In a clinical judgement analysis, we used linear regression models to reflect the impact of clinical and non-clinical cues on priority decisions, by comparing the stated prioritization policies of 30 clinicians with their actual policies as revealed by an appraisal of 50 'paper patients'. Correspondence was modest for some cues, e.g. 25 doctors said they accounted for age, but age only had a significant bearing in the derived decision models of two doctors. Correspondence between the derived and expressed weights was greatest for clinical angina grade and the presence of left main stem stenosis. Correlation between the rank order of importance between the two models was poor for most of the cues, and statistically significant only for smoking. However, stated policies made it appear that lifestyle factors such as smoking habit would influence prioritization decisions for most clinicians but policies derived from actual prioritization decisions seldom related to lifestyle or demographic variables. There were significant differences in the degree of correlation between the two models according to the experience of the clinician. However, correspondence was not significantly better for doctors with cardiological training than those without. The overall contribution of demographic and lifestyle factors to decision making appears to be small, suggesting that they should be omitted from prioritization guidelines.