pCAM-ICU: Pediatric Confusion Assessment Method for the ICU

The original study enrolled 68 critically ill children with developmental age of at least five years and obtained 146 paired assessment-days against psychiatrist DSM-IV-TR diagnosis. pCAM-ICU sensitivity was 83%, specificity 99% and inter-rater kappa 0.96; assessments are not interchangeable with a denominator of 146 independent children.

This systematic review examined five paediatric intensive-care diagnostic-accuracy studies and judged all to have high risk of bias. For pCAM-ICU it reported approximately 78% sensitivity and 98% specificity in 68 children and called for direct, methodologically stronger comparisons.

This study compared paediatric delirium tools with expert DSM-IV-TR diagnosis in 64 children aged five years or older and 214 assessments. For pCAM-ICU, first-assessment sensitivity was 76.9% and specificity 98%, while sensitivity across repeated assessments fell to 52.3%, showing that covariates and analysis unit materially affect performance.

This Brazilian study included 116 children aged 5-17 years across eight PICUs and used neurologist or psychiatrist DSM-5 diagnosis. Sensitivity was 90.9%, specificity 99.3% and inter-rater agreement perfect, but delirium prevalence was only 7% and children under pharmacological sedation were excluded.

This prospective Australian PICU study compared routinely delivered pCAM-ICU and CAPD assessments with DSM-5 diagnosis. pCAM-ICU sensitivity was 75.0% and specificity 84.9%, notably less specific than in tightly controlled validation studies and therefore informative about real-world delivery.

This meta-analysis pooled four pCAM-ICU studies and reported sensitivity of 0.73 and specificity of 0.98. Age, sample characteristics and mechanical ventilation moderated performance, supporting pCAM-ICU as highly specific but not a reliable standalone rule-out test.