• Analise Peleggi, Bryan Strub, Soo-Jeong Kim, and Carol M Rockhill.
• University of Washington, Department of Psychiatry and Behavioral Sciences, Seattle, WA, United States of America.
• Am J Emerg Med. 2022 May 1; 55: 89-94.

ObjectiveViolence and aggressive behaviors among youth are a leading cause of Emergency Department (ED) mental health (MH) encounters. A consistent method is needed for public health research, to identify ED encounters associated with aggression. The aim of this study was to develop such a screening procedure.Data SourcesElectronic records and administrative claims data related to MH related ED encounters at one of Pediatric Health Information System (PHIS) Children's Hospitals in the United States from January 1, 2019 to December 31, 2019.Study DesignThe authors selected a combination of ICD-10 codes to screen MH ED encounters for aggression; and then conducted a chart review to compare characteristics of groups that screened positive vs. screened negative, and groups with confirmed vs. without confirmed aggression.Data Extraction MethodUnique ED encounters associated with a MH related ICD-10 code from a one-year period at the study institution were extracted (n = 3092 MH ED encounters). Encounters with any aggression-associated codes were identified as "screen-positive" (N = 349). From the remaining "screen-negative" encounters, 352 unique encounters were randomly selected as a comparison group. Both groups were chart reviewed to investigate the accuracy of the screening method.Main FindingChart review confirmed aggression in 287 of 349 screen-positive and 48 of 352 select screen-negative, chart-reviewed encounters. Additional codes were added, with a goal of finding the combination of codes with the highest accuracy. The resulting screen had sensitivity, specificity, positive and negative predictive values of 0.901, 0.817, 0.818, and 0.864, respectively.Principal ConclusionsThis paper presents a screening method for identifying ED encounters related to aggression. A replication study will be necessary to validate the method prior to applying to large claims data. If validated, it will support future research on this important population.Copyright © 2022 Elsevier Inc. All rights reserved.

15 keywords...

hide…