Study objective: Emergency clinician-performed ultrasonography holds promise as a rapid and accurate method to diagnose and exclude deep venous thrombosis. However, the diagnostic accuracy of emergency clinician-performed ultrasonography performed by a heterogenous group of clinicians remains undefined.
Methods: Prospective, single-center study conducted at an urban, academic emergency department (ED). Clinician participants included ED faculty, supervised residents, and midlevel providers who completed a training course for above-calf, 3-point-compression, venous ultrasonography. Patient participants had suspected leg deep venous thrombosis and greater than or equal to 1 predefined sign or symptom. Before any imaging, clinicians classified patients as low (<15%), moderate (15% to 40%), or high (>40%) pretest probability of deep venous thrombosis, followed by emergency clinician-performed ultrasonography. A whole-leg reference venous ultrasonography was then performed and interpreted separately in the radiology department. Patients were followed for 30 days. The criterion standard for deep venous thrombosis(+), required thrombosis of any leg vein on a reference ultrasonograph and clinical plan to treat.
Results: We enrolled 183 patients, and 27 (15%) had deep venous thrombosis(+). The sensitivity and specificity emergency clinician-performed ultrasonography was 70% (95% confidence interval [CI] 60% to 80%) and 89% (95% CI 83% to 94%), respectively, with overall diagnostic accuracy of 85% (95% CI 79% to 90%). The posterior probability of deep venous thrombosis(+) among the 88 low-risk patients with a negative emergency clinician-performed ultrasonographic result was 1 of 88, or 1.1% (95% CI 0% to 6%), and the posterior probability of deep venous thrombosis(+) among 14 high-risk patients with a positive emergency clinician-performed ultrasonographic result was 11 of 14, or 79% (95% CI 49% to 95%).
Conclusion: The overall diagnostic accuracy of single-visit emergency clinician-performed ultrasonography performed by a heterogeneous group of ED clinicians is intermediate but may be improved by pretest probability assessment.