Predictive Accuracy of Paediatric Trauma Score, our experience at Children Hospital, Lahore
Aim: To find predictive accuracy of pediatric trauma score (PTS) in terms of mortality in the Children’s hospital and institute of child health Lahore.
Study design: Cross sectional study.
Setting: Paediatric Surgery department, CH & ICH, Lahore.
Duration of study: 1st January 2019 to 30th June 2019.
Methods: All patients presented with trauma, aged 1-15 years were included in study. All patients with major burn, associated cardiac injuries (e.g., stab heart), victims who required monitoring and hospitalization less than 24 hours, and those referred to other hospitals were excluded from the study. Initial assessment of all patients was done and paediatric trauma score was evaluated in the emergency room. Patients were managed and 24 hours follow-up was done to predict the mortality.
Results: From 290 children, 207(71.38%) were males and 83 (28.6%) were females. Mean age of the children was 7.53 years. The most common mode of trauma was Car accident in children 67(23.1%), followed by Auto rickshaw vs. motorcycle n= 50 (17.24%), and auto rickshaw vs. pedestrian 40(13.79%). Mean Hospital Arrival Time in Minutes was 100.31 minutes. Mean paediatric trauma score amongst all children was 7.22. Overall mortality was reported in 27(9.3%) children. PTS < 8 turned out to be statistically significantly related to mortality (P value < 0.001). Patients with diagnostic test characteristics of PTS < 8 to predict mortality were having sensitivity 91.0%, specificity 56.1%, positive predictive value 38.4%, negative predictive value 95.4% and the diagnostic accuracy of 64.1%.
Conclusion: By Paediatric trauma score we can predict about the necessity of critical interventions in patients and also, we can predict percentage of mortality in paediatric trauma patients. However, accuracy of paediatric trauma score seems to be moderate in detecting most severe paediatric trauma cases.
Keywords: PTS, mortality, hospital arrival time, morbidity, major burn predictor.