Design: Systematic review of measurement properties using COSMIN methods, QUADAS-2 and COSMIN-adapted GRADE certainty ratings
Population: 41 studies from 8,378 records, covering 15 language versions of six paediatric delirium tools in acute care
Primary outcome: Measurement properties (criterion validity, reliability, convergent validity) and certainty of evidence
Effect: High-certainty evidence for criterion validity of CAPD and Confusion Assessment Method tools with moderate-certainty reliability; SOS-PD moderate-certainty criterion validity with high-certainty convergent validity and reliability in fewer studies; CDAS and PEDS evidence limited to single studies

Delirium in the paediatric intensive care unit is common, under-recognised and associated with worse outcomes, and most units that screen for it have chosen a tool on availability rather than evidence. This systematic review assessed the measurement properties of every available option, appraising 41 studies from 8,378 records and covering 15 language versions of six instruments.

The candidates are the Cornell Assessment of Pediatric Delirium (CAPD), the preschool and paediatric Confusion Assessment Method for the ICU, the Sophia Observation withdrawal Symptoms - Pediatric Delirium scale (SOS-PD), the Pediatric Delirium Scale and the Child Delirium Assessment Scale. CAPD has by far the largest evidence base, and high-certainty evidence supports the criterion validity of CAPD and the Confusion Assessment Method tools, with moderate-certainty evidence for their reliability. SOS-PD had the most favourable overall certainty profile — moderate-certainty criterion validity, high-certainty convergent validity and reliability — but was assessed in fewer studies. The remaining two rest on single studies each.

The limitations are as useful as the ranking. Evidence concentrates on criterion validity and interrater reliability; measurement error, cross-cultural validity and performance in subgroups were rarely assessed. For a unit outside the countries where these tools were developed, cross-cultural validity is precisely the property that matters, and it is largely unstudied — an argument for choosing a tool with a validated version in the language your nurses actually use.

Screen for delirium routinely; the question is which tool, not whether
CAPD is the safest default on breadth of evidence; SOS-PD has the better certainty profile on fewer studies
Avoid the Pediatric Delirium Scale and Child Delirium Assessment Scale as primary tools — single studies each
Check that a validated version exists in the language your nursing staff will use
Expect no gold standard: criterion validity here is measured against imperfect reference diagnoses

Why it matters

Most units chose their delirium tool before this evidence existed, and two of the six options rest on a single study each.

The statistics, in plain English

Certainty ratings here describe how confident we can be in the estimate, not how well the tool performs — high-certainty evidence for criterion validity means the validity has been consistently measured, not that the tool is highly accurate. The authors note there is no perfect gold standard for paediatric delirium, so every criterion validity figure is measured against an imperfect reference, which tends to understate a good tool and flatter a poor one.