Diagnostic accuracy of triage tests to exclude pulmonary embolism

We performed a study in 403 prospectively included patients with suspected pulmonary embolism to compare the accuracy of a combination of the SimpliRED D-dimer assay and an intuitive clinical probability estimate with either one alone. Based on a conjoint diagnostic refer, ence standard, including ventilation-perfusion lung scintigraphy and pulmonary angiography, the prevalence of pulmonary embolism was 31%. We demonstrated a high sensitivity (98%, 95% CI 95-100) and negative predictive value (94%, 95% CI 79-99) for the combination of the two tests. These figures were more favorable than for either test alone. The specificity of the combination was lower (11%, 95% CI 9-12) and consequently the proportion of patients in whom further diagnostic tests would have been avoided was only 8%. We conclude that the combined use of the SimpliRED test and the clinical probability estimate attains a higher sensitivity than either test alone. However, there remains a risk of false negatives and the exclusion efficiency is limited.