Risk assessment is supposed to be about looking at how likely it is that a person will come to harm, and then putting measures in place to reduce these risks to acceptable levels. As practiced in British Steel and its subsequent incarnations, the risk was assessed by looking at the likelihood of occurrence, the frequency of exposure, the number of people exposed and the degree of possible harm.
What I often found missing, however, was the probability that the maximum harm will occur, thereby leading to a skewing of the level of risk. For instance, at one time I had a frequent discussion with my boss about the holding of hand rails when going up and down stairs. While I don’t argue that holding the hand rail is a good idea whether you’re going up or down the stairs, I could not make my boss see that the level of risk was far greater when going DOWN the stairs, than going UP.
Theoretically it would be possible to state that the risk for both types of action was more or less equivalent if you take likelihood of tripping, the frequency of exposure, the number of people and the maximum level of harm to be the same. However, this line of reasoning ignores the fact that whilst the maximum level of harm, being death or severe disability, is the same for both types of action, the likelihood that you will be killed or severely disabled is far greater when you fall down the stairs than if you tripped going up the stairs.
I don’t know why I could not make my boss that going up and down stairs should be seen as two completely different types of action, each requiring their own risk assessment, rather than aspects of the same action, where the overall assessment of the risk is taken to be the highest of the whole activity.
Presumably this confusion may derive from a “better safe than sorry” approach, but sometimes I wonder whether it also leads to spending too many resources of staying out of trouble as a blanket rather than an activity-specific approach.