The title of this Safety Moment is derived from a scene in Shakespeare’s Hamlet in which the hero disparages the meaning of what is written in the book in his hand.
Yet words do matter. And, in the context of a process safety program, some method is needed of deciding which of the many “events” that take place require a formal incident investigation. A catastrophic incident that leads to fatalities, serious injuries and property loss will definitely require that a full-scale investigation be conducted. But what about those events that are near misses or that had the potential to be much worse? Which of these should be investigated?
The OSHA standard says the following,
The employer shall investigate each incident which resulted in, or could reasonably have resulted in a catastrophic release of a highly hazardous chemical in the workplace.
Such guidance is of little help due to its inherent circularity.
- Should this incident be investigated?
- Yes, if it could it have caused a release of a highly hazardous chemical.
- Which means that we will have to conduct an investigation to find out.
Moreover, it is often the apparently trivial event that is of profound importance. For example, one of the worst incidents in the nuclear power industry was the Three Mile Island meltdown. One of the factors in that event was that an important reading on the control room panel was covered by a piece of paper and thereby overlooked. Does this mean that every time an instrument cannot be read a full-scale investigation needs to be conducted? Hardly. No company has sufficient resources to conduct a formal incident investigation on more than a tiny fraction of the events that take place.
What is needed is some means of defining the types of event so that they can be analyzed as to their impact. Provided below are some thoughts to do with words in the context of incident investigation and process safety management. By carefully thinking through the meanings of these words management and the process safety management team will be better able to judge which incidents deserve a full investigation.
- An incident investigation and analysis is conducted in order to identify ways of improving management systems in order to prevent similar incidents from recurring.
- An event is an out-of-the-ordinary occurrence. It may be a harbinger of bad news (say the failure of a pump) or it may be good news (say a new production record).
- An incident is an event that cause the system to cross a safety boundary as defined in Safety Moment #17: Safe Limits. An incident is always an undesirable event.
- A near miss — also referred to as a near hit — is an event that could have had serious consequences but that was prevented from causing further problems by the response of the instrumentation or the technicians. An example would be the failure of a level controller on a storage tank leading to a very high level in that tank that was noticed by the technician before the tank overflowed.
- A high potential incident is one that is just one step away from a catastrophic event (but this definition suffers from the circularity discussed above).
- A passive event occurs in the background and places the system in an unsafe condition, but nothing actually happens. An example would be the internals of a relief valve being plugged with process materials.
- The word accident should not used because the word implies surprise and lack of controllability. There is nothing anyone can do about accidents. The whole point of an incident investigation and analysis program — and indeed of Process Safety Management programs in general — is that all aspects of an operation are under control of management.
- The term root cause analysis means different things to different people, probably because there really is no such thing as a true “root cause”. An event has causes; those events have causes; those events in turn have causes; and so on and so on. Where to stop? Going back to Hamlet, But you must know your father lost a father, That father lost his, . . .
Where does the sequence of events end?
. . . .