Abstract
Studies which have attempted to demonstrate a relationship between perioperative anxiety and postoperative pain have yielded equivocal results. It is suggested that this is due, at least in pan, to an inadequate conception of anxiety which conflates several distinct, albeit related, components: the affective, the cognitive and the behavioural. By reformulating anxiety along these lines it becomes possible to generate testable predictions concerning factors which may influence postoperative pain response as well as allowing existing results (both positive and negative) to be explained.