted as error mainly on the basis that a death or serious injury had been the outcome.
- A considerable amount of research has been done on error in other settings. This research acknowledges that error takes different forms and that judging whether or not a particular behaviour is an error will depend on the context in which it occurs. Other sectors of safety treat error as an attribution which is prone to various types of bias and acknowledge that safety failures (crash in our language) involve multiple contributors, with human behaviour often the last, just before the failure. Evidence from other sectors shows that often the behaviour was normal or reasonable, but the circumstances meant that behaviour led to the failure. We have not learned from these other sectors and until we do, we are doomed to be on this dangerous spiral.
- The author makes the case that road safety cannot learn from other sectors like aviation because it is too different. This is simply not correct. For a start, the aviation sector, like road and maritime transport, but not rail, comprises both highly regulated and lightly regulated sub-sectors. Commercial aviation is highly regulated at all levels, whereas general aviation is much less regulated and unfortunately has significantly less happy safety record. All these sectors share the same basic problems of managing risk or the likelihood of an unexpected, undesired event causing harm. All need to understand the contributors to safety failures in designing strategies to improve safety.
I have worked on safety in all these sectors. Undoubtedly road safety currently deals with safety differently to other transport sectors. Safe System and its principles would not be seen as sufficient in these other sectors. Systems thinking, usability, user-centred system design are fundamental to safety in other sectors, but not road safety. If anything, the road sector represents a higher risk than these other transport sectors due the millisecond level opportunity to intervene to prevent crashes. We are crazy if we don’t learn from other sectors.
The way forward
Road safety must continually review and challenge itself and to guard against complacency. We must not ignore the current increasing trends in road deaths and hospitalisation for serious injury. These are telling us that we need to think again. It is not enough to argue, as in the Opinion piece, that we just have to try harder to implement Safe System. The direction that Safe System has sent us, that of mitigating the consequences of crashes, is not wrong, but it is certainly not sufficient. Systems thinking approaches will help us broaden our search for solutions and will also acknowledge the need to understand the circumstances that cause human error in our road system and address these circumstances through effective user-centred design.
Obviously, the issue of human error is much larger than I have room to discuss here, but it is an important one for road safety. In recognition, the ACRS has in train events and resources to help members expand their knowledge in this area. Specifically, this includes a Webinar on ‘What is error? The role of human error in the road system’ to be held on 25th March, 2025 and there will be a significant section on the road user and the role and management of human error in the Road Safety Body of Knowledge currently being developed by the ACRS. More immediately, the Members Forum of the ACRS website has opened an area for members to exchange ideas and views about human error and systems thinking. Please feel free to post your questions or views.
In summary, we must dispense with the first principle in Safe System as it is fundamentally incorrect. People do not always make errors and are not always fallible, but they become fallible when we set up the road system to make it confusing, distracting and overloading for road users. We must expand the focus of road safety to include action to eliminate or reduce as far as possible human errors that are brought about by poorly designed elements of the road system including roads, vehicles, rules and enforcement and training. This is a collective responsibility. We need to work to eliminate design deficiencies in the road system that make driving difficult especially during the performance of activities critical for safe driving and, that in turn, cause them to crash.
Prof (Em) Ann Williamson
ACRS President