P061: Preventable adverse drug events in Canadian emergency departments

Abstract

Introduction: Adverse drug events (ADEs), unintended and harmful events associated with medications, cause or contribute to 2 million emergency department (ED) visits in Canada each year. Our objective was to determine the proportion of preventable ADEs by event type, severity, drug and drug class, and describe associated factors. Methods: We reviewed the charts of ADE patients enrolled in 1 of 3 prospective studies conducted in 3 tertiary care and 1 urban community ED. In the parent studies, researchers enrolled patients by applying a systematic selection algorithm to minimize selection bias, and physicians and pharmacists evaluated patients prospectively to evaluate causal associations between the drug regimens and patient presentations. After completion of the prospective study, a research pharmacist and physician independently reviewed the charts of all ADE patients, abstracted data using an electronic form and applied 3 preventability algorithms. The main outcome was a probably or definitely preventable ADE defined as avoidable by adhering to best medical practice, appropriate monitoring, taking a history of prior ADEs, compliance with recommended therapy, and avoidance of errors. Reviewers discussed discordant ratings until reaching consensus. We used kappa scores to evaluate between rater agreement, and investigated risk factors for preventability using logistic regression. Sample size was based on enrolment into the parent studies. Results: We reviewed the charts of 670 patients diagnosed with 725 ADEs. We excluded 44 patients with incomplete assessments. The inter-rater agreement in categorizing ADEs as preventable was 0.51 (95%CI 0.42-0.59). We deemed 61% (95%CI 57-65%) of ADEs preventable. Of preventable events, 30% were due to non-adherence, 24% to adverse reactions, and 15% to an excessive dose, and 29% required hospital admission. Among preventable events, 8% were due to warfarin, 5% hydrochlorothiazide, 3% acetylsalicylic acid, and 3% insulin. On multivariate analysis, mental health diagnoses were associated with preventable ADEs (OR 2.1, 95%CI 1.3-3.3, p=0.002). Conclusion: In this large multi-centre cohort, preventable events made up the majority of ADEs, and utilized substantial hospital resources. Strategies to reduce ED visits due to ADEs should target improving adherence behavior, and developing interventions for patients with mental health diagnoses and on high-risk medications.