Overuse and practitioner variability in ordering patterns of Doppler ultrasound for the assessment of deep vein thrombosis in a tertiary care center

Introduction: A low probability Wells score in combination with negative D-dimer testing can exclude proximal deep vein thrombosis (DVT) with high sensitivity, reducing the need for ultrasonographic investigation. In properly... [ view full abstract ]

Introduction: A low probability Wells score in combination with negative D-dimer testing can exclude proximal deep vein thrombosis (DVT) with high sensitivity, reducing the need for ultrasonographic investigation. In properly selected cases, the positivity rate of venous Doppler ultrasound for proximal DVT approaches 15%. The objective of this study was to determine whether patients were being optimally selected for ultrasound in the work-up of proximal DVT and to describe practice variability within our institution

Methods: We retrospectively reviewed all lower extremity venous Doppler ultrasounds performed at our university affiliated hospital in Montreal, Canada from September 2014 to February 2016 for evidence of acute thrombosis. When performed, D-dimers were cross-referenced with patient identification and time of study. The proportion of positive studies for proximal DVT was compared by years in practice and by specialty.

Results: 2030 studies met inclusion criteria. Acute deep venous thrombosis was identified in 334 studies (16.5%): 6.6% proximal and 9.9% isolated distal. D-dimers were only ordered prior to 113 studies (5.8%), 80 of which were positive (70.8%). The proportion of studies with an acute proximal DVT did not differ by physician specialty or years in practice; however, there was marked variability in positivity rate between individual physicians (range 0%-42.9%). Of the 79 practitioners who ordered 10 or more examinations, 18 (23%) achieved a positivity rate above 15%; however, 23/79 (29%) had positivity rates below 5%, including 15/79 (19%) who had no positive studies. 

Conclusion: In the investigation of acute proximal DVT we found evidence of systematic underuse of D-dimer and an overreliance on Doppler ultrasonography. Within specialties, we noted substantial variability in positivity rates between practitioners with both high and low performers. Within our institution we expect that addressing practitioner variability will help optimize resource utilization while minimizing the risk of false positive studies.