Screening and Risk Algorithms for Detecting Pediatric Suicide Risk in the Emergency Department.
Academic Article
Overview
abstract
IMPORTANCE: The Joint Commission requires hospitals and behavioral health care organizations to identify patients at risk of suicide (National Patient Safety Goal 15.01.01). Risk algorithms and in-person screening for suicide risk show utility in identifying patients at risk of suicidal behavior, yet there is little research examining their comparative performance in children and adolescents. OBJECTIVE: To assess the performance of suicide risk screening and risk algorithms in identifying the risk of suicide attempts among pediatric patients in the emergency department (ED). DESIGN, SETTING, AND PARTICIPANTS: This retrospective cohort study included youths aged 10 to 18 years presenting to the ED of a northeastern US state between September 1, 2019, and August 31, 2021. Patients were screened for suicide risk using the Ask Suicide-Screening Questions survey and the Columbia-Brief Suicide Severity Rating Scale. Electronic health records from this same cohort containing data from May 31, 2017, to the date of their first encounter within this period were extracted to train a risk algorithm. To observe the presence or absence of a suicide attempt, patients were followed up from their first ED encounter for a minimum of 6 months and a maximum of 2.5 years, through March 2, 2022. Data were analyzed from May 2023 to December 2024. EXPOSURE: Assessments from suicide risk screening and a risk algorithm. MAIN OUTCOMES AND MEASURES: The occurrence of a suicide attempt following a patient's first suicide risk screening or first visit in the screening period, if not screened. RESULTS: Among 19 653 patients included in the analysis, the median age was 14.3 (IQR, 12.1-16.2) years, and 10 007 (50.9%) were female. Four hundred ninety-five patients (2.5%) were treated for a suicide attempt. Among patients screening positive for suicide risk in testing samples (mean, 8.1% [95% CI, 7.6%-8.6%]) and patients in the top 8.1% of the distribution on the algorithm, the algorithm correctly identified a mean of 50.7% (95% CI, 47.3%-54.1%) of those who attempted suicide in contrast to 36.5% (95% CI, 31.9%-41.2%) identified by screening. The algorithm uniquely identified 127% more youths who attempted suicide (125) than did screening (55). CONCLUSIONS AND RELEVANCE: In this cohort study of pediatric patients, the risk algorithm was superior to screening across all performance metrics and could substantially assist health care organizations' efforts to meet the Joint Commission's National Patient Safety Goal to reduce the risk of suicide.
