ObjectivesAn evidence-based pathway for pulmonary embolism testing was implemented in two academic emergency departments as part of a prospective management study (the PEGeD study). This study aimed to identify factors associated with emergency physicians not following (deviating from) the PEGeD pulmonary embolism testing pathway. MethodsThis was a health records review of cases from the PEGeD study which enrolled emergency patients with suspected pulmonary embolism. Emergency physicians documented the Wells score on hard-copy PEGeD pathway forms which guided the use of diagnostic imaging. Patient visits were classified as having pulmonary embolism testing adhering to or else deviating from the PEGeD pathway. Patient data were collected from electronic medical records. We calculated adjusted odds ratios (aORs) for prespecified predictors of deviation: patient age, patient sex, arrival day of week, arrival time of day, documented hypotension, higher Canadian Triage and Acuity Score (CTAS) allocation, active cancer, and a history of venous thromboembolism. The multivariable logistical regression analysis was clustered by individual physician. ResultsIn total 1570 PEGeD forms were received, 78 were excluded and 1492 patients were included for analysis. The mean age was 55, 62% female, 27% presented at the weekend, 44% presented after 4 pm, 19% with cancer history, 13% with prior venous thromboembolism, 3% had a systolic blood pressure less than 100 mmHg and 46% had a CTAS score of 1 or 2. The treating physician deviated from the PEGeD pathway in 81/1492 (5.4%, 95% CI 4.4, 6.7%)) patients, of whom 7 were diagnosed with pulmonary embolism. Deviation from the PEGeD pathway was associated with a CTAS score of 1 or 2 (aOR 2.02; 1.26, 3.24) and prior venous thromboembolism (aOR 1.85; 1.04, 3.30).ConclusionsEmergency physician deviated from the PEGeD pathway infrequently. Physicians should question whether imaging is needed when D-dimer blood testing has already excluded pulmonary embolism.