Palliative and Supportive Care

Original Articles

Pediatric delirium: Evaluating the gold standard

Gabrielle Silvera1 c1, Julia Kearneya2, Chani Traubea1, Thomas M. Atkinsona2, Katarzyna E. Wykaa1a3 and John Walkupa1

a1 Weill Cornell Medical College, New York, New York

a2 Memorial Sloan-Kettering Cancer Center, New York, New York

a3 City University of New York, New York, New York


Objective: Our aim was to evaluate interrater reliability for the diagnosis of pediatric delirium by child psychiatrists.

Method: Critically ill patients (N = 17), 0–21 years old, including 7 infants, 5 children with developmental delay, and 7 intubated children, were assessed for delirium using the Diagnostic and Statistical Manual–IV (DSM–IV) (comparable to DSM–V) criteria. Delirium assessments were completed by two psychiatrists, each blinded to the other's diagnosis, and interrater reliability was measured using Cohen's κ coefficient along with its 95% confidence interval.

Results: Interrater reliability for the psychiatric assessment was high (Cohen's κ = 0.94, CI [0.83, 1.00]). Delirium diagnosis showed excellent interrater reliability regardless of age, developmental delay, or intubation status (Cohen's κ range 0.81–1.00).

Significance of results: In our study cohort, the psychiatric interview and exam, long considered the “gold standard” in the diagnosis of delirium, was highly reliable, even in extremely young, critically ill, and developmentally delayed children. A developmental approach to diagnosing delirium in this challenging population is recommended.

(Received December 26 2013)

(Accepted January 23 2014)


  • Delirium;
  • Child psychiatry;
  • Pediatric critical care


c1 Address correspondence and reprint requests to: Gabrielle Silver, Consultation Liaison Child Psychiatry, Weill Cornell Medical College/New York Presbyterian Hospital, 525 East 68th Street, Box 140, New York, New York 10065. E-mail: