Selecting hazards: Good and bad hazard choices and how to inform them

There are frequently passionate discussions within safety programs on what qualifies as a hazard or whether a particular event is rather a: cause, hazard, mishap, accident or outcome. Often there is no clear right and wrong answer but some hazards are better choices than others. These choices will ultimately influence the overall number of hazards a project will have to manage and the ability of that project to determine some overall measure of residual risk at the close. But the issue should not be a numerical consideration of whether 1000 hazards is more appropriate than 10, or whether either option indicates some relative degree of merit, but which set is a better choice and why? Two aspects, which should direct what ´hazards´ should be chosen are: the coverage represented by the totality of hazard, and the volume of hazards that have to be physically managed (and maintained) throughout the program and beyond. Too few hazards and there is a question of visibility of critical elements of the design, too many hazards and there is a question mark over the appreciation of the overall risk presented and distribution of that risk; an excessive volume of hazards will also contain some complex interrelationships between those hazards which may also prove problematic to manage. Bow-Tie diagrams are one method used to model hazards, but they are fairly one-dimensional and do not help in deciding which ´hazards´ should be placed as the ´centre´ of the bow-tie. This paper considers an alternative approach to modelling the accident-hazard scenario using Ishikawa diagrams which, by explicitly modelling the relationships between cause and effect events, inform the decision process on which events are a better choice to inform the safety program.