Assessing the performance of paediatric early warning scores to predict critical deterioration events in hospitalised children (the DETECT study): a retrospective matched case-control study.

backgroundPaediatric Early Warning Scores (PEWS) enhance patient safety, by focused monitoring of vital signs to identify children at risk of deteriorating. However, there is an acknowledged need for standardisation. The aim of this study was to compare the performance of seven PEWS (Alder Hey, Bedside, Bristol, Irish, Newcastle, Scottish and the proposed National PEWS for England (v3)) utilised in clinical practice in the United Kingdom and Ireland. The primary outcome was occurrence of a critical deterioration event (CDE) in hospitalised children, and secondary outcome was 72-hour hospital mortality.

methods250 patient episodes were identified over a 12-month period. Cases were matched 2:1 with controls; using age range and admitting specialty. PEWS were calculated, along with performance characteristics. Maximum PEWS were calculated at 24, 12, 6 and 4 h prior to CDE or discharge, and area under the receiver operating curve (AUC) used to measure performance. Sub-group analysis explored performance within 3 specialities observed to have increased risk for deterioration. Kaplan-Meier survival curves compared time to event data using the identified optimum PEWS performance cut-point.

resultsThe median age of patients experiencing CDE was 8 months. AUCs across all PEWS in predicting CDE, ranged from 0·87 to 0·95. Optimum cut-offs for each PEWS were identified. Kaplan-Meier curves for cumulative risk of time to CDE according to the PEWS stratification, demonstrated CDE was significantly less likely for patients below the cut-off values (log-rank test, p < 0·001).

conclusionsAll seven PEWS assessed demonstrate excellent predictive ability for CDE, in a heterogenous cohort. For evaluation of PEWS performance, CDE is a more appropriate outcome measure than hospital mortality, due to low mortality outside PICU. A standardised PEWS allows consistency, benchmarking and opportunity for continuing recalibration.