Abstract
<title>Abstract</title> <p>Multi-informant assessment is central to child and adolescent psychiatry, yet adolescent and caregiver reports of suicide-related symptoms may not be interchangeable. We examined parent–adolescent agreement and the clinical relevance of discrepant reports among patients aged 11–18 years hospitalized with depressive disorders. The primary analytic cohort included 229 participants with paired first-baseline assessments and complete age, sex, Nurses’ Global Assessment of Suicide Risk (NGASR), and Columbia-Suicide Severity Rating Scale (C-SSRS) data. Agreement was quantified using Spearman correlation, an absolute-agreement intraclass correlation coefficient, quadratic weighted kappa, and item-level positive and negative agreement. Confirmed endorsements, excluding caregiver “unknown” responses, defined suicide-report patterns in 205 participants. NGASR and C-SSRS were treated as same-window clinical anchors rather than prospective outcomes. Agreement for the platform-exported suicide domain was modest (Spearman ρ = 0.321, ICC = 0.318, weighted κ = 0.317). Item-level κ was 0.321 for recent suicidal ideation or planning and 0.303 for lifetime suicide attempt. Confirmed endorsement patterns were neither informant (15.1%), parent only (11.2%), adolescent only (21.0%), and both informants (52.7%). In 205 classifiable participants, confirmed endorsement by each informant was associated with NGASR in the joint model (B = 2.53 for adolescents and B = 1.77 for parents). However, these same-window associations partly reflected overlapping content. Agreement estimates were consistent across sensitivity analyses of timing, age range, and scoring. Parent and adolescent reports overlapped but were not interchangeable. Disagreement should prompt clinical clarification, although these data do not establish future-risk prediction.</p>