How can safety be improved in environments where pretty much everything has already been prescribed, automated, controlled or fool-proofed? This is a question many organisations with well matured safety management systems struggle with. They have advanced reporting schemes in place, rich sets of lag indicators, numerous safety professionals that are skilled in hazard identification and the implementation of controls. Their people are well trained, properly certified, and subject to frequent reminders of safe work methods and the like. Yet, they continue to suffer safety incidents. What’s up with that?

Most of the organisations I speak with are bewildered about the unwavering flat-lining of their incident rates, and they scratch their heads about the rising cost of workers compensation claims. What else can they do? More training? More controls? More procedures? Perhaps, but where should they focus these efforts? The remaining incidents do not follow predictable patterns or any lead indicators, such as the number of safety inspections and observations. It seems the existing strategies have been emptied of their potential to drive further improvement. Yet a significant surd remains. An absurdity perhaps. 

I believe that this mystery to a large degree is based on an misunderstanding of what safety is. Safety is not something organisations have. The various components that make up a system, cannot be safe in and of themselves.  Put differently, safety is not a product or something static you can point at and say 'there it is'. By all means, parts, people and processes can be checked for their compliance with external standards, the structural integrity in relation to themselves, or the failure modes that they may be subject to. Take a ladder as an example. It can be checked for its compliance with technical standards international guidelines. It can be risk assessed for pinch points and structural integrity, and controls can be put in place accordingly. But does that make the ladder safe? It may make it reliable and up to scratch, and that way become a more predictable aspect of the environment. However, having a reliable ladder does not guarantee safety in the processes that involves the ladder.

When organisations operate as if safety is about compliance with standards, adherence to strict rules, or as if safety was something each component had, they inevitably get stuck in deviation management. Safety professionals turn into an internal police of sorts. Campaigns are held about the importance of complying. And the reporting systems mushroom with deviation data that no one seems to have sufficient time or knowledge to analyse meaningfully. 

Safety is something people and organisations do. Safety emerges from the interaction between various parts of the system. Whether the ladder is involved in success or failure depends on the activity it is involved in, the degree to which it allows users to achieve their goals without adding their spontaneous adaptations to the process, or how the ladder is exposed to things that were not predicted at the design stage. Safety, and incidents, emerges from the quality of the myriad of relations that any object, person or process is subject to during the course of getting things done. 

When organisations realise that safety is something they do - an emerging capacity to be successful across varying conditions - they change the questions they ask around safety. The question of who did what wrong, is gradually replaced by a need to ask and understand what people are dependent on to achieve successful outcomes, and what conditions and constraints make it difficult. And then they do something about it. The focus shifts from constraining people and processes, to supporting them.

For organisations that have depleted the potential of traditional strategies, changing how they understand and manage safety may be the only remaining option. This does not mean that they should stop verifying compliance or upholding training standards. It does, however, involve a sound scepticism that compliance and prescription is ever going to be enough. But to overcome, they need to think about safety differently. 

Comment