Evaluating the efficacy of the SICK score in predicting clinical outcomes among paediatric patients presenting to the emergency department of a tertiary care centre in Kerala

Abstract

Background: Early risk stratification in the paediatric emergency department (ED) is crucial for optimal resource use and prevention of avoidable mortality. The signs of inflammation in children that kill (SICK) score is a bedside tool based on observable physiological parameters. However, evidence regarding its role in guiding level-of-care decisions and its association with hospital stay is limited. This study evaluated its ability to predict pediatric intensive care unit (PICU) admission and its correlation with length of stay. Methods: This prospective observational study was conducted over 18 months in a tertiary pediatric ED in Kerala. A total of 340 children aged 1 month to 12 years were consecutively enrolled. SICK scores were calculated at presentation. Patients were triaged to ward or PICU based on clinical assessment. Data were analyzed using descriptive statistics, independent t-test, chi-square test, Pearson correlation, and ROC curve analysis. A p value <0.05 was considered significant. Results: Of 340 children, 40 (11.8%) required PICU admission. PICU patients were younger (43.8 versus 74.0 months; p<0.001). Mean SICK score was higher in the PICU group (5.23±1.69 versus 2.62±2.11; p<0.001). Hospital stay was longer in PICU patients (9.13±2.61 versus 2.62±1.15 days; p<0.001). SICK score showed a moderate positive correlation with length of stay (r=0.324; p<0.001). ROC analysis showed good discrimination (AUC=0.820; 95% CI: 0.756–0.884), with an optimal cut-off ≥3.5 (sensitivity 90.0%, specificity 71.7%). Conclusions: The SICK score is a reliable tool for predicting PICU admission and correlates with hospital stay, supporting its use in paediatric ED triage, especially in resource-limited settings.