Accuracy of Risk Assessment Tool in Predicting Pneumonia’s Outcome among Egyptian Children: Hospital Based Study

Abstract

Aim:To determine possible factors associated with lethal outcome of pneumonia and to assess the accuracy of Pneumonia Severity Index (PSI) and Pediatric Risk of Mortality (PRISM) score in predicting mortality from pneumonia. Study Design: A retrospective analytical study Place and Duration of the Study: Pediatric Emergency Department (PED) of the pediatric hospital (Abu El-Reesh) Egypt, during a period from April 2010 to April 2012. Methodology: Children ≤5 years admitted to the PED diagnosed having pneumonia were included in the study (n=236). Data were retrieved from the electronic records and consisted of; hospital data, personal data, provisional and definite diagnosis, presenting clinical symptoms and signs, outcome and measurements of blood counts and serum biochemical markers. Results: Non-survivors constituted 26.7% of the studied group. Non-survivors significantly had a higher median PRISM score (18; IQR 11 for non-survivors compared to 8; IQR 6 for survivors, P =.000), have a longer median length of stay (8 days; IQR; 1 day for non-survivors compared to 4 days; IQR 2 days for survivors, P =.000), higher PSI score (61; IQR 39 for non-survivors compared to 41; IQR 20 for survivors, P=.000).Only longer LOS, higher PRISM score were independently associated with mortality. ROC curve analysis revealed area under the curve (AUC) of 0.857 for PRISM score (95% CI 0.80–0.91) and 0.73.6 for PSI score (95% CI 0.66–0.81). A PRISM score ≥ 12.5 is 81.4% sensitive and 73.3% specific in predicting mortality. Conclusion: Case fatality rate is quite high. PRISM scoring is accurate in predicting mortality among pneumonia pediatric patients and thus useful in decision making concerning management of these cases.