ABSTRACT 106

Abstract

Background and aims: Paediatric Early Warning tools are increasingly used as aids to identify in-patients at risk of deterioration. Most tools are based on local consensus opinion and unvalidated. A modified Bristol PEW tool was introduced at a UK tertiary paediatric hospital in 2006. Aims: To use logistic and mutivariate regression to evaluate which components of the PEW were most predictive of developing a significant adverse event (SAE); respiratory or cardiac arrest, unplanned admission to PICU or unexpected death. Methods: Odds ratio and significance were calculated using 5 year reported data for 15 components of a PEW. A 7 component multiple regression model was developed giving odds ratio and confidence intervals. Results:Table: No title available.Conclusions: Clinically tiring, acidosis, age and tachycardia were the most important individual predictors of deterioration. Conversely, if the nurse was worried that a patient was deteriorating, there was a lower risk of a SAE, because the nurse would actively seek out urgent medical attention. Further education must be targeted specifically towards ward nurses so they recognise the common patterns of paediatric deterioration, and reduce the incidence of missed deterioration.