Learning Curve of Intensivist-Performed 2-Point Compression Ultrasonography for Diagnosis of Deep Vein Thrombosis in the Intensive Care Unit

Abstract

PURPOSE: Physician performed 2 point compression ultrasonography (CUS) for diagnosis of proximal deep vein thrombosis (DVT) has had reasonable accuracy when performed in the emergency department and may be useful in an intensive care setting. The accuracy of this test is subject to operator experience and training. We assessed the learning curve of intensivist performed 2 point CUS in the ICU. METHODS: Study physicians included 3 critical care fellows and 1 attending, who received an hour training in lower extremity CUS. The study included adult ICU patients with lower extremity ultrasound ordered for suspicion of DVT. Protocol consisted of identification and compression of femoral and popliteal veins at a single point. Results were reported as compressible (negative), non-compressible (positive), or indeterminate and were compared to formal study performed and interpreted by radiology department. Radiologist interpretation was used as the reference standard. Learning curve of the study staff was evaluated by comparing the test characteristics of the first 45 tests vs. the latter 47 tests. RESULTS: A total of 92 studies were performed on 85 patients (5 had two studies and 1 had three studies). The study population was grouped into two categories: early group (first 45 tests) and the late group (latter 47 tests). There was no difference in baseline characteristics between the 2 groups. The early group had 7 DVTs with 1 false positive and 2 false negatives giving a sensitivity of 71%, specificity of 97%, and an accuracy of 93%. The late group had 7 DVTs with 1 false positive and 1 false negative giving a sensitivity of 86%, a specificity of 98%, and an accuracy of 96%. Agreement (measured by kappa) was better in the late group (0.83 vs. 0.73). CONCLUSION: Our findings demonstrate a clear learning curve for intensivist performed 2 point US to detect DVT. CLINICAL IMPLICATIONS: Training and volume are necessary to improve diagnostic accuracy in intensivist-performed 2 point US.