The Preschool Confusion Assessment Method for the ICU: Valid and Reliable Delirium Monitoring for Critically Ill Infants and Children

Abstract

Objectives: Delirium assessments in critically ill infants and young children pose unique challenges due to evolution of cognitive and language skills. The objectives of this study were to determine the validity and reliability of a fundamentally objective and developmentally appropriate delirium assessment tool for critically ill infants and preschool-aged children and to determine delirium prevalence.

Design and setting: Prospective, observational cohort validation study of the PreSchool Confusion Assessment Method for the ICU in a tertiary medical center PICU.

Patients: Participants aged 6 months to 5 years and admitted to the PICU regardless of admission diagnosis were enrolled.

Measurements and main results: An interdisciplinary team created the PreSchool Confusion Assessment Method for the ICU for pediatric delirium monitoring. To assess validity, patients were independently assessed for delirium daily by the research team using the PreSchool Confusion Assessment Method for the ICU and by a child psychiatrist using the Diagnostic and Statistical Manual of Mental Disorders criteria. Reliability was assessed using blinded, concurrent PreSchool Confusion Assessment Method for the ICU evaluations by research staff. A total of 530-paired delirium assessments were completed among 300 patients, with a median age of 20 months (interquartile range, 11-37) and 43% requiring mechanical ventilation. The PreSchool Confusion Assessment Method for the ICU demonstrated a specificity of 91% (95% CI, 90-93), sensitivity of 75% (95% CI, 72-78), negative predictive value of 86% (95% CI, 84-88), positive predictive value of 84% (95% CI, 81-87), and a reliability κ-statistic of 0.79 (0.76-0.83). Delirium prevalence was 44% using the PreSchool Confusion Assessment Method for the ICU and 47% by the reference rater. The rates of delirium were 53% versus 56% in patients younger than 2 years old and 33% versus 35% in patients 2-5 years old using the PreSchool Confusion Assessment Method for the ICU and reference rater, respectively. The short-form PreSchool Confusion Assessment Method for the ICU maintained a high specificity (87%) and sensitivity (78%) in post hoc analysis.

Conclusions: The PreSchool Confusion Assessment Method for the ICU is a highly valid and reliable delirium instrument for critically ill infants and preschool-aged children, in whom delirium is extremely prevalent.

Figure 1

This CONSORT figure demonstrates participant flow including: initial screening, exclusions, eligibility, and the final enrolled cohort for the investigation. ^aEight patients were determined upon initial assessment by the reference standard to have cognition less than expected for 6 months of age after consent was obtained. Withdrawal was discussed with the family, no further delirium evaluation was completed, and obtained data was destroyed.

Figure 2

The Preschool Confusion Assessment Method for the ICU (psCAM-ICU). The short-form (clinical) psCAM-ICU is used to assess for delirium in infants and children who are at least responsive to voice. Following the psCAM-ICU algorithm, once three features have been identified as “YES” or present then the patient has delirium and the assessment is complete. The cardinal features of an acute change or fluctuating mental status and inattention (Features 1 and 2) are required for delirium diagnosis. The presence of an acute altered level of consciousness (Feature 3) or a sleep wake cycle disturbance (Feature 4) may provide further evidence of system dysregulation, but do not alone assure presence of delirium. Compared to the reference standard (child psychiatrist using DSM criteria), the short form of the psCAM-ICU, requiring fewer assessment components within Features 2 and 4, performed with a high specificity of 86% (84, 88) and sensitivity of 78% (75, 80), similar to using the more elaborate diagnostic criteria in the long-form (research) psCAM-ICU.