Pulmonary embolism: Are we there yet?

INTRODUCTION: Clinical prediction rules (such as Wells model) are a reliable assessment tool for diagnostic work-up of suspected pulmonary embolism (PE). When used as part of a clinical algorithm and in combination with a D-Dimer, the model can safely exclude PE in low-risk groups and indicate when further investigations are unnecessary. The purpose of this study was to investigate the level of adherence to local diagnostic imaging guidelines for suspected PE and to ascertain the impact of interventions. METHODS: Retrospective search of all patients referred from the Emergency Department (ED) of Royal Perth Hospital for computed tomography pulmonary angiography (CTPA) or V/Q scan between 11 September 2005 to 10 March 2006 (pre-intervention) and 1 January 2008 to 31 March 2008 (post-intervention) was conducted. The guidelines on 'Diagnostic Imaging Pathways' were considered as gold standard. Interventions included orienting ED doctors to guidelines and modified request forms for mandatory completion of Wells score. A prevalence- and bias-adjusted kappa (PABAK) score analysed the level of agreement between documentation on notes (R-score) and stamp (S-score). RESULTS: Thirty-five per cent (n = 187) and 22% (n = 109) deviated from the pathway pre-intervention and post-intervention, respectively (13% absolute reduction; P = 0.017). Stamp compliance was only 55% despite mandatory filling requirement. PABAK for 'PE as most likely diagnosis' was 0.25 for V/Q group and - 0.26 for CTPA. In addition, 44/60 (73%) had an intermediate or high S-score, yet only 11 of those 44 had a matched intermediate to high R-Score. CONCLUSIONS: Interventions reduced inappropriate practice but did not eliminate it completely. Compliance issues may be managed in the future via the introduction of electronic request linked to decision support.