Is safety pyramid a myth? Study suggests new approach to injury prevention

Injury statistics in recent years revealed that while minor injuries have declined steadily, serious injuries remained the same, casting doubts on the validity of the Heinrich Safety Triangle concept. A new study indicates a different model for injury prevention — going beyond the numbers.
Large companies seeking to reduce serious injuries and deaths must spend more time looking for potential dangers hidden in their workplace procedures, a new study says.

The study, by Behavioral Science Technology (BST), a safety management consulting firm, says the risks at many worksites can only be discovered, and corrected, by systematically reviewing work methods and analyzing lost-time injury data.

The study comes a year after BST chairman, Thomas Krause, told a conference of the American Society of Safety Engineers that the traditional understanding of the relationship between minor and serious injuries, embodied in the Heinrich’s Safety Triangle, may no longer be valid.

The safety triangle, or pyramid, states that reducing minor injuries (the large number at the bottom) leads to a proportionate reduction in severe injuries and deaths (the small number at the top). However, statistics over the past 10 years have shown that minor injuries have steadily declined while the number of serious injuries and deaths has not changed.

“The assumption that you could reduce injuries at the bottom and see at the same time reductions all the way up to the top is flawed and turns out not to be the case,” Krause says.

Similar injuries, different causes
The assumption, he adds, was based on two mistaken notions: that all mishaps, whether minor or serious, have similar causes and that all lost-time injuries have the same potential to be serious.
 
In fact, similar injuries may have completely different causes, and accidents differ greatly in their potential to be serious, Krause says. For example, a person falling off a railcar may end up bruised. It’s a minor injury, but the potential for a fatality is very high. Another person, on the other hand, may get a bruise by bumping into someone in the hallway.

“It’s the same bruise, and the data will look identical if you just look at the numbers. It will look like a bruise, but the events are radically different,” he says. “It doesn’t hold that if you prevent people bumping into each other in the hallway, you’re going to reduce fatalities.“

So when a company simply tracks the number of lost-time injuries, he adds, it’s impossible for anyone to tell how many of them could have been serious.

The BST study suggests a new paradigm, one based on the notion that serious injuries have different underlying causes than minor ones. These underlying factors — missing controls, lax procedures, badly designed equipment — create high-risk situations that are likely to lead to a major incident.

Thus identifying and addressing these high-risk situations, or “precursors,” is the key to preventing major accidents. A precursor is any high-risk practice that has not been recognized and corrected. It could, for example, be a safety control that is routinely ignored. In such a case, the company could go for years with very low lost-time injury rates. Then a worker is killed.

“We’re blind to the indicators of fatality. We’re looking at the wrong information,” Krause says. “And that scenario — of good recordable rates and then a big event that’s bad — happens over and over.”

To identify and address precursors, companies need not only to examine their procedures, safety observations and audits but also to analyze incident data to distinguish between the small number of incidents that had the potential to be serious and all the rest, which did not.

“Being able to tell that makes all the difference in the world, because if you’re having a lot of events that are potentially serious, the likelihood that you’re going to have a serious event is very high,” Krause says, adding senior executives often wrongly regard a serious or fatal incident as a fluke.

“If you randomly take a fatal event in an organization, the possibility that the circumstances leading up to that event never happened before but just happened today — it’s possible, but it’s probably one in ten thousand. The greatest likelihood is, when there’s a serious event, it means something.”

Beyond lost-time injuries
Yan Lau, senior manager of OHS policy, employment and immigration with the Government of Alberta, says safety statistics in the province show a similar trend as in the U.S. — from 2006 to 2010, the lost-time claim rate fell by 37.3 per cent, while during the same time fatalities rose from 124 to 136.

He agrees with Krause that injuries differ greatly in their potential to be serious. Most injuries are caused by over-exertion and affect the musculoskeletal system, he says.

“They’re sprains and strains, and the majority of those could not end up as fatalities. Then there are some types of hazards that quite often end up in serious injury or fatality, like a fall from a height or contact with heavy equipment,” he says.

Lau supports the study’s recommendation that companies track the number of events that they judge could have been serious or fatal as a metric separate from time-lost injuries.

The BST study was conducted in partnership with seven major companies, including ExxonMobil, PotashCorp, Shell, BHP Billions and Maersk.

Robert Whiting, senior project manager and subject specialist at the Canadian Centre for Occupational Health and Safety, says the study serves as a reminder to managers not to be complacent and not to rely simply on a low injury rate — a mistake often made because they are focused on the cost of lost-time injuries.

“A lot of it has to do with claims management because that’s where a lot of the money is. Serious accidents can have a huge cost, but if they’re rare, it may be tempting not to pay as much attention to them. But the law says you have to consider every hazard, and they don’t differentiate between serious ones and less serious ones.”

Whiting says major and minor incidents often differ in that a small mishap usually has a short, simple chain of causation — leave oil on the floor and someone slips. In contrast, the precursor steps of a serious incident are usually much more complicated.

“The cause may be hidden in the process. There may be flaws or limitations hidden in the process, or it may be that, as with some airplane accidents, slight changes in the design of a control system will change the chance that somebody will make a mistake,” he says. “These things are subtle, and they take a long time to find.”

And for companies where serious accidents are rare, identifying the cause is likely to be very difficult, Whiting adds. Where severe accidents are more frequent, a safety manager may be able to detect a pattern and determine cause. “That’s why you need to do a deeper analysis.”

Lau says that the two models take different approaches and neither is completely correct. The message of the safety pyramid is that companies should not dedicate a lot of resources to preventing the relatively few events that result in serious injury when there are so many ways to control total incident loss.

“The new model (high risk exposure — precursors — outcomes) is an example of using the common ‘logic model’ to link input to outcomes for explaining the occurrence of serious injuries.  Both models serve a purpose of explaining some observations,” he says.

Lau agrees with the study that more research is needed to determine which safety performance measure will best indicate workplace safety.

In addition to systematically identifying and addressing precursors, the study concludes, staff at all levels must be educated on the lessons of the new paradigm and learn to recognize that a good safety record based on lost-time injuries is not a sure sign of a safe workplace.