Errors: An Inherent Part of Human-System Performance - Foundations for Designing User-Centered Systems - page 283

Fig. 10.1 The influence of
the blunt end on the sharp end
of an incident

Blunt End

Regulatory

Organizational

Teams

Operators/
Users

Sharp End

Technology

It is only when you start to give deeper consideration to the circumstances in
which the error occurred that you really appreciate how hard it is to decide who (or
what) is really to blame. If a doctor makes a medication error after having worked
for 24 h continuously, for example, can we really blame the doctor? We know that
fatigue adversely affects cognitive performance, yet the system still put the doctor
in a situation where they were likely to suffer from fatigue. Similarly, we know
that the way that system controls are laid out can either help or hinder perfor-
mance. If a design is inappropriate, can we really blame the users?
In fact, the actions that are performed at the lowest level (usually where the user
interacts with the technology) are only a small part of the picture. This level of
interaction is often referred to as the sharp end of the system. Actions at the sharp
end are often influenced by what happens at the so-called blunt end of the system,
as shown in Fig. 10.1 . The idea of a sharp end and a blunt end comes from Woods
et al. ( 1994 ). This figure illustrates that final users are often seen as causes at where
users meet the task, but that there is a lot of structure behind them that influences
the situation as well; structure that is harder to change but that has a large amount
of influence.
Decisions and actions taken at the regulatory level can affect what the operators
do. In the USA, for example, the FAA's regulations state that pilots must not perform
deliberate violations, unless the violation is needed to put the aircraft into a safe
state. Similarly, standard operational procedures (which are generally defined at the
organizational level) usually define what the operators can (or cannot) do.

10.2 Studying Error

The issue of data collection is fundamental to the study of human error. If we could
reliably predict exactly when an error was going to occur, it would be a simple
matter to warn the person or people involved so that they could avoid it.