The effect of outcome information on health professionals' spontaneous learning

Objectives: Previous studies have demonstrated the presence of an outcome bias in medical decision making, which can be represented by a tendency to overweight outcome information when evaluating the quality of a decision. This study investigates whether the outcome of a previous decision on a medical case affects a later decision on a similar case. Methods: Thirty-six practising doctors and 36 nurses were presented with two superficially different but structurally identical dichotomous diagnostic problems, at intervals of about 6 weeks. The first problem was followed by the delivery of information about the outcome that could be either positive or negative. Results: Of the doctors, 39% of those who received information indicating an adverse outcome in the first case modified their diagnosis in the second case, whereas none of the doctors who were given information indicating a positive outcome made an alternative diagnosis. Of the nurses, 22% of those given information of a positive outcome and 56% of those given information of a negative outcome in the first case made an alternative decision in the second case. Conclusions: The results show that the outcome of a previous single case can be overweighted in the process of making a later decision about a similar case, to the point that information on outcome alone can modify a decision that the health provider originally thought to be optimal, according to his or her experience and the available evidence. Hence, outcome bias not only affects the evaluation of a decision, but can also affect learning by modifying later decisions.