The study sample included 207 cases for ISH injuries among pediatric patients aged 4–17. Nearly half (n = 100) of the sampled cases were captured by codes for ISH with a drug poisoning mechanism (T36-T50) (Table 1). The second largest group (n = 82) was captured by codes for ISH injuries involving cutting or piercing. The 7 cases with unspecified mechanism were captured via the code T14.91 (suicide attempt).
The study reviewers found sufficient information in the attending physician’s notes to confirm 184 of the 207 sampled cases as ISH injuries (Table 2). Thus, the estimated PPV for the study definition’s ability to capture provider-documented ISH injuries was 88.9, 95% CI (83.8, 92.8%). There were 5 additional cases in which sufficient information for confirming an ISH was found outside the attending physician notes (e.g., psychiatric attending or behavioral health nurse), suggesting that the PPV could be as high as 91.3, 95% CI (86.6, 94.8%).
The attending physicians documented 79 (38.2%); 95% CI (31.5, 45.2%) cases as ISH injuries with intent to die (i.e., suicide attempt) (Table 2). An additional 16 cases were classified by the study reviewers as suicide attempts using information documented in notes from psychiatric consultations and behavioral health nurse assessments. Based on review of the entire encounter record, the estimated percentage of ISH ED discharge records with documented intent to die (i.e. suicide attempt) was 45.9, 95% CI (38.9, 52.0%).
More than half of the medical records documented suicidal ideation. Based on the review of the entire record for the sampled encounters of care, we estimated that 55.6% (95% CI (48.5, 62.4%)) of the ED discharge records for ISH documented patient expressions of suicidal ideation (Table 2). More than two-thirds of patients had histories of depressed mood disorder (n = 143, 69.1%) and mental health treatment (n = 144, 69.6%). The most commonly documented risk factors for ISH were relationship stressors and lack of social support (n = 138, 66.7%), school-related stressors (n = 108, 52.2%), bullying (n = 44, 21.3%), sexual abuse (n = 36, 17.4%), and physical abuse (n = 25, 12.1%) (Table 3).
The medical records indicated that 119 (57.5%) of the patients were currently in therapy, 132 (63.8%) were taking medications for mental/behavioral conditions, and 71 (34.3%) were on medications for other health conditions. Fifteen patients were maintained on 72-h hold; in another 106 cases a 72-h hold was ordered but then discontinued. In more than 90% of the cases there was documentation of a plan in place for treatment or follow-up service after discharge. Overall, 153 (73.9%) patients were discharged routinely to home/self-care but 23 of them were held initially for observation; the remaining 54 (26.1%) were discharged/transferred to psychiatric or other inpatient units.
Chi-square analyses demonstrated there were significant differences in the prevalence of risk factors associated with self-harm behaviors between those who had documented intention to die (i.e. attempted a suicide) and those who did not (Table 3). Adolescents who attempted suicide were more likely to report suicidal ideation (p < .0001), experience school-related stressor (p = 0.038), experience relationship stressors or lack of social support (p = 0.0006), be victims of physical abuse (p = 0.005), have a history of suicide attempts (p < .0001), have a history of expressed suicidal thoughts (p = 0.008), or have a family history of suicide or suicide attempt (p = 0.0004).
Half (50.5%) of patients with documentation confirming a suicide attempt had a service marker for observation versus 10.7% of patients with confirmed self-harm but no documentation of suicide attempt (p < 0.001). In contrast, there was no statistically significant difference in observation status between those with (33.0%) and without (23.9%; p = 0.15) documented suicidal ideation.