Improving Effectiveness of Thrombophilia Work-up

Background: Thrombophilia testing is often performed in both the inpatient and outpatient settings. However, inappropriate ordering of laboratory tests may increase costs, patient burden, and unnecessary referrals. Providers may order tests not generally recommended by guidelines due to a lack of standardized testing and uncertainty of clinical significance. Such testing results in minimal clinical utility. We hypothesized that revising existing Cerner order sets in the electronic health record (EHR) and adding clinical decision-making support could improve the effectiveness of thrombophilia work-up at Stony Brook University Hospital (SBUH) by decreasing the number of inappropriate tests ordered. Methods: For this first “Plan-Do-Study-Act” (PDSA) cycle, our focus was on minimizing the ordering of tests that have minimal to no use case in routine thrombophilia work-up: beta-2-glycoprotein IgA, anti-cardiolipin IgA, phosphatidylserine IgG/IgM/IgA, MTHFR mutation testing, serum viscosity, and protein S activity. The tests considered appropriate were anti-cardiolipin IgG and IgM, beta-2 glycoprotein IgG and IgM, factor II mutation, factor V Leiden, free protein S antigen, and antithrombin III, as these tests have a generally accepted role when thrombophilia evaluation is indicated. We first examined the frequency with which the above defined inappropriate tests were ordered from January 1, 2023 through June 22, 2024 at SBUH for patients 18 years of age and older in both the inpatient and outpatient settings. Orders by pediatric providers were excluded. We calculated the percentage of inappropriate tests out of the total number of tests ordered for thrombophilia work-up. The electronic medical record (EMR) order sets were revised to include only one outpatient order set and one inpatient order set for thrombophilia work-up. A separate order set was created for neurology department use which included laboratory testing relevant to stroke work-up. Clinical decision support was included to offer guidance on testing in arterial thromboses, venous thromboses, and testing performed after starting anticoagulation. We then re-examined the above metrics after implementing the new order sets from June 23, 2024 to July 25, 2024. Since the pre- and post-evaluations were over different time frames, the weekly percentage of inappropriate orders out of total orders for thrombophilia work-up was used for comparison. Groups were compared using Welch's unpaired two-tailed t-test. Results: A total of 9489 orders (120 orders per week) were reviewed pre-implementation and 478 orders (96 orders per week) were reviewed post-implementation. The average percentage of inappropriate orders out of total thrombophilia orders showed a significant decrease, from 21.9±4.9% pre-implementation to 12.0±2.9% post-implementation (p = 0.0004). Conclusions: Although preliminary, the statistically significant decrease in the percentage of inappropriate tests in just one month since implementation of the new order sets is promising. Since no additional educational initiatives or other interventions were done during this timeframe, the change can be attributed to the redesigned order sets with embedded clinical decision support. These data support the consolidation of order sets and the elimination of nonstandard orders. We plan to reassess at subsequent time points to ensure the trend continues. Future directions will include assessing and improving the appropriateness of indications for thrombophilia testing while continuing to ensure appropriate tests are ordered.