When is a risk worth taking?
In one form or another, most asset expenditure decisions are at their core an exercise in risk management. When is a residual risk low enough to be acceptable without further mitigation? Prioritising the allocation of limited funds can require the use of quantitative risk management techniques to compare the relative safety improvement benefits offered by different projects or solutions.
The range of likelihoods available in a corporate risk tool may span about five bands, with the least likely, “Rare”, possibly defined as “less than 1 in 20 years”. Some risks are much less likely than this and these rarer than “Rare” events often cannot be prevented without very substantial investment. Such investments may intuitively be unjustified, however many engineers and managers are rightfully reluctant to accept a high safety risk on the basis of intuition.
The Work Health and Safety Act and Regulations require a person conducting a business or undertaking to do what is ‘reasonably practicable’ to ensure health and safety.
A duty holder must consider what is possible for ensuring health and safety and any possible measures must be carried out unless the cost is grossly disproportionate to the risk. One approach to determining whether or not the cost is grossly disproportionate is to apply the principle of As Low As Reasonably Practicable – ALARP.
For a risk to be ALARP, one must demonstrate that the cost involved in reducing the risk further is grossly disproportionate to the benefit. Infinite time, effort and money could be spent in the attempt to reduce a risk to zero but businesses need a framework to make decisions to spend a finite amount of money to achieve acceptably low, even if non-zero, residual risk. The level of such acceptability is set by society.
A recent example of an ALARP assessment arose when considering the need for replacement of a low voltage (400 volt) electrical switchboard located on a rotating clarifier bridge. The switchboard complied with the Wiring Rules in force at the time of its installation, however did not comply with the current standard. The non-compliance presented an increased risk to an electrician of inadvertent direct contact with a live part, leading to electric shock, whilst carrying out troubleshooting. Also, the frequency of interaction by electricians with this switchboard was much higher than usual, attributed to the age of the switchboard and associated equipment.
An assertion was made that if replacement of the switchboard would reduce the risk by reducing the frequency of interaction, and also due to the incorporation of the latest safety measures, then did this not mean that not all possible measures had been carried out unless the switchboard was replaced? It was necessary to establish whether the current risk was As Low As Reasonably Practicable (ALARP), and whether the cost of further reducing this risk by replacing the switchboard, would be grossly disproportionate.
Maintenance records established that there were up to ten maintenance interactions per year. The most conservative of three sources indicated that the lifetime risk of fatality for an electrician due to electrocution could be 2.3 per 1000 full time equivalent (FTE) workers. An hourly fatality rate for electricians was derived. An escalation factor was applied to allow for the increased risk of live troubleshooting vs typical “construction” electrical work.
Guidelines on the use of ALARP provide levels of likelihood of fatality that are categorized as Intolerable, Tolerable if ALARP, and Broadly Acceptable. These are depicted in figure 1.
Multiplying the hourly fatality rate by the historical annual frequency of maintenance interactions and by an assumed duration per interaction, gave a value for annual frequency of fatality which was very close to the lower end of the “Tolerable if ALARP” range shown in figure 1. Relevant Good Practice was being applied, and other Risk Reduction Measures had already been implemented. All that remained was to demonstrate that the cost of lowering the risk any further was grossly disproportionate.
The quantifiable safety benefit of replacing the switchboard was the reduction in frequency of interaction. This was translated into a notional cost-benefit using the published Australian mean value of a statistical life (VSL). The cost of replacing the switchboard was found to be approximately 1000 times greater than the statistical value of the safety benefit. Even allowing for the (lack of) sensitivity of this result to the various (conservative) assumptions made, or to the statistical data points selected, it was clear that the cost was grossly disproportionate to the statistical value of the risk reduction that would be achieved.
Whilst the example given may not appear as dramatic as the failure of a dam, or explosion of a major hazard facility, it is the type of scenario which may apply to dozens or even hundreds of discrete assets in a water utility. Grossly disproportionate investment in one risk may lead to lack of investment in addressing other risks which may in totality be more significant. Hence tools for ranking such risks are valuable.
Authors/Contributors: Craig Johnston and Peter Quinn