To Test or Not to Test? Quantifying and Reducing Inappropriate Hypercoagulable Workups in the Inpatient Setting

Background: Since the proposal of Virchow's Triad in 1856, the concept of hypercoagulability has been useful in analyzing causes of thromboembolic events. In the inpatient setting, evaluation of underlying causes of thrombosis is often complicated by unclear indications for testing, unreliability of individual assays, and confounding factors such as thrombosis itself and use of anticoagulant medications. We endeavored to quantify the occurrence of inappropriate hypercoagulable work-ups (HCW) in the inpatient setting in our institution, and then to implement a resident-focused educational intervention with the goal of limiting this unnecessary and potentially harmful use of resources.

Methods: We retrospectively reviewed all patient charts containing an order for Factor V Leiden assay in 2017, 61 in total. If sufficient data were available, Hematology had not been consulted, and family history was not positive for thromboembolic events, orders were classified as inappropriate based on the recommendations of Pruthi¹: temporary risk factors, active anticoagulation therapy, arterial thrombosis, or indefinite anticoagulation already indicated. We calculated the percentage of total charts with inappropriate orders. For our quality improvement intervention, we used a pre- and post-test design to assess knowledge before and after an educational presentation on HCW given to internal medicine and neurology residents. The pre- and post-tests were identical and consisted of 12 questions. SPSS statistical software was used to compare results using Chi-Square analysis. A p value <0.05 was considered significant.

Results: A total of 37 charts were evaluated after exclusions, and 29 (78.4%) contained inappropriate orders for HCW. Of these, 16 (55.2%) were ordered as work-up of arterial thrombosis in patients with CVA or TIA but no cardiac shunt. 4 (13.8%) of the inappropriate orders followed provoked events. Apart from Factor V Leiden, additional assays were ordered with the following frequencies: Prothrombin gene mutation 92%, Anti-cardiolipin antibodies 78%, Homocysteine 43%, Antithrombin 42%, Protein C 42%, Protein S 42%, Anti-B-2-glycoprotein antibodies 25% and Dilute Russell's Viper Venom Test 22%. For our intervention, 85 resident tests were analyzed. An increase in correct responses occurred with statistical significance for questions regarding workup of CVA, as well as acute and chronic venous thrombosis. After the intervention, fewer residents stated that HCW would change their management of patients following a thromboembolic event (p<0.05). Residents correctly identified on pre-tests that anticoagulation would skew results of HCW. However, statistically significant improvement did occur in identifying appropriate timing of HCW between the pre- and post- test results (p<0.001).

Conclusion: Thromboembolic events are commonly encountered in the inpatient setting, and often lead to inappropriate HCW, particularly in the setting of CVA and TIA without an intracardiac shunt. The set of assays ordered may also demonstrate high variability, which suggests uncertainty about which tests are most appropriate. HCW education in our resident population resulted in a statistically significant increase in correct responses to hypothetical clinical scenarios. This may predict decreased rates of inappropriate HCW, but further study with this endpoint would be beneficial. In addition, further interventions should be investigated, including hard stops and order sets in electronic health records. Through studies of this nature, which our data suggest are needed, unnecessary and potentially harmful use of resources could be reduced.

1 Pruthi RK. Optimal utilization of thrombophilia testing. Int J Lab Hem. 2017;39(Suppl. 1):104-110. https://doi.org/10.1111/ijlh.12672