IntroductionThrombophilia testing is costly, and results can be misleading in the setting of acute venous thromboembolism (VTE) or anticoagulation use. It is frequently ordered despite limited impact on clinical decision making. American Society of Hematology (ASH) Choosing Wisely guidelines recommend against thrombophilia testing in patients with VTE in the setting of major transient risk factors (surgery, trauma, prolonged immobility).MethodsA best practice alert (BPA) was implemented in our electronic medical record system to promote awareness of the ASH Choosing Wisely guidelines. The BPA showed the text of the ASH Choosing Wisely guidelines onscreen whenever thrombophilia testing was ordered for a patient with a diagnosis of pulmonary embolism (PE) or deep venous thrombosis (DVT). It was up to the clinician to decide whether the Choosing Wisely guidelines applied in a particular clinical scenario and whether to follow the BPA recommendation and cancel the thrombophilia test or to override the BPA and order testing.Thrombophilia testing orders which would trigger the BPA were protein C and S activity levels, lupus anticoagulant, dilute Russell viper venom test, homocysteine levels, Factor V Leiden, prothrombin gene mutation and activated protein C resistance.The BPA began displaying to providers in the outpatient setting on 7/1/2016 and in the inpatient setting on 9/1/2016. The pre-BPA period was defined as the 12 months prior to activation of the outpatient BPA (7/1/2015 to 6/30/2016). The post-BPA period was defined as the 7 months from the activation of the inpatient BPA (9/1/2016 to 3/31/2017). We reviewed the number of BPAs triggered as well as the actions taken by providers (follow or override). We also compared the total number of thrombophilia tests in the outpatient and inpatient settings in the pre- and post-BPA periods.ResultsThe BPA appeared onscreen 140 times in the outpatient setting and 35 times in the inpatient setting during the study period. The BPA was followed 17.1% of the time in the outpatient setting and 100% of the time in the inpatient setting. Thrombophilia testing volumes stayed unchanged in the outpatient setting (471.5 tests per month pre-BPA vs. 471.6 tests per month post-BPA, p=1.00). In contrast, there was a significant decrease in the average number of thrombophilia tests per month (101.1 vs. 73.3, p=0.038) in the inpatient setting after the implementation of the BPAs. During the study period, inpatient censuses were not statistically different in the pre-BPA and post-BPA periods.DiscussionWe found that the thrombophilia BPA was followed at a very high rate in the inpatient setting compared to the outpatient setting. The BPA is meant to remind providers of the ASH Choosing Wisely guidelines, but providers must still use clinical judgment to determine whether the guideline applies in a given clinical scenario. The high rate of compliance in the inpatient setting suggests that in many cases, thrombophilia testing is being ordered on inpatients with transient risk factors, and that a brief reminder message can effectively change provider behavior. On the other hand, BPA compliance was relatively low in the outpatient setting, though this could be appropriate based on the clinical scenarios which outpatient providers encounter. Future analyses will include manual chart review to determine the appropriateness of BPA overrides in the outpatient setting, as well as a thematic analysis of BPA override comments.There was a significant reduction in the average number of inpatient thrombophilia tests ordered per month in the post-BPA period compared to the pre-BPA period. Although the inpatient thrombophilia BPA had a high rate of compliance, it was only shown 35 times during the study period. Thus the BPA messages alone were unlikely to account for the significant change in total volume of thrombophilia testing. Prior research on blood transfusion BPAs suggests that BPAs can have an educational impact, causing providers to become more aware of guidelines and thus actually reducing the frequency of triggered BPAs over time. Thus, each BPA message may have a multiplicative effect, preventing future unnecessary tests in addition to the original order. DisclosuresNo relevant conflicts of interest to declare.