However, I would shy away from using the 'accident chain' term in the
proactive sense because, for me, it seems to represent a linear process.
i.e that action A leads to B to C and so on, and if we can only break the
chain then the accident is prevented. Whilst this may be the case for a
single specific accident, it does not necessarily hold true in a systemic
sense.
Let me explain that better. What Reason's 'Swiss Cheese Model' and Sakuma
san's 'Rain drop model' illustrate is that just because you break a single
link in the chain does not mean that the latent failures cannot link up
with others. In terms of rain drops running down a glass window, if you
remove a single rain drop it may stop it linking with a drip of water, but
then the rain drop next to it may link up instead.
What I am trying to say (not very eloquently I know) is, that whilst in
retrospect accidents may seen to be a linear chain of events, in prevention
terms we need to look at multiple consequences of our actions.
It is not simply enough to recognise one accident chain and say 'I must not
do X to break the chain', individuals also have to be able to say 'but I
must also not do Y or Z because that may lead to another undesirable
consequence'. Take the Chernobyl accident as an example. To do the
experiment that was taking place at the time of the well known disaster
involved disabling some safety systems to avoid a shut down. Whilst a shut
down would not be an accident as such, it serves as an example of an
'undesirable outcome' to the original experiment. As the test went on,
other things went wrong and in the absence of the disabled safety systems,
the reactor went critical and failed.
Action A - disabling the system broke the perceived 'undesirable outcome'
chain as intended. However, other actions such as failing to stabilise the
reactor created a new accident chain (which the disabled system (Action A)
could had broken) leading to another, much worse, undesirable outcome.
Does anybody get what I am on about here?
In other words, as human components of any system, we are all in countless,
interwoven accident chains and so we muct not fixate on avoiding single
ones.
In aviation terms, it not just about preventing another Dryden, Kegworth or
Cali etc, it is about preventing all the possible variations on a theme -
aspects from every accident that has gone before.
What was the old cliche about no new accidents, only variations on a theme...
Sorry to go on a bit!
Graham Braithwaite
University of New South Wales
Sydney, Australia