For ages now I've been longing to discover someone who is willing to put the intellectual effort into a practical way to support decision makers using risk and uncertainty concepts in work design and work health and safety. In short, I'd like to help more people discover the work of Terje Aven in Norway. Here's why.
Business as usual isn't
Over the last month I've been working closely with people in a very safety focused industry. Their current issue is getting to grips with the detail of forms related to risk assessment and how to rate risk. Everyone does this in WHS and we all get stuck sometimes, but the forms do eventually get filled in to varying standards. Sometimes there's a backlog when the tough questions get asked because someone has started to think about what's behind the rating. I was working with Greg on one of these forms and he was quite confident that an activity that scored in the risk category that included the stark word 'fatality' could be reduced to zero and therefore to a lower category. In Australia at least, the law, if someone dies, calls this a Category 2 offence under the WHS Act. If this was you as the duty holder or a contributor to this risk, a wrong decision here may put you in jail for 5 years or certainly result in fines of $300, 000. Suddenly ticking that box or signing off on risk becomes more difficult. Greg didn't appear to have that intuition, you might be less sure. In looking into Greg's discomfort with this, I've discovered the work of Terje Aven at the University of Stavenger in Norway. Terje has a fascinating series on risk and uncertainty in work health and safety which should be required reading for any safety professional (references below).
So what's the problem?
I am a massive fan of safety by design. As the daughter of an award winning engineer in robotics, for most of 50 years I've been watching the magic of great design and how that protects what people value. I'm passionate about our need as people who work and business owners to learn more about how to apply safety by design practice. That means understanding risk and uncertainty in design and in particular, work design. In a legal situation, Greg's confidence in his proposed risk controls implies it is feasible to have a situation which gives 100% control of all the conditions which might result in someone losing their life. So I asked Greg, would he be able to make a bet with me? Would he sell his house and take the chance that he was wrong just once about someone dying at work? Would he do that every day for the 5 years it takes for the likelihood of a 5 - 50 year risk to start kicking in? Suddenly Greg is less confident in sliding this C class risk to B class risk, from one that needs the highest level sign-off, to one which is closer to business as usual.
Uncertainty, risk and ALARP
To get to as low as reasonably practical, (ALARP) you first must deal with the uncertainty around the safety results of business activity and if you can specify it accurately enough, the risk around each and every activity in the workplace. You have to make your decision based on what you do and do not know thoroughly about the consequences and the 'chance' of those consequences turning up as an event or occurrence. The resources required to do that are part of another discussion but always finite. To get to ALARP you need to understand that there is not only risk or uncertainty around the activity, but also a personal risk around getting risk level wrong because not enough is known. Where there is uncertainty about activity consequences and therefore risk, there is a second type of risk coming from decision paralysis and the short and long term consequences for the decision maker of getting that risk level wrong.
Probability 101 is about exact repetition
You might have done probability in maths at high school or even at university. Or maybe not. The basic type of probability works on the assumption that the event which the maths is applied to is reproducible again and again. This is like Ground Hog Day, where every day the same thing happens again and again, the same way with no variation. Every repetition does not affect the next repetition. The cleanest way to calculate safety risk in this way is in a laboratory where every single repetition happens independently in the same way and the interactions between all the factors is kept consistent. To come up with a clear result, the exact results need to be known. If it is electrical safety that Greg is concerned with, then no one is going to calculate risk in this way.
Getting clear: uncertainty, risk and risk levels
Greg's form requires him to warrant that, through a variety of means, which add up to a perfect result, the event ‘fatality’ will never occur. Risk of death from the viewpoint of an electrician and his family must be fully controlled and prevented. For the insured business, the risk of death of an employee means partly what the family want, partly what the regulations require and partly what other stakeholders such as insurers and the community at large will accept as unfortunate but reasonable. Actuaries don't look at risk in exactly the same way the family or colleagues will. It seems callous, but insurers may pay up and move on.
Who owns the problem comes first
Imagine yourself in Greg's position. It looks like he only has to tick a box or write a word. He has to
1. take several views on risk, the view of the regulator, the business entity, the insurer, his boss's view, his own view and that of the worker who will is actually doing the activity.
2. work out how risk averse each of these parties are
3. assess his level of confidence that each of the players has used accurate and complete information on all of the different consequences
4. decide on an appropriate means to judge the likelihood or probability of each consequence or combinations of those occuring. These are tightly aligned with relative risk.
- risk aversion - how much would you be willing to pay to offset the cost of the event
- comparing the event to
Here are two of Terje's definitions:
Risk: uncertainty about and severity of the consequences of an activity OR the two-dimensional combinaton fo consequences (of the activity studied) and the associated uncertainties (what will the consequences of the activity be)
Uncertainty means that we do not know what the consequences of the activity will be or the value of unknown consequences