Clinical experience and pre-test probability scores in the diagnosis of pulmonary embolism

Background: The Geneva and Wells pre-test probability scores are intended to replace empirical assessment of patients with suspected pulmonary embolism (PE). The effect of clinical experience on the inter-rater variability of these scores, and on empirical judgement, is unknown. Aim: To determine whether medical staff appointment grade affects the inter-rater variability of these pre-test probability scores, or empirical assessment, in patients with suspected PE. Design: Questionnaire survey. <Methods: Doctors were grouped by grade (mean number of years since graduation +/- SEM): house officers 0.7 +/- 0.2, registrars 6.3 +/- 0.6, consultants 25 +/- 4 and applied pre-test probability scores to actual case scenarios. Results: The Geneva score was the most consistent method of determining pre-test probability and was unaffected by clinical experience (Geneva kappa=0.73, Wells kappa=0.38, empirical kappa=0.23, p < 0.001 ). With empirical judgement, inter-rater variability was inversely proportional to clinical experience (house officers kappa=0.37, registrars kappa=0.24, consultants kappa= 0.16, p < 0.05). Discussion: The Geneva score was the least variable method and can be applied by junior or senior doctors. Using empirical judgement, junior doctors were more likely to agree on the pre-test probability of PE than were their more senior colleagues. This may imply that as physicians gain experience, they recognize that the diagnosis of PE can be difficult to assess and are reluctant to exclude it on clinical grounds.