Prediction of Suicide Attempts and Suicide-Related Events Among Adolescents Seen in Emergency Departments

Key Points

Question  How do the Ask Suicide-Screening Questions (ASQ) and the Computerized Adaptive Screen for Suicidal Youth (CASSY) instruments compare in predicting suicide attempts (SAs) among adolescents?

Findings  In a longitudinal cohort study of 2740 youths seen in pediatric emergency departments with a 3-month follow-up, both ASQ and CASSY showed strong predictive validity and similar sensitivity and specificity in predicting future SAs. Both ASQ and CASSY performed similarly among patients with physical symptoms; the CASSY more accurately predicted SAs for those with psychiatric symptoms.

Meaning  This study suggests that for universal screening, both instruments perform well among patients with physical symptoms, but for the small subset of youths with psychiatric symptoms, the CASSY has greater predictive validity.

Importance  Screening adolescents in emergency departments (EDs) for suicidal risk is a recommended strategy for suicide prevention. Comparing screening measures on predictive validity could guide ED clinicians in choosing a screening tool.

Objective  To compare the Ask Suicide-Screening Questions (ASQ) instrument with the Computerized Adaptive Screen for Suicidal Youth (CASSY) instrument for the prediction of suicidal behavior among adolescents seen in EDs, across demographic and clinical strata.

Design, Setting, and Participants  The Emergency Department Study for Teens at Risk for Suicide is a prospective, random-series, multicenter cohort study that recruited adolescents, oversampled for those with psychiatric symptoms, who presented to the ED from July 24, 2017, through October 29, 2018, with a 3-month follow-up to assess the occurrence of suicidal behavior. The study included 14 pediatric ED members of the Pediatric Emergency Care Applied Research Network and 1 Indian Health Service ED. Statistical analysis was performed from May 2021 through January 2023.

Main Outcomes and Measures  This study used a prediction model to assess outcomes. The primary outcome was suicide attempt (SA), and the secondary outcome was suicide-related visits to the ED or hospital within 3 months of baseline; both were assessed by an interviewer blinded to baseline information. The ASQ is a 4-item questionnaire that surveys suicidal ideation and lifetime SAs. A positive response or nonresponse on any item indicates suicidal risk. The CASSY is a computerized adaptive screening tool that always includes 3 ASQ items and a mean of 8 additional items. The CASSY’s continuous outcome is the predicted probability of an SA.

Results  Of 6513 adolescents available, 4050 were enrolled, 3965 completed baseline assessments, and 2740 (1705 girls [62.2%]; mean [SD] age at enrollment, 15.0 [1.7] years; 469 Black participants [17.1%], 678 Hispanic participants [24.7%], and 1618 White participants [59.1%]) completed both screenings and follow-ups. The ASQ and the CASSY showed a similar sensitivity (0.951 [95% CI, 0.918-0.984] vs 0.945 [95% CI, 0.910-0.980]), specificity (0.588 [95% CI, 0.569-0.607] vs 0.643 [95% CI, 0.625-0.662]), positive predictive value (0.127 [95% CI, 0.109-0.146] vs 0.144 [95% CI, 0.123-0.165]), and negative predictive value (both 0.995 [95% CI, 0.991-0.998], respectively). Area under the receiver operating characteristic curve findings were similar among patients with physical symptoms (ASQ, 0.88 [95% CI, 0.81-0.95] vs CASSY, 0.94 [95% CI, 0.91-0.96]). Among patients with psychiatric symptoms, the CASSY performed better than the ASQ (0.72 [95% CI, 0.68-0.77] vs 0.57 [95% CI, 0.55-0.59], respectively).

Conclusions and Relevance  This study suggests that both the ASQ and the CASSY are appropriate for universal screening of patients in pediatric EDs. For the small subset of patients with psychiatric symptoms, the CASSY shows greater predictive validity.

The rates of suicidal ideation (SI), suicide attempts (SAs), and suicide have increased among adolescents over the past 15 years, with a sharp increase among Black and Hispanic youths.^1-4 There is a growing consensus that screening patients seen in emergency departments (EDs) for suicidal risk is an important component of adolescent suicide prevention.⁵ One-fifth of youths visit an ED at least annually.⁶ Patients presenting to the ED are more likely than the general population to be at risk for suicide; suicide decedents aged 10 to 24 years were nearly 7 times more likely than living controls to have visited an ED within 30 days prior to their death.⁷ A significant proportion of youths seen in pediatric ED settings for physical symptoms have positive screening results for suicide risk.^8,9 Moreover, there has been an increase in presentations to pediatric EDs for adolescent suicidal behavior that has accelerated since the onset of the COVID-19 pandemic.^1,10,11 The last clinical contact for a substantial proportion of patients with SAs and for decedents is an ED visit.^7,12,13 Screening in ED settings may also help to address racial disparities in suicidal behavior and access to care.^2,3,14

There is less consensus about how to screen for suicide risk. The Emergency Department Study for Teens at Risk for Suicide (ED-STARS) was developed to address this question. ED-STARS, based in the Pediatric Emergency Care Applied Research Network (PECARN), assessed youths for suicidal risk and observed them longitudinally to assess how best to screen for and predict subsequent suicidal behavior. A previous study described the development and validation of the Computerized Adaptive Screen for Suicidal Youth (CASSY).¹⁵ In this study, we compare the performance of the CASSY with a widely used measure for screening for suicidal risk among patients in EDs, the Ask Suicide-Screening Questions (ASQ).¹⁶

Although the CASSY was developed explicitly to predict suicidal behavior, the ASQ was originally developed to identify concurrent risk for SI. Nevertheless, both measures show evidence of predictive validity.^13,15,17,18 The extant findings from studies of these 2 instruments are not easily compared because these studies differed by age range of patients, method and content of outcome assessment, and geographic distribution of EDs.^15,17,18 The 2 studies demonstrating the predictive validity of the ASQ focused on youths aged 8 to 18 years, using record review of a return to the ED for SI or SA (suicide-related event [SRE]), and were conducted in a single urban ED. The initial validation of the CASSY was based on a network of 13 EDs among adolescents aged 12 to 17 years. The primary outcome for the validation of the CASSY was an SA assessed by follow-up interview.

Herein, we compare the performance of the ASQ and the CASSY with respect to the prediction of our primary outcome of SAs and a secondary outcome of visits to the ED or hospital for an SRE in the 3 months after baseline based on participant and parental report. The performance of the ASQ and the CASSY are compared in subgroups defined by age, sex, race, ethnicity, and presenting chief symptom (physical vs psychiatric). We hypothesized that both instruments would perform significantly better than chance in the prediction of SAs and SREs. In light of racial and ethnic disparities in access to care and in suicidal behavior, we wanted to confirm that both instruments would perform as well for Black and Hispanic youths as for White youths in the prediction of suicidal behavior.^2,3,14,19,20 The ASQ has demonstrated equivalent psychometric properties for Black and White youths in terms of concurrent suicide risk.²¹ We hypothesized that the CASSY would show greater discriminating value than the ASQ in predicting SAs because of the adaptive and dimensional nature of the CASSY and because the CASSY always administers 3 of the ASQ items.

These data are from ED-STARS, a multicenter, random-series, prospective cohort study supported by PECARN.²² Study 1 of ED-STARS is a longitudinal follow-up of 2075 adolescent patients, aged 12 to 17 years, seen in 1 of 13 PECARN EDs from June 26, 2015, through July 31, 2016, and was used to develop the CASSY.¹⁵ The present study’s evaluation of the ASQ and the CASSY is based on study 2 of ED-STARS, which consists of 2740 adolescents seen in 14 PECARN EDs and 1 Indian Health Service ED between July 24, 2017, and October 29, 2018, who completed a CASSY, an ASQ, and a 3-month follow-up assessment. Derivation of the sample size is provided in eAppendix 1 in Supplement 1. For study 2, the EDs were in the West (4 [26.7%]), Southwest and Central (2 [13.3%]), Midwest (4 [26.7%]), and Mid-Atlantic and New England (5 [33.3%]) regions of the US. By design, 40.3% (1105 of 2740) of the study 2 sample presented with psychiatric chief symptoms. Adolescents completed baseline assessments in the ED on a computer tablet. Interviewers, blinded to baseline data, conducted 3-month computer-assisted telephone follow-up interviews. Adolescent participants received $15 at baseline and $25 or $35 at each follow-up in the form of Amazon.com e-gift cards (hard copy mailed if requested). We obtained institutional review board approval to increase the incentive at follow-up to $35 for nonresponders to obtain more generalizable data. We obtained institutional review board approval from all sites (Morgan Stanley Children’s Hospital of NewYork-Presbyterian, The Children’s Hospital of Philadelphia, Cincinnati Children’s Hospital Medical Center, Children’s National Medical Center, Children’s Hospital of Colorado, Hasbro Children’s Hospital, Children’s Hospital of Wisconsin, C.S. Mott Children’s Hospital–University of Michigan, Nationwide Children’s Hospital, Primary Children’s Hospital, Texas Children’s Hospital, University of Arizona Medical Center, University of California Davis Children’s Hospital, Children’s Hospital of Pittsburgh, and Whiteriver Indian Hospital [Indian Health Service]). Parents or guardians provided written informed consent, and adolescents provided written assent. Parental consent or assessments were conducted in English or Spanish. This study followed the Standards for Reporting of Diagnostic Accuracy (STARD) reporting guideline and used a prediction model to assess outcomes.

The ASQ consists of 4 yes-or-no items that assess recent SI, burdensomeness, and lifetime suicidal behavior; a “yes” answer or a nonresponse to any of these items results in a positive screening result for suicide risk.^16,23 The ASQ showed very high sensitivity (96.9%), specificity (87.6%), and concurrent validity for the identification of clinically significant SI as assessed by a positive score on the Suicidal Ideation Questionnaire (area under the receiver operating characteristic curve [AUROC] = 0.92).¹⁶ A study of 15 003 youths aged 8 to 18 years, recruited by a combination of universal and targeted screening, showed a sensitivity of 77% and a specificity of 85% for predicting return to the ED for an SRE, as assessed by record review.¹⁷ A subsequent study in the same ED using universal screening demonstrated a sensitivity of 67%, a specificity of 84%, and an AUROC of 0.75.¹⁸

The CASSY was developed with data from study 1 and validated with 2754 adolescent ED patients (study 2).¹⁵ In study 2, adolescents completed the CASSY at baseline, which, as a computerized adaptive test, drew from a pool of 72 items that covered a broad range of risk and protective factors for suicidal behavior. Simulated adaptive testing from the complete response patterns revealed a mean of 11 of the 72 items (mean [SD], 15.3% [5.6%]; range, 5%-21%) per participant. The CASSY always administers 3 ASQ items—past week SI, death wish in past few weeks, and lifetime history of SA—as “anchor items” to ensure coverage of suicidal items in each administration.²⁴ Mean (SD) CASSY scores, which are the predicted likelihood of SA within the 3-month follow-up period, were 0.18 (0.11) for those who did make an SA and 0.05 (0.07) for those who did not make an SA. The CASSY had an AUROC of 0.89 in study 1 and 0.87 in study 2 for predicting SAs assessed by 3-month follow-up interviews. Using the 80% specificity threshold from study 1, the CASSY had a sensitivity of 82.4% and specificity of 72.5% (eAppendix 2 in Supplement 1).

Demographic information was obtained from the parent or legal guardian and included self-reported race and ethnicity (American Indian or Alaska Native; Asian, Native Hawaiian or Other Pacific Islander; Black or African American; Hispanic or Latino; White; or multiracial), educational level of parents, and receipt of public assistance. Adolescents also answered 37 to 59 items that characterized the sample on SI and suicidal behavior, nonsuicidal self-injury, depression and anxiety, positive affect, alcohol and drug use, fighting, bullying and victimization, history of concussion, and connectedness to family, friends, and school (eTable 1 in Supplement 1).

Our primary outcome, an SA between baseline and 3-month follow-up, was defined by (1) an adolescent or parent report of adolescent ED visit or hospitalization with an SA and/or (2) an adolescent responding “yes” to either “In the past 3 months, have you made a suicide attempt?” or “In the past 3 months, have you tried to harm yourself because you were at least partly trying to end your life?” from the adapted Columbia–Suicide Severity Rating Scale.²⁵ A secondary outcome, an SRE, was a visit to an ED or hospital for SI or SA based on participant or parental interviews.

Statistical analysis was performed from May 2021 through January 2023. Descriptive statistics were used to summarize baseline characteristics overall for those with and those without primary (SAs) and secondary (SREs) outcomes by 3 months. Baseline characteristics were compared between those retained or not retained for follow-up. Continuous variables were compared using t tests, and categorical variables were compared using χ² tests. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated for prediction of the main and secondary outcomes by the ASQ and by the CASSY at various cut points. The value of the CASSY at which sensitivity plus specificity is maximized was considered the optimal cut point. Sensitivity and specificity of predicting the outcomes by the ASQ and by the CASSY at this optimal cut point were calculated within demographic subgroups and subgroups of adolescents by chief symptoms. The AUROC values for the ASQ and the CASSY were calculated and compared overall and within these same subgroups using the method of DeLong et al.²⁶ For these subgroup analyses, α was adjusted using the method of Benjamini and Hochberg²⁷ with a false discovery rate set at 5%. The instruments’ specificities were compared between male and female youths and between youths with physical symptoms and youths with psychiatric chief symptoms using χ² tests. The CASSY scores were collected and provided by Adaptive Testing Technologies, which was blinded from outcome and other baseline data.²⁸ These scores were provided to the Data Coordinating Center at the University of Utah, where validation analyses were performed. All analyses were conducted using SAS, version 9.4 (SAS Institute Inc).²⁹ All P values were from 2-sided tests, and results were deemed statistically significant at P < .05. Because the CASSY includes 3 of the 4 ASQ items as fixed anchor items, comparison of the AUROC values for the CASSY and the AUROC values for the ASQ essentially tests the improvement in predictive accuracy of the additional adaptively administered items in the CASSY over the 3 ASQ items only.

Of the 6513 eligible adolescents approached for participation, 4050 (62.2%) were enrolled. Adolescents with complete baseline evaluations (3965 [97.9%]) were eligible for follow-up (eTable 2 in Supplement 1; Figure). Among 3933 youths who had completed both baseline CASSY and ASQ instruments, 3-month follow-ups were obtained for 2740 adolescents (69.7% retention; 1705 girls [62.2%] and 991 boys [36.2%]; mean [SD] age at enrollment, 15.0 [1.7] years) from adolescents and parents (2434 [88.8%]), adolescents only (120 [4.4%]), or parents only (186 [6.8%]) (Table 1). The study population included 105 American Indian or Alaska Native adolescents (3.8%); 62 Asian, Hawaiian, or Other Pacific Islander adolescents (2.3%); 469 Black or African American adolescents (17.1%); 678 Hispanic adolescents (24.7%); 1618 White adolescents (59.1%); 161 multiracial adolescents (5.9%); and 325 adolescents (11.9%) with unknown race. A total of 1100 of 2690 participants’ families (40.9%) received public assistance, and 793 participants (28.9%) reported a previous SA. Those who were not retained were more likely to be Black, to have parents with a lower educational level, to live in households receiving public assistance, and to have presented with a psychiatric symptom.¹⁵

Of the 2740 participants with 3-month follow-up, 234 (8.5%) had an SA, an SRE, or both. Specifically, 163 (5.9%) had at least 1 SA, 166 (6.1%) had an SRE, 95 (3.5%) had both an SA and an SRE, 68 (2.5%) had an SA only, and 71 (2.6%) reported an SRE only (Table 1).

There were no significant differences between the ASQ and the CASSY with respect to sensitivity (0.951 [95% CI, 0.918-0.984] vs 0.945 [95% CI, 0.910-0.980]), specificity (0.588 [95% CI, 0.569-0.607] vs 0.643 [95% CI, 0.625-0.662]), PPV (0.127 [95% CI, 0.109-0.146] vs 0.144 [95% CI, 0.123-0.165]), or NPV (both 0.995 [95% CI, 0.991-0.998]) in predicting an SA within 3 months, respectively, with similar findings for prediction of an SRE. Table 2 and eTable 3 in Supplement 1 show the comparison of the ASQ and the CASSY when either the specificity or the sensitivity of the CASSY is set to that of the ASQ for SAs and SREs, respectively. Across demographic and clinical stratifications, there were no differences between the CASSY and the ASQ with respect to sensitivity, specificity, PPV, or NPV.

The CASSY’s AUROC was significantly higher than that of the ASQ for predicting both SAs (0.867 [95% CI, 0.845-0.888] vs 0.769 [95% CI, 0.750-0.789]; P < .001) (Table 3) and SREs (0.841 [95% CI, 0.818-0.864] vs 0.754 [95% CI, 0.732-0.777]; P < .001) (eTable 4 in Supplement 1). The AUROCs for both instruments were significantly higher than that obtained by predicting future SAs based on age, sex, race, and ethnicity (AUROC, 0.55 [95% CI, 0.51-0.60]). The AUROC for using ASQ item 3 only (past-week SI) was similar to the AUROC obtained using the 4-item ASQ (AUROC, 0.77 [95% CI, 0.74-0.80]).

For predicting SAs, the CASSY showed higher AUROCs than the ASQ for demographic stratifications by age, sex, race, and ethnicity (Table 3). For patients presenting with physical chief symptoms, the AUROCs for the CASSY and the ASQ were both high and not significantly different (0.938 [95% CI, 0.914-0.961] vs 0.878 [95% CI, 0.807-0.949], respectively; P = .07); for those with psychiatric chief symptoms, the CASSY had a higher AUROC than the ASQ (0.724 [95% CI, 0.681-0.767] vs 0.568 [95% CI, 0.547-0.588], respectively; P < .001). For both measures, the specificity was lower in female youths than male youths (CASSY, 0.571 [95% CI, 0.546-0.595] vs 0.758 [95% CI, 0.731-0.785], respectively; ASQ, 0.524 [95% CI, 0.500-0.549] vs 0.690 [95% CI, 0.661-0.719], respectively; P < .001 for both), and for those presenting with psychiatric vs physical symptoms (CASSY, 0.248 [95% CI, 0.221-0.275] vs 0.877 [95% CI, 0.861-0.893], respectively; ASQ, 0.182 [95% CI, 0.158-0.206] vs 0.827 [95% CI, 0.809-0.846], respectively; P < .001 for both). In comparisons within each measure, the AUROCs were similar among Black, Hispanic, and White youths.

The median CASSY scores were 0.00 (range, 0.00-0.03) for those who presented with physical symptoms and 0.10 (range, 0.05-0.16) for those who presented with psychiatric symptoms. The individual ASQ items are presented for comparison in Table 4.

In this prospective, multicenter cohort study of adolescent patients seen in the ED, we compared the performance of the ASQ, a 4-item screening tool for suicide risk, with the CASSY, a computerized adaptive test that always includes 3 items from the ASQ and, on average, 8 additional items. Among patients with physical symptoms, accounting for most ED presentations, the AUROCs for predicting SAs in both measures were high and indistinguishable. There were no differences between the CASSY and the ASQ with respect to sensitivity, specificity, PPV, or NPV overall or for any stratification for either primary or secondary outcomes. However, the CASSY had a higher AUROC overall, across several demographic strata, and among those who presented with psychiatric reasons for the prediction of SAs and SREs.

The main advantages of the ASQ are that it is brief (4 items), free of charge, has widespread use and validation in multiple settings and age groups, and is integrated into youth suicide risk clinical pathways.^30-37 In a sample in which most youths present with physical symptoms, the performance of the ASQ will be indistinguishable from the CASSY. Although the ASQ was initially validated against a concurrent measure of SI, its ability to predict future SAs is logical given the association between more severe SI and future suicidal behavior.³⁸ Although the AUROCs for the prediction of SAs were similar for the ASQ item 3 and the 4-item ASQ, we caution against a single-item measure in light of findings that SI is not as strong a predictor of future SAs for Black youths compared with White youths.¹⁹ In addition, predicting an SA is not the only outcome of interest, given that youths with SI are highly likely to need a mental health referral.³⁸

The primary advantage of the CASSY is that it is a dimensional measure that provides both a continuous severity score and an estimate of the probability of an SA in the next 3 months. This continuous severity score can be used to track changes in suicidal risk over time. Clinical pathways can be developed based on stratifications of the estimated probability of an SA. Finally, the CASSY allows end users flexibility to make trade-offs between sensitivity and specificity.³⁹ The CASSY shows stronger predictive validity in patients with psychiatric chief complaints.

The main disadvantages of the ASQ are that it is a dichotomous measure and that its predictive validity for those presenting with behavioral health complaints is lower than that of the CASSY. The main disadvantages of the CASSY are that it is not free, it requires a license and a computer interface, and, while brief, it is longer than the ASQ (4 items vs a mean of 11 items). Some sites may be reluctant to document an estimated probability of an SA for medicolegal reasons. Both measures showed lower specificity in predicting SAs for female youths vs male youths and lower specificity in predicting SREs for physical vs psychiatric symptoms.

The ASQ and the CASSY will both function effectively as suicide risk screening tools for youths presenting to EDs. For universal screening, among most patients in the ED, the ASQ and the CASSY both showed equally high sensitivity and specificity. Both instruments show similar performance for Black, Hispanic, and White youths. No cost, simplicity in administration, brevity, and its integration into a care pathway favor the ASQ. Flexibility to customize sensitivity and specificity, a dimensional output that provides the likelihood of predicting an SA, and higher accuracy for predicting SAs for those with behavioral health symptoms are the main advantages of the CASSY. Using these tools as part of a clinical pathway,^30,35 which may include further assessments with the ASQ Brief Suicide Safety Assessment²³ or the Columbia Suicide Severity Rating Scale,²⁵ may provide efficient triaging for those at greatest risk. Studies conducted among adults suggest that both tools should be paired with brief interventions to promote safety planning and linkage to services that can reduce the risks for a subsequent SA.⁴⁰

This study has some strengths, including a sample that is geographically, racially, and ethnically diverse and drawn from 15 different ED settings; parental consent and interviews offered in English or Spanish; moderate acceptance and follow-up rates; and the assessment of outcomes by clinical interview rather than record review.

This study also has some limitations, including recruitment primarily from academic medical centers, as well as greater attrition among Black patients, those presenting with psychiatric symptoms, and those whose parents were poorer and had less education. Although specificity was low in both instruments for predicting future SAs, most individuals with a positive screening test for suicidal risk will have at least one lifetime psychiatric disorder, so a mental health referral is often indicated.³⁸ Future studies can examine clinician and patient preference with respect to instrument features and performance as part of a clinical pathway designed to engage patients with SI in treatment and prevent future SAs.

In this cohort study, both the CASSY and the ASQ performed well in predicting SAs and SREs among patients presenting with physical symptoms. The CASSY performed better than the ASQ in predicting future SAs among patients with psychiatric concerns, who constitute a relatively small but consequential and growing proportion of pediatric ED attendees. The similarity of the 2 measures in sensitivity, specificity, PPV, and NPV suggest that both instruments will perform well for universal screening in EDs. Emergency department clinicians will need to decide which measure is best suited to their milieu and patient mix. Both the ASQ and the CASSY are worthy of consideration for identifying youths in the ED at risk for suicide.

Accepted for Publication: December 28, 2022.

Published: February 15, 2023. doi:10.1001/jamanetworkopen.2022.55986

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2023 Brent DA et al. JAMA Network Open.

Corresponding Author: David A. Brent, MD, Department of Psychiatry, University of Pittsburgh School of Medicine and UPMC Western Psychiatric Hospital, 3811 O’Hara St, Bellefield Towers, Room 311, Pittsburgh, PA 15213 (brentda@upmc.edu).

Author Contributions: Drs Gibbons and Casper had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Brent, Grupp-Phelan, Bridge, Gibbons, Cwik, Casper, King.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Brent, Grupp-Phelan, Gibbons, Chernick, Mistry, Page.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Bridge, Gibbons, Duffy, Page, Casper.

Obtained funding: Brent, Grupp-Phelan, Gibbons, Cwik, King.

Administrative, technical, or material support: Brent, Gibbons, Chernick, Cwik, Shenoi, Mahabee-Gittens, Patel, Melzer-Lange, Keller, King.

Supervision: Rea, Cwik, Mahabee-Gittens, Mistry, Rogers, Hickey.

Conflict of Interest Disclosures: Dr Brent reported receiving grants from the National Institute of Mental Health (NIMH) (supported the development of intellectual property for the BRITE app, the As Safe As Possible intervention, the Computerized Adaptive Screen for Suicidal Youth [CASSY] measure, a suicide risk machine learning algorithm, and the Screening Wizard screening tool), the American Foundation for Suicide Prevention (AFSP), Once Upon a Time Foundation, and The Beckwith Institute; royalties from Guilford Press, eRT Inc from the electronic self-rated version of the Columbia Suicide Severity Rating Scale, and UptoDate; personal fees from Healthwise; honoraria from the Klingenstein Third Generation Foundation for scientific board membership and grant review; and serving as a member of the AFSP scientific board outside the submitted work. Dr Horowitz reported being a co-creator of the Ask Suicide-Screening Questions. Dr Bridge reported serving as a member of the scientific advisory board for Clarigent Health outside the submitted work. Dr Gibbons reported being a founder of Adaptive Testing Technologies outside the submitted work. Dr Chernick reported receiving grants from Columbia University during the conduct of the study. Dr Shenoi reported receiving grants from the National Institutes of Health (NIH) during the conduct of the study. Dr Mahabee-Gittens reported receiving grants from the National Institute of Environmental Health Sciences outside the submitted work. Dr Patel reported receiving grants from the NIMH during the conduct of the study. Dr Cohen reported grants from the NIH during the conduct of the study. Mr Page reported receiving grants from the NIH during the conduct of the study. Dr Casper reported receiving grants from the NIMH and the Health Resources and Services Administration during the conduct of the study. Dr King reported receiving personal fees from the AFSP; holding equity in Oui Therapeutics and Vytyl Health Management Inc; and sharing intellectual property rights for the Computerized Adaptive Screen for Suicidal Youth. No other disclosures were reported.

Funding/Support: This study was supported by grant U01MH104311 from the NIMH for the Emergency Department Screen for Teens at Risk for Suicide study (Drs Brent, King, and Grupp-Phelan). The Pediatric Emergency Care Applied Research Network (PECARN) is supported by the Health Resources & Services Administration (HRSA) of the US Department of Health and Human Services (HHS), in the Maternal and Child Health Bureau, under the Emergency Medical Services for Children program through the following cooperative agreements: DCC, University of Utah; GLEMSCRN, Nationwide Children’s Hospital; HOMERUN, Cincinnati Children’s Hospital Medical Center; PEMNEWS, Columbia University Medical Center; PRIME-, University of California at Davis Medical Center; CHaMP Node, State University of New York at Buffalo; WPEMR, Seattle Children’s Hospital; and SPARC, Rhode Island Hospital/Hasbro Children’s Hospital. This work was additionally supported in part by grant ZIAMH002922 from the Intramural Research Program of the NIMH. Dr. Bridge receives research grant funding from the NIMH, the Patient-Centered Outcomes Research Institute, and the Centers for Disease Control and Prevention. Dr Chernick was supported by grant K23HD096060 from the National Institute of Child and Health Development. Dr Rea was supported by a grant from the California Health Care Foundation. Dr Mahabee-Gittens was supported by grants R01ES027815 and R01ES030743 from the National Institute of Environmental Health Sciences.

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Group Information: The PECARN Collaborators are listed in Supplement 2.

Disclaimer: This information or content and conclusions are those of the authors and should not be construed as the official position or policy of, nor should any endorsements be inferred by, the HRSA, HHS, or US Government.

Data Sharing Statement: See Supplement 3.

Additional Contributions: The authors thank the PECARN research coordinators at study sites, the telephone follow-up interview team at the University of Michigan, study research assistants, and the adolescents and their caregivers who participated.