Attitudes, access or application: what is the reason for the low penetration of decision support systems in healthcare?

The bulk of this paper will deal with the results of a survey which we have undertaken among clinicians. However, for the lessons of that survey to be learned, we have to understand the reasons
why computerized decision support systems are desirable in the first place. There is no space in this paper to discuss those reasons, but we wish to set the scene by asserting that the three major factors are:
1. The huge knowledge explosion in healthcare (the half-life of a paper is about 8 years);
2. The complexity of the decisions facing clinicians and policy-makers (simple decisions typically involve choosing one course of action from a menu of millions of courses of actions);
3. The limitations of unaided human information processing (Miller’s Magical Number 7 ± 2-).

There is a large literature on sub-optimal decision-making (and even serious errors) by clinicians that shows that what one would predict in theory actually occurs in practice.

We do not assert that using computers with their large memories, powerful inferencing mechanisms, consistency and general reliability is the only way to improve health decision-making. We
do, however, assert that this is currently the most likely candidate for doing so.