The Safety Pyramid
The Safety Pyramid
or Do You Really Still Believe in Santa Claus
Over the last month or so a leading safety group has been conducting a discussion thread around the question of safety facts and fictions. Their premise was that our world of Safety is full of half truths and magic mushrooms – significant decisions are being made whilst we are fumbling around within these mists of truth. For quite some time I have held similar views. Indeed one of the Safety Presentations developed a couple of years ago we called, “Myths, misdirection and misperception: there misapplication to safety outcomes”. It has always been a very popular and entertaining song ‘n dance. Has it impacted on organisations taking a more scientific view toward their safety interventions? Who knows………
Anyway, that is not to say we should not continue to consider some of these myths, and how they might be impacting on the manner in which we go about doing things. Regular subscribers shall have seen the position I have taken on LTIFR in the past. I don’t intend to revisit that discussion – past copies of Transforming Safety can be accessed via TransformationalSafety.Com. Related to it though, is the question that I often need to address, around a belief that the number of LTI’s occurring in a workplace are predictive of the likelihood that a major disaster or workplace fatality shall occur. This is certainly one of the most emotionally “believed” myths out there within our Safety World. Yet there is absolutely no truth to it.
Such is made more evident via a comment made to me recently:- “David….I don’t know what is going on here? We have been working on reducing our near misses and LTI’s and they are the lowest they’ve ever been, yet we continue to kill just as many people every year….” Sadly, we do know what is going on here! This very senior global CEO has been seduced by one of the myths of safety. There are enormous global safety consultancies who base their entire business model on this foundation. Some do it knowingly – and I consider that criminal.
I am reminded of a story I once heard. There were two builders who constructed some very impressive houses. They had all the “bells and whistles” you could think of. Everything and anything was available. It was like being on one permanent vacation. Unfortunately, there was a down side. One of the builders had not completed a proper geographical survey. Approximately twelve (12) months after completion there was a minor earthquake – one of those houses fell down around the ears of the owners. As it turned out the house had been built upon a foundation of sand. What we have here are two equally impressive homes, that from the inside, and out, appear absolutely fantastic. We can’t easily see the fact that one of them is founded on flawed foundations. No matter what “improvements” that the owner might make, they were doomed to failure, and all the money and emotional investment lost. The place was always going to fall down when the “system” was put under any significant pressure. Our safety systems are no different. If they are built upon foundations of sand, then we shall be forever battling against the entropic decay that the poor foundation accelerates. Oh, and by the way, the optimal geographical survey for a safety system is a recognised safety culture and safety leadership review of the practices going on within the “system”. (before anyone accuses me of plagiarism, the source for the above story is Matthew 7, 24-27).
So the myth we are concerned about has evolved, over time, via some very questionable quasi science (and I am being generous there). What I suspect you might be thinking about by now is this thing which has become known as the “Safety Triangle” or “Accident Pyramid”. There are now all sorts of derivations as different commentators have tried to put their own “brand” on it. In the example at right you would believe, based upon this Pyramid, that for every 30,000 at risk behaviours, you are likely to experience a major incident (other versions put fatality at the pinnacle), and so forth up the Pyramid. Nothing could be further from the truth. It is just a convenient nonsense – seems to make some intuitive sense (and that is what makes it so dangerous) – although there is no real science to support the intuition.
The implication is that there is a direct mathematical relationship between the number of near misses occurring on a site and the likely number of major incidents, LTI’s etc. At the very best, one must say this is misleading. At its worst we have to say that it is very dishonestly pedaled about. The higher up the Pyramid you go, the less valid is any relationship at all.
I take a slightly less confronting view. My own position is that there are so many “Safety Professionals” out there who have limited qualifications within safety science itself, that they too have been seduced into thinking there is something factual about the Pyramid. Even those people though, you would think, might express some surprise at the base ten multiples being so “perfect”?
To understand what is going on here we need to take a small history lesson. It was actually Bert Heinrich (Liberty Mutual Insurance) who presented the first Pyramid back in the 1930’s. Heinrich is considered the “father” of the BBS approach to safety. He came up with a pyramid that gave figures to risk exposure – what a surprise, he was an insurance guy. Further research over the years has shown that the make-up of the Pyramid, and even the shape of it, is unique to each and every organization – even despite the very fact that it is NOT predictive. In 1969 Frank Bird Jr modified the pyramid for who knows what reason – maybe he wanted his name in lights? Then, some years later, Connoco Philips modifed the pyramid – yet again. All equally snake-oil. Yet it is this very snake oil that has become the foundation of the traditional behavioral approach to safety management.
No two (2) companies share the same Pyramid. Bet no-one ever told you that? That alone should be ringing alarm bells. Indeed, the major factor which has been shown to impact the shape of an organizational pyramid is the prevailing safety culture functioning inside the business (there’s that ‘foundation’ again). That observation is not all that new, Bob Eckhardt made the comment as far back as 2003. He also suggested that the accident pyramid (like much of Heinrich’s early work) went unchallenged for so long, that it developed a sense of “fact” and “reality” around it. I would certainly agree with Bob!
Now another issue here is that any relationship between the variables in the Pyramid is VERY loose. It is actually a very weak correlational relationship; for those of the readership with any background in fundamental statistics, you shall know that a weak correlational relationship lacks any real “meaning” at all. Now allow me to highlight this by using an example I often present within particular safety presentations. What is the most dangerous sport in the world? What I mean by that is, what sport suffers the majority of deaths whilst participants are partaking of it? Likely answers; parachuting, high speed car racing, skiing, or bull riding. I suspect you could come up with a few more of your own.
The correct answer is Lawn Bowls. Yes, the pastime that Sir Francis Drake was playing when advised that the Spanish Armada had been spotted off the Elizabethan coast of England. I can already hear the silent howls of incredulous derision at this comment. Do the homework! This is a statistical fact based upon sound correlational mathematics. In fact, the mathematical relationship with regard to the lawn bowls example is much stronger than those which exist within the Pyramid. Why is it then that we so easily see the flaw in the lawn bowls example, and can’t see it in the Pyramid? That’s a much longer journey into the world of cognitive psychology for another day. Of course, the only reason that lawn bowls is such a “dangerous” sport is the general age demographic of those who invariably play it. So many people do die whilst playing bowls, but it is not the act of playing bowls that significantly contributes to the death. It is a wide range of other factors. As has already been noted above, with respect to accident causation the existence of a positive resilient safety culture is of far more value than the vagaries associated with the Pyramid.
At this point I shall briefly also address the relationship that seems to get the most attention. The supposed link between “unsafe acts” and “fatality” etc. A quite highly regarded safety guy (Fred Manuele) made the point in the Journal of Professional Safety, that “Many accidents that result in severe injury are unique and singularly occurring events in which a series of breakdowns occur in a cascading effect.”
Let’s add the words of another of the safety guys who I always pay attention to, the late Dan Petersen, who said “…if we study any mass data, we can readily see that the types of accidents resulting in temporary total disabilities are different from the types of accidents resulting in permanent partial disabilities or in permanent total disabilities or fatalities. The causes are different.”
So in essence some of the worlds foremost safety commentators, both current and past, suggest there is no relationship at all – and we need to get past trying to find a nice convenient button-hole to put it in.
Let us not forget these sorts of observations showed no relationship to the Texas City Disaster. Remember they had some of their best safety performance (based on Pyramid data) for some years prior to the explosion which killed fifteen (15) people.
Let us not forget these sorts of observations showed no relationship to the Deepwater Horizon Disaster. Remember they were actually celebrating seven (7) years of excellent safety performance (based on Pyramid data) at precisely the time the installation caught fire and killed eleven (11) people. I have heard it said they were actually cutting the cake when the place blew up. I suspect that is somebody trying to make a point. If it was, then it is well made.
By maintaining these myths within our safety systems we continue to be at-risk of blowing up the cake. Unfortunately the wider systems within which we operate have way too much invested in the status quo – and seem comfortable enough to just invent bigger and bigger sticking plasters (band aid’s). At the end of the day do you just want to patch the wound, or do you want to cure the illness.
The Risk Engineering Journal has an excellent article debunking these pyramids. To have a read just Click Here.
To read more about the author of this article just Click Here.
“Most organizations operate in failure states and that just remains invisible because bad stuff is not happening. We might call that the ‘normalization of deviance’ and, make no mistake, it will kill.”
“I was fortunate to attend Transformational Safety’s Anatomies of Disaster Program. This was amongst the most powerful two days I have ever spent in a room. From the outset David Broadbent set the scene by dedicating the program to the late Rick Rescorla – the man who is credited with saving over 2700 lives on 9/11. Throughout the two days David would often respectively reflect and remember those who had died, or been injured, in the disasters we explored. He would say, and I will never forget, “…we must always remember those that lost their lives lift us up into the light of understanding”. I learnt so much. HRO, Resilience Engineering, Critical Incident Stress Management (CISM) and more. Those of us who were there are still talking about it…… Thankyou David