G242(P) Does the sepsis screening tool used in the paediatric emergency department highlight children at risk of sepsis and what is the relationship of initial bpews score to patient outcome

Aims This study of children seen in the paediatric emergency department set out to determine how useful the sepsis screening tool was in alerting clinicians to unwell potentially septic children, and to establish if the initial BPEWS (beside paediatric early warning score) was predictive of admission. Methods A retrospective study of the electronic patient records for patients who had triggered on the sepsis screening tool over a 4 month period. Information gathered included initial BPEWS with breakdown for each parameter, and the outcome – discharged or admission (admission included children observed in the clinical decision unit or use of the nursing outreach service). Of those who required admission, further information was gathered as to final diagnosis, whether antibiotics were given (IV or oral) and the microbiology results. Results In the study period 719 patients triggered on the sepsis screening tool of which 713 were analysed. 172 of these were admitted (24%), and of these only 1 had an invasive bacteraemia. Respiratory infections were a common cause of illness with 19% (33) having upper respiratory tract infection and 17% (29) having lower. 16% (27) of those admitted were ultimately diagnosed with viral induced wheeze. 63 (37%) were positive for one or more viruses on respiratory or stool samples. Correlation analysis of the BPEWS showed that the higher the initial BPEWS, the more likely that children would be admitted. There was a correlation of 95% between these variables and chi squared analysis of the relationship showed a significant p value of Conclusion Whilst large numbers of children triggered on the sepsis screening tool, 76% were not admitted and thus raises questions as to the sensitivity of the tool. Increasing BPEWS was found to be predictive in identifying children who require admission. This suggests that perhaps the tool requires some adaptations, such as using high BPEWS as a significant trigger, to allow it to be more sensitive and applicable for this population and workforce.