MP36: Reducing utilization of unnecessary coagulation tests by emergency providers

Abstract

Background: Curbing unnecessary laboratory testing represents a significant opportunity for cost reduction in the Canadian health care system. A Choosing Wisely report cited a 31% decline in the number of tests ordered in a Canadian emergency department (ED) after implementation of recommendations. The international normalized ratio (INR) remains frequently ordered in emergency departments without an appropriate indication. Aim Statement: We aimed to reduce the number of INR tests completed in the St. Joseph's Healthcare Hamilton Emergency Department by 50% by April 30, 2019. Measures & Design: We conducted the study in an urban, academic ED employing the Epic electronic health record (EHR). We tailored interventions according to the Hierarchy of Effectiveness to address root causes revealed by analysis of our baseline ordering behaviour. Interventions included provider education around evidence-based ordering indications and removal of the INR from our “chest pain” bloodwork panel. Our outcome measure was the weekly number of INR tests completed per ED visit. Process measures included the proportion of INR tests ordered for inappropriate indications on monthly audits of 20 charts where an INR was completed. Balancing measures included average ED length of stay for patients receiving INR testing. Evaluation/Results: We collected outcome, process, and balancing measures through the EHR and analyzed this data using statistical process control charts. Over the nine-month study period, we decreased weekly INR tests from 248.4 to 115.0, a reduction of 56% which met criteria for special cause variation. This amounts to a cost savings of $43,008 per year. ED length of stay for patients receiving INR testing did not change significantly. Discussion/Impact: Our interventions were successful in realising our 50% target reduction in INR tests without an increase in ED length of stay from repeat venipuncture. This result is in keeping with similar efforts in other Canadian EDs. Our interventions could likely be spread to other settings where an INR is included as part of a “chest pain” panel. This may represent a substantial cost reduction opportunity on a national scale. Further work is needed in order to assess long term sustainability, which can be supported by employing high effectiveness mechanisms such as automation of optimal behaviour.