This study successfully applied the adult-verified SPRC/SAMHSA Decision Support Tool to identify low-risk pediatric patients with SI who may be safely discharged home from the ED with follow-up psychiatric assessment and individualized instructions. In addition to comparing the tool's score with the psychiatrist’s disposition decision, we searched for other factors that influenced the decision to admit to hospital or discharge from the ED. Older age emerged as being significantly associated with a higher risk for admission (P < 0.001). Having a history of psychopathology was also significant for admission (P = 0.01), in correlation to Zaltzman's data on adults that showed that major risk factors for suicide were male sex, psychopathology, prior suicide attempt, and accessibility to means for suicide8.
Unlike reports of females being significantly more likely to attempt suicide out of a sense of hopelessness, loneliness, rejection, and guilt, as well as conflicts with parents and peers, sex was not a significant factor in the decision to admit to hospital in our study9. Two patients in our study were transgender, and both were admitted to hospital. García-Vega et al. observed that there were higher levels of SI and SA in people with gender dysphoria than in the general population.10
The largest sources of referral to hospital admission in our cohort were the parent(s) and self-referral (P < 0.001). Most patients referred by the school were discharged (118/138, 86%, P < 0.001). This latter finding is probably due to the current policies of the Israeli Ministry of Education which requires school staffs to complete a suicide risk assessment for any case of suicidal expression by a pupil11.
Out of 93 patients categorized by the decision support tool as being at low risk, only 5 (5.4%) were admitted to hospital (P < 0.001), and all due to social issues which required welfare intervention. In contrast, 85 (42%) of the high-risk patients were hospitalized. These findings are in line with a previous study on adults.7 Our results suggest that pediatric patients with SI can be discharged home if they are classified as low risk (a score of zero) by the Decision Support Tool, while those classified as high risk should have a psychiatric evaluation during the ED visit in the greater likelihood that they will require hospitalization. When we applied a regression analysis of the tool's scores, we found that a previously diagnosed mental condition, past SA, thoughts of carrying out a plan, suicidal intent, and substance abuse were significant factors which could independently predict admission. Irritability/agitation/aggression did not emerge as independent predictors for admission, possibly due to the wide range of behaviors that can be classified as "irritability"12. Furthermore, the physical conditions in the ED (bright lights, overcrowding, noise) may influence the patient and the physician when grading irritability.
There are several limitations to this study, beginning with its retrospective design which relies upon accuracy of the documentation during the ED visit and lacks follow-up information as well as any knowledge of the outcome after the index visit. Family status documentation was also lacking, including remarriage of one or both of biological parents or loss of a parent, two factors known to be related to suicidal behavior 13,14. Second, the psychiatric assessments in the study were done by 3 different psychiatrists and clinician bias may have affected the disposition decision. Furthermore, the study was carried out in a single institution with a specific urban population, thus limiting generalizability.