Abstract

If we are to believe the great medical teacher William Osler that medicine is a science of uncertainty and an art of probability [1], then errors in judgement are inevitable and it is incumbent on those involved in clinical research, teaching and practice to seek ways of handling them. All too often, the behaviour of clinicians tends to suggest that they are neither aware of the implications of the inherent errors in their measurements or observations, nor are they always honest with themselves in the confidence expressed in their decisions [2]. Decision analysis provides a framework for handling uncertainties but its effective use in the clinical sciences requires not only a knowledge of the nature and function of clinical tests but also of the sources of potential error.

© 1993 Optical Society of America