Over the last two decades, safety improvements across a number of industries have largely flatlined (as measured in fatalities and serious injury rates, for instance) despite a vast expansion of safety investment, compliance and paperwork.
A typical factory floor riddled with daily safety challenges
The cost of compliance and bureaucratic accountability demands are mind-boggling with every employee working on average eight weeks per year just to be compliant.
It has also stopped progressing safety.
‘Safety differently,’ an approach developed by Sidney Dekker, is about halting or pushing back on the ever-expanding bureaucratisation and compliance of work. It sees people not as a problem to control, but as a resource to harness.
Today’s standard model of safety, systems are already safe and need protection from unreliable human beings? Sidney says that’s an illusion.
It’s not true that the only thing we need to do to make systems safer is to provide more procedures, more automation and tighter monitoring of performance. Emails from managers imploring people to stop making errors. Imploring people to follow the rules. Saying ‘if we just ask everybody to try a little harder, we’ll have a safe system’.
Sidney says what you need to do is to invert the perspective. Safety is not the absence of errors and violations. We need to see safety as the presence of something. Presence of what?
When you get into the messy details, what you see is under difficult circumstances people can still make things go right because of their adaptive capacity. Resilience is this people’s adaptive capacity. Resilience is the ability to bounce back. To accommodate change and to absorb disruptions without catastrophic failure.
Recent research back this up: the risk of fatalities and life-changing events hide in normal, daily routine practices.
John Green, Health and Safety Director at AECON discusses what Safety Differently is and why he feels there is a need for change.
The ‘Safety Differently’ movement views accidents like the BP Deepwater Horizon as too much focusing on near misses instead of critical issues, and so finds fault in Heinrich’s idea that minor accidents predict major ones.
Safety should be rather an ethical responsibility for people, assets and communities, instead of a bureaucratic accountability to managers, boards and regulators.
Safety Differently doesn’t just want to stop things from going wrong, but is curious about discovering why things go well and helping organisations enhance the capacities in their teams, people and processes that make it so.
According to Sidney, organisations looking to excel at safety must do the following:
Erik Hollnagel, approaching from a different perspective – health, points to a key distinction of ‘Work as imagined vs Work as done’ and how Safety-I is out of date and why we need to switch to Safety-II.
Safety efforts usually aim to eliminate or reduce unacceptable risk and harm. According to this definition, called Safety-I, a system is safe if as few things as possible go wrong. A problem with this approach is that safety management is based on evidence from random snapshots of failed system states.
Resilience engineering argues that safety should be viewed differently with emphasis on things that go well. According to this definition, called Safety-II, a system is safe if as much as possible goes well. Safety management and the understanding of safety should be based on a systematic understanding of how performance succeeds, rather than on how it fails.
According to Erik, Work-As-Done focuses on how people adjust their work so that it matches the conditions. Instead of only looking at the few cases where things went wrong, we should be looking at the many instances where things went right and try to understand how that happened.
We need to stop looking at problems in isolation. We need to stop using separate vocabularies, models, methods, organisational focus and organisational roles for each problem. This is the situation now with safety, quality, and profitability as examples. It is convenient in the short term but detrimental in the long term. We need to stop solving problems in isolation.
What caused a particular accident is not answered by listing things that would have prevented it. Erik founded the ‘Developing the resilience potentials’ idea digging deep into Safety-II, when a system is safe if as much as possible goes well similarly to Sidney’s ‘Safety Differently’ thematic but through different perspectives.
You must be logged in to post a comment.
Well, here we are again! Safety Differently is dangerous, as it is largely unproven. I am all for de-cluttering procedures, and processes that do not add to safety on the ground, but simply fill some bureaucratic void or perceived need. I am certainly all for stopping Serious Injuries & fatalities (SIFs) as argued in this journal previously in 2014 & 2017.
What is said here today about Heinrich’s Triangle is true: Focusing on the causes of Minor Injuries does not stop catastrophes or SIFs which tend to have entirely different precursors. Focus ALL your injury avoidance efforts at the top of the triangle and you will encompass the minor incidents as well as SIFs.
Hollnagel’s Work as Imagined (WaI) vs Work as Done (WaD) is nothing more than risk appraisal and assessment. Very simple and very clear, It is not magic or new.
What makes the Safety 2 movement dangerous, is that it is full of evangelicals trying to throw the baby out with the bathwater, based on UNPROVEN hypotheses. Nothing has been published anywhere since this movement started in 2014 to show it has had any effect on safety performance whatsoever in any country (i.e. Australia where it comes from, and the USA where the evangelicals preach, or anywhere else to my knowledge). In fact, in my view, it is the introduction of another useless fad, many of which have been introduced over the past 32 years that have taken peoples eye of the safety ball, and had absolutely no impact on the UKs incident rates at all.
So, in sum Caveat Emptor!
Mmm, turning a blind eye and expediently ignoring known known’s over the years sustaining a position of denial that fatigue has anything to do with more major errors, mishaps and accidents, on whatever scale, is not going to break-down the glass ceiling inhibiting progress either. “Work Exposure Limits” not just limited to animal vegetable or mineral – particulates and chemicals as, Biopschosocial suggests but, functionality in terms of preserving, conserving and restoring optimal and sustainable mental and physical wellbeing without addressing the repetitive stressors exhibiting in presenteeism, predictably foreshortening the working life-cycle of human resources, no further progress will be made.
On the subject of blind-eye’s, 58% of DSE operators still present, “carrying-on regardless” of their repetitive stress injuries, asthenopic or myopic adaptations in order to cope, tolerate and persevere poor accessibility of standard sub-optimally calibrated, adapted, operator customised or optimised screen ergonomics although, declared a Global Pandemic by the WHO.
It’s about time we started looking at safety differently and I believe this is a way forward. I do believe that there is too much focus on failure rather than success and that treating people like idiots is a folly . How many times do we have to see signs telling us to ‘hold the handrail’?
People keep people safe, not pieces of paper.
I do though disagree with the statement , “Heinrich’s idea that minor accidents predict major ones”. He never suggested any such thing. Only a correlation in numbers and their relationship. It’s this urban myth , encouraged over a hundred years, that has fed this mentality about stopping near misses and you will stop a fatality.
I can understand why under Author Bio no one has put there name to this. Heinrich as probably 99% of the readers of the article, and clearly not the author or editor pointed to the fact there was a correlation between minor incidents and the chances of experiencing a major one. For many practitioners this would be the foundations of culture – If you get the little things right, hopefully you wont have to deal with too many big things going wrong!
In the UK (& elsewhere) we are fortunate enough to have the principal of SFAIRP – the reason being you can have great processes in place but slot in an independent thinking organism (think AI) and it will adapt and find the shortest route available to complete a task given sub-optimal conditions (heat, light, fatigue, tools, comfort, mental state, management, supervision, personal & employer incentives – work as imagined vs work as done). During optimal conditions (99/100) things go to plan. Yes we need to change and adapt our approaches and we need pioneers who can do this and create evidence, to Dominic Coopers’ point. For those with a human factors background we have a lot of evidence in error tables, not eradication tables so we know we will never eliminate all error. But who is to say that the plateau isn’t just that? Have we gone as far as we can given all the best in human endeavour?
Most processes can be made safer with greater investment, some with less. So, is resolving the plateau problem less about human endeavour, practices and processes and more about financially viability – or, in common parlance – so far as is reasonably practicable?