O1021: Effectiveness of the HEADS-ED Screening Tool in Predicting Need for Admission in Children Seeking Emergency Department Care with Acute Mental Health Concerns

Background: Most emergency departments (EDs) lack standardized protocols for pediatric mental health assessment, yet front-line clinicians need to be able to efficiently and accurately identify children who require additional mental health assessment and/or admission. The HEADS-ED is a tool that enables standardized and efficient healthcare provider evaluations of acute mental health concerns. In a previous single center study, a high-risk HEADS-ED score (total score >8 and/or suicidality score=2) was found to have an 82% sensitivity and 87% specificity to predict hospitalization. We assessed the generalizability of this performance by evaluating the tool’s predictive properties in a larger sample of children receiving ED-based mental healthcare.

Methods: A cross-sectional study was conducted in two pediatric EDs in Alberta, Canada. Children < 18 years with mental and behavioral health presentations were enrolled between February 1, 2021, and June 30, 2022. The HEAD-ED score, including action recommendations, assigned by an ED MD or a member of the mental health team, was used to determine the need for additional assessment by the mental health team including psychiatry consultation and possible admission. Participant demographic and ED visit data were collected after the ED visit. We calculated the sensitivity, specificity, positive and negative predictive values along with 95% confidence intervals, for the HEADS-ED tool to predict admission at the index ED visit using the a priori defined cut-point (total score ≥8 or suicide score of 2).

Results: We included 1,000 participants in the study. The most common diagnoses were suicidal ideation (40.8%, 408/1000), mood disorder (24.4 %; 244/1000), and anxiety-related disorders (21.5%; 215/1000). The HEADS-ED tool was used for 71.4% mental health evaluations (714/1000) and the median score was 5.0 (4.0, 7.0). Participants with high-risk HEADS-ED scores included 16.9% (121/714) with a HEADS-ED total score ≥8, 28.6% (204/714) with a suicide risk score of 2, and 36% (255/714) that met both high-risk criteria. A greater proportion of participants in the high-risk group had a previous mental health ED visit (29.9% [75/251] vs. 19.5% [88/451]; difference=10.4%; 95% CI of the difference: 3.6%, 17.1%). A HEADS-ED score ≥8 or a suicide score of 2 had a 79.1% (95% CI: 69.0, 87.1) sensitivity and 70.2% (95% CI: 66.5, 73.8) specificity to predict hospital admission. Score sensitivity did not differ based on sex or age. Positive and negative predictive values were 26.7% (95%CI: 21.3, 32.5) and 96.1% (95%CI: 93.9, 97.7), respectively.

Conclusion: In this study, HEADS-ED tool performance was similar to previous reports, and a low-risk evaluation reliably predicted discharge at the index visit. Tool performance may have been affected by the action recommendations given to providers for patients identified as high-risk. Further research looking at refining tool performance across differing clinical presentations is warranted.