Abstract / Summary
BackgroundPulmonary embolism (PE) is a common and potentially life-threatening condition encountered in acute medical practice.Computed tomography pulmonary angiography (CTPA) remains the standard imaging modality for diagnosis, but increasing utilisation raises concerns regarding radiation exposure, contrastrelated complications, and healthcare resource use.This study evaluated diagnostic pathways for suspected PE through a retrospective clinical audit and practitioner surveys conducted before and after an educational intervention. MethodsThis study comprised a retrospective clinical audit of 1499 consecutive patients who underwent CTPA for suspected PE, an initial cross-sectional survey of 48 healthcare professionals, and a repeat cross-sectional survey of 33 healthcare professionals conducted after an educational intervention.Collected variables included demographics, Wells score, National Early Warning Score 2 (NEWS2), oxygen saturation, heart rate, D-dimer results, CTPA findings, alternative diagnoses, and in-hospital mortality.The practitioner surveys assessed knowledge of D-dimer testing, use of age-adjusted thresholds, and factors influencing clinical decision-making.Of 1499 CTPA episodes, 1150 had an interpretable D-dimer result.Eight episodes without a definitive CTPA reference-standard outcome were excluded, leaving 1142 episodes for the diagnosticperformance analysis. ResultsPE was confirmed in 206 of 1499 patients, giving an overall prevalence of 13.7%.Among 1142 episodes included in the diagnostic-performance analysis, the standard threshold demonstrated a sensitivity of 87.8% (95% CI 82.3%-91.7%)and specificity of 14.8% (95% CI 12.7%-17.2%).The age-adjusted threshold demonstrated a sensitivity of 86.2% (95% CI 80.5%-90.4%)and specificity of 23.4% (95% CI 20.8%-26.2%).False-positive classifications decreased from 813 to 731, an absolute reduction of 82 cases and a relative reduction of 10.1%, while false-negative classifications increased from 23 to 26.The initial practitioner survey identified important knowledge gaps.Age-adjusted D-dimer was used regularly by 21 (44%) of respondents and occasionally by 13 (27%), while six (12%) were unfamiliar with the concept.Following the educational intervention, regular use increased to 18 (54.5%),occasional use was reported by eight (24.2%), and unfamiliarity decreased to approximately one (3%).Awareness of the local Ddimer assay also improved.Clinical decision-making was influenced by fear of missing PE in 32/48 (66.7%) of respondents, while 24/48 (50.0%) reported that clinical impression could outweigh formal scoring systems.In the initial survey, 29/48 respondents (60.4%) were unsure which D-dimer assay was used locally. ConclusionsRetrospective application of the age-adjusted threshold in this CTPA-selected cohort increased specificity and reduced false-positive classifications, with a small reduction in sensitivity and three additional falsenegative classifications.Because the analysis was not restricted to patients with PE-unlikely or non-high pre-test probability and did not assess actual imaging utilisation or follow-up safety, these findings do not establish the safety of age-adjusted D-dimer as a PE rule-out strategy.The repeat practitioner survey demonstrated higher assay knowledge and self-reported use; however, causality cannot be inferred from the unpaired survey design.