Identifying undiagnosed high-risk suicidality cases through comorbidity-adjusted risk modeling

Area of Science:

• Pediatric Emergency Medicine
• Mental Health Research
• Health Informatics

Background:

• Pediatric suicidality is a critical public health issue, underreported and posing challenges for intervention.
• Identifying at-risk pediatric populations is crucial for effective prevention and resource allocation.
• Understanding comorbidities associated with suicidality aids in early detection and targeted support.

Purpose of the Study:

• To determine the prevalence of mental health comorbidities linked to pediatric suicidality.
• To utilize propensity score matching to identify undiagnosed high-risk pediatric suicidality cases.
• To enhance the accuracy of identifying youth experiencing suicidal ideation, self-harm, or attempts.

Main Methods:

• Analysis of ICD-10 diagnosis codes for pediatric patients (ages 6-18) in an emergency department setting.
• Defined suicidality cases including ideation, self-harm, and attempt subtypes.
• Employed propensity score matching to link diagnosed suicidality cohorts with undiagnosed at-risk individuals.

Main Results:

• 2.9% of emergency department encounters met the suicidality case definition.
• Identified 55 significant comorbidities, with many not aligning with current diagnostic codes.
• High probabilities of suicidality diagnosis were associated with personality disorder, gender dysphoria, bipolar disorder, depression, and schizophrenia spectrum disorders.
• Propensity matching correctly identified 53.4% of true positive suicidality cases.

Conclusions:

• Propensity score matching effectively identifies undiagnosed pediatric suicidality cases.
• Comorbidity profiles of matched cases closely resemble those with diagnosed suicidality.
• This method offers a valuable tool for improving the identification and support of at-risk youth.