Ultrasound imaging obtained by emergency department physicians to diagnose deep vein thrombosis: accuracy, safety, and efficiency.

OBJECTIVES

To assess the accuracy, safety, and efficiency of ultrasound images obtained by emergency physician to diagnose deep vein thrombosis (DVT). patients with acute heart failure (AHF) attended in a hospital emergency department (ED).

MATERIAL AND METHODS

Prospective multicenter cohort study. We assigned patients suspected of having DVT to an intervention or control group. Emergency physicians took ultrasound images that were later evaluated by a radiologist in the intervention group. In the control group, images were evaluated only by the radiologist. We analyzed patient, physician, and episode variables. Test results, times until imaging, and 30-day adverse events were also analyzed. Sensitivity, specificity, positive and negative likelihood ratios, and agreement between physicians and radiologists (κ statistic) were calculated.

RESULTS

A total of 304 patients (209 in the intervention group and 95 controls) were included. The groups were comparable. The overall prevalence of DVT was 35.5% (95% CI, 30.3-41.0). The sensitivity of ultrasound images obtained by emergency physicians was superior in relation to experience: 71.4 (95% CI, 50.0-86.0) for those in a training course, 75.0 (95% CI, 80.0-95.4) for those with at least 2 months' practical experience, and 94.7 (95% CI, 82.7-98.5) for routine users. Specificity statistics for the 3 levels of physician experience were 83.3 (95% CI 55.2- 95.2), 100 (95% CI 83.0-100), and 96.6 (95% CI 88.4-99.0), respectively. The positive and negative likelihood ratios for ultrasound imaging by physicians were 27.94 and 0.054, respectively. The κ statistic was 0.80. Mean (SD) time until a physician took ultrasound images was 1.81 (1.46) hours versus 4.39 (1.81) hours until a radiologist obtained images (P = .007). Three deaths occurred within 30 days. They were not attributable to recurrence or bleeding.

CONCLUSION

Ultrasound images taken by emergency physicians to diagnose DVT are accurate and safe and may be efficient. However, routine experience with ultrasound is necessary.