Assessment of Selective and Universal Screening for Suicide Risk in a Pediatric Emergency Department

Key Points Question Are results of universal and selective screening for suicide risk implemented in a pediatric emergency department associated with future suicidal behaviors? Findings In this cohort study of 15 003 youths aged 8 to 18 years, positive screens were significantly associated with subsequent suicide-related hospital visits compared with standard emergency department procedures. Screening also more than doubled the detection of suicide risk compared with treatment as usual. Meaning These findings suggest that screening for suicide risk in pediatric emergency departments is an effective approach to identify risk for subsequent suicide-related emergency department visits.


eAppendix 2. MD-SPIN PROCEDURE
As a way to comply with the Joint Commission's recommendation and in an attempt to address the growing rate of youth suicide, a pediatric emergency department in an urban hospital implemented suicide risk screening as a standard of care.
The implementation of systematic suicide risk screening began with the formation of a task force consisting of nurses, doctors, researchers, and other stakeholders. The task force's first objective was to select a suicide riskscreening tool; the committee wanted to select a tool that had strong psychometrics and that was both brief and easy to implement. Based on the latter criteria the group decided to use to Ask Suicide-Screening Questions tool better known as the ASQ 13 . Early studies on the ASQ indicated that the screener had good psychometrics with high sensitivity and specificity.
After the screening tool was selected, medical staff were given training on how to administer the ASQ and talk to patients about suicide. The ASQ screening tool was then built into the electronic health record, making it simple for nursing staff to administer and record in the busy ED setting. Due to concerns surrounding workflow and acuity, the ASQ was initially just administered to youth ages 8 to 18 presenting to the pediatric ED with a psychiatric chief complaint.
The medical staff was able to implement the ASQ with relatively little interruption of the ED workflow, and found that the ASQ was valuable in recognizing patients who may be at an increased risk for suicide. Due to the successful implementation of the ASQ in psychiatric patients, in January of 2017, the ED staff decided to implement universal suicide risk screening; this meant that every patient age eight and older would be screened with the ASQ.
The nurses administered the ASQ along with the other routine screenings they conduct at triage. Screening at triage was ideal because than medical providers can prepare to adjust their plan of care if a positive screen occurs. If a patient did screen positive on the ASQ, the medical team was notified and the ED social work team conducted a risk assessment.
Data on ASQ screens were extracted monthly from the EHR. Data included the patient's age, race, ethnicity, sex, chief complaint, principle diagnosis, date of ED visit, health co-morbidities, ED disposition, and response to each of the ASQ items. When universal screening began in January of 2017, the same data was extracted from the universal population. A database was created for all ASQ screenings completed over the three and a half years. No patients were excluded on the basis of gender, race, or sex.

Relative Risk for Death by Suicide
Relative risk for death by suicide could be calculated only for the selective screening sample, as there were no deaths due to suicide recorded in the universal screening sample within our data. Similarly, analysis of relative risk for death within the selective screening sample should be considered exploratory given the low total number of deaths by suicide. Relative risk is nonetheless presented in order to provide a preliminary estimate of the magnitude of association between positive screen and subsequent death by suicide within the selective screening sub-sample.

eAppendix 3. Relative Risk for Death by Suicide
There were three deaths by suicide within the selective screening sub-sample during the follow-up period, as well as one additional undetermined death. Of these deaths, two of those that died by suicide had screened positive on the ASQ, and the third had screened negative. The undetermined death likewise screened negative. The relative risk for confirmed death by suicide in the selective screening subsample was RR(95% CI)=4.50(0.41-49.57), and with inclusion of the undetermined death, RR(95% CI)=2.25(0.32-15.96). There were no deaths by suicide among the universal screening sub-sample within our follow-up period. All three deaths occurred at least a year after the initial screening, M(SD)=833.33(210.07) days from the index visit.
Notably, of the three confirmed deaths by suicide, two had first come to the ED with chief complaints unrelated to suicide (i.e., wheezing and externalizing behaviors). While the youth with externalizing behaviors screened negative on the ASQ, the one with a chief complaint of wheezing screened positive, and had a documented history of being seen for homicidal ideation. Both of the patients who screened positive on the ASQ were admitted or transferred from the ED during this index visit, whereas the one who screened negative was discharged. Taken together, of these three deaths by suicide, suicide risk was identified at the index visit for one based on chief complaint and the ASQ, one by the ASQ alone (i.e., would have otherwise not been identified), and one was missed entirely, or perhaps did not yet experience suicidal ideations or behavior.  4. Frequency data for relative risk calculations, using an "either/or" approach in which suicide risk is defined as a positive ASQ screen and/or a suicide-related presenting problem.