Evaluating the Lethal Means of Non-Fatal Suicide Attempts Presenting to Washington State Emergency Departments Before and During the COVID-19 Pandemic

Introduction: During the COVID-19 pandemic, concerns about increasing mental health conditions have emerged due to intersecting stressors associated with COVID-19 morbidity and mortality as well as disease mitigation efforts such as stay-at-home orders, school closures, and social distancing. However, little is known regarding the COVID-19 pandemic’s impact on suicidal behavior in Washington. Methods: This observational longitudinal study utilized 7,765 Washington Emergency Department (ED) visit records from 2019 to 2021 to describe the lethal means utilized for non-fatal suicide attempts and investigate if the utilization of lethal means for suicide attempts has changed during the COVID-19 pandemic. Data on lethal mean utilization were collected from visit records using a novel, automated classification approach leveraging ICD-10 diagnostic codes and clinical free-text fields. Results: Poisoning was the most identified lethal mean among non-fatal suicide attempt visit records during the study period (n = 5833), followed by Other (n = 1010), No Lethal Mean detected (n = 601), Suffocation (n = 356), and Firearms (n = 68). There were a few instances where non-fatal suicide attempt visit counts and proportion of all ED visits increased relative to before the COVID-19 pandemic such as for poisoning (+13.7%, Visit Ratio = 1.23 [1.1, 1.37]), other (+28.8%, Visit Ratio = 1.39 [1.04, 1.86]), and total suicide attempts (+16.3%, 1.26 [1.14, 1.39]) during Winter 2021 (2/21 – 3/20/2021). School age youth (12 – 17 years) appeared to be most impacted as non-fatal suicide attempts accounted for the largest share of ED visits (across all subgroups examined), and they experienced the greatest number of statistically significant elevated visit ratios for suicide attempts during the COVID-19 pandemic (n = 12). Conclusions: This study developed a novel approach to identify the utilization of lethal means for non-fatal suicide attempts using population-based ED data. This approach could be valuable in monitoring real-time changes in population-level suicidal behaviors and evaluating the impact of lethal mean restriction policies.