Performance of Cornell Assessment of Pediatric Delirium Scale in Mechanically Ventilated Children

Accuracy of delirium diagnosis in mechanically ventilated children is often limited by their varying developmental abilities. The purpose of this study was to examine the performance of the Cornell Assessment of Pediatric Delirium (CAPD) scale in these patients. This is a single-center prospective observational study of patients requiring sedation and mechanical ventilation for 2 days or more. CAPD scale was implemented in our unit for delirium screening. Each CAPD assessment was accompanied by a physician assessment using Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-V) criteria. Sensitivity analysis was performed to determine the best cut-off score in our target population. We also evaluated ways to improve the accuracy of this scale in patients with and without developmental delay. A total of 837 paired assessments were performed. Prevalence of delirium was 19%. Overall, CAPD score >= 9 had sensitivity of 81.8% and specificity of 44.8%. Among typically developed patients, the sensitivity and specificity were 76.7 and 65.4%, respectively, whereas specificity was only 16.5% for developmentally delayed patients. The best cut-off value for CAPD was 9 for typically developed children and 17 for those with developmental delay (sensitivity 74.4%, specificity 63.2%). Some CAPD questions do not apply to patients with sensory and neurocognitive deficits; upon excluding those questions, the best cut-off values were 5 for typically developed and 6 for developmentally delayed children. In mechanically ventilated patients with developmental delay, CAPD >= 9 led to a high false-positive rate. This emphasizes the need for either a different cut-off score or development of a delirium scale specific to this patient population.