More seriously, lay definitions of the term “accident” imply that these events are somehow caused by blind or random chance; as if they were an act of God, a total fluke, a freak occurrence. In fact, this is often far removed from the truth.

Although every accident is unique, the study of accidents in a wide range of sectors has shown that there are certain commonalities, and it is therefore possible to formulate some clear conclusions about organisational and system safety.

One of the key findings has been that there is very rarely a single cause of an accident. A single act, event, omission or failure very rarely “causes” an accident. Accidents are complex combinations of multiple causal and contributory factors.

A causal factor is a factor that has a direct, one-to-one link to the accident. If a causal factor is removed from a sequence of events, the accident does not occur. A contributory factor is a factor with a probabilistic relationship to the accident – removing a contributory factor reduces the chances that an accident will occur, but does not directly prevent it.

Hence, accidents are not random, but are the culmination of a complex series of causal and contributory acts, events, omissions and failures. Some of the failures which cause an accident will be active failures, meaning failures at the front end of operations, which often have an immediate effect.

Others will be latent failures, “resident pathogens” (Reason, 1990), which lie dormant, unnoticed and unrecognised, until an unforeseen circumstance reveals the gap in the safety defence. Typically, the unforeseen and unpredictable interactions between events and circumstances mean that it simply would not have been possible to predict the accident in advance (Perrow, 1997).

In reality, if it were possible to predict the nature, outcome, location and timing of an accident with any degree of accuracy, then organisations and systems, and the people who work within them, would never have them in the first place. Unfortunately, the advantage of hindsight, which is conferred automatically on all accident investigators by virtue of their role, sometimes obscures this very simple fact.

In addition, luck does play a role in determining the outcomes of adverse events. It may be a cliché, but luck comes in two flavours: good luck and bad luck. The causes of an incident and a more serious accident may be identical, but perhaps because of a stroke of good luck, the more serious accident outcomes were averted.

As an example, consider a failure to the main rotor gearbox on a helicopter. If a helicopter’s main rotor gearbox fails a few seconds after take-off, then the outcome will probably be a premature landing, with minimal damage and minor injury.

The same failure, with absolutely identical causes, but occurring at 6,000 feet above the North Sea during winter conditions, will have very different and far more serious outcomes.
If it is merely providence which has prevented an incident from becoming an accident,

then there is clearly a case for conducting an investigation. “Incidents that by chance fall short of developing into major accidents should attract an equal intensity of investigation if they are to serve as sources of insight into causes and allow future accidents to be prevented that may not benefit from the same fortuitous chance” (RAE, 2005, p. 7).

Related post

No comments:

Post a Comment