ABSTRACT
Construction of a decision tree requires detailed information on the chances of the various outcomes occurring following implementation of any decision about clinical care. In addition, it is also necessary to know the sensitivity and specificity of diagnostic tests or the efficacy and safety of the various treatment strategies. A utility (value in terms of quality of life) has to be assigned to each potential outcome. This is highly subjective by definition as no two people will assign precisely the same utility to a given outcome. In an ideal situation, the patient should have an input into the decision-making process by discussing the utility that he or she places on each possible outcome. The expected utility for each branch of the decision tree (based on the estimated utility of the outcome and the probability that this will occur) can be calculated and hence the most optimal treatment strategy may be chosen (see example).
