Diagnosing Lower-Extremity Deep Vein Thrombosis: Can an Emergent Referral System be Improved?

Abstract

Introduction

—Clinical diagnosis of deep vein thrombosis (DVT) can be difficult. This fact, along with our litigious environment and the expediency of obtaining venous duplex ultrasound (VDU) studies, has led to the overuse of this examination. To improve the appropriateness of VDU orders from the emergency department (ED), a quality improvement project was initiated with the ED and vascular lab.

Methods

—A prospective study was conducted involving ED patients referred for lower-extremity VDU to diagnose DVT. The registered vascular technologist (RVT) assessed the patients before the examination by using a pretest probability (PTP) scoring system for DVT. The results were shared with the ED physicians at regular intervals. For the last 6 months of the project, the RVT made a pretest clinical impression of test outcome. The VDU results, the distribution of pretest scores and the results of the RVT's pretest impression were analyzed.

Results

—A total of 291 ED patients with suspected lower-extremity DVT were assigned a PTP score. Of all positive studies, 72% (36/50) had a high PTP score. A moderate PTP score lacked accuracy for predicting DVT, with 24% (12/50) of the positive and 42% (102/241) of negative exams in this group. A total of 36% (104/291) of the patients had a low PTP score, and the incidence of DVT was 2% (2/104) in this group. Of the total 291 patients, 177 patients had a RVT pretest clinical impression performed. The RVT impression was accurate for the presence of DVT 83% (24/29) of the time. The RVT correctly predicted that there was no DVT in 92% (136/148) of the negative exams. Furthermore, in the moderate PTP group, where the PTP score was not as predictive of the presence or absence of DVT, the RVT impression was accurate 89.5% (60/67) of the time.

Conclusions

—Although a moderate PTP score is not a reliable predictor of a positive VDU, a high score validates the need for a VDU, whereas a low score does not warrant the exam. If ED practitioners consistently used a pretest clinical scoring system, the number of negative examinations could be significantly reduced. In addition, the overall accuracy of the RVT's clinical impression demonstrates that experience in DVT diagnosis can likely improve patient selection for VDU.