Wednesday, 21 May 2014

It's all human error

Over the past 12 months, human error has been responsible for:

The term "human error" has become a catch-all for a smorgasbord of occurrences such as unintelligible hand-writing, fatigue, double-counting, pressure to under-report, etc. However, the commonality of these news reports is that there was a human at the sharp end and a suggestion that "human error" is the counterpart of "technical error" or "equipment failure".

One of the problems with using the term "human error" is that "equipment failure" is often also "human error" separated by space and time. For example, if an engine fails on an aircraft (a technical error) then one could trace this failure back to poor engine design or manufacture or installation or maintenance, all of which are "human error". Seen in this light the only true non-human causes of accidents are "acts of God" such as earthquakes or tsunamis. (Even with the latter, the (human) decision to lower the Fukushima nuclear plant from 35m to 10m above sea level contributed to this post-tsunami disaster.) Technical or equipment errors are also strongly implicated in many human errors.

Thiopentone and Co-amoxiclav (or perhaps the other way 'round?)
For example, drug manufacturers are at liberty to package their drugs in whatever vial colour and design they choose. This leads to an increase in the number of people getting the wrong drug. New medical devices (pumps, anaesthetic machines, defibrillators, monitors) are not stress-tested in conditions where error is likely. These design errors lead to human error.

Death knell for "human error"?

Does the term "human error" serve a purpose? At a very coarse level it may tell us if there was a human at the sharp end, but other than that it suggests an unwillingness to delve deeper into the accident. It also implies that if we could get rid of the "human", by increased automation for example, systems would be safer. The term "human error" should therefore be replaced, not by one new term but by the correct (as far as is known at the time) term(s) which best explain the accident causation. This may be performance variability and/or organisational variability and/or equipment design...

Further reading

A very well-written post on human error by Steven Shorrock at  "The use and abuse of ‘human error’"

No comments:

Post a Comment