Background: Pediatric patients present to the emergency departments (EDs) with a wide range of clinical manifestations, ranging from mild to severe. A systematic approach is crucial to identify those at high risk of deterioration. However, the predictive value of such predictors remains unclear.
Objectives: Our study aims to evaluate different mortality predictors used in pediatric emergency departments (PEDs) regarding the diagnostic accuracy metrics, including sensitivity, specificity, and diagnostic odds ratio.
Methods: We comprehensively searched multiple databases and included all cohort studies, case-control studies, and randomized controlled trials from January 2000 to December 2024 with pediatric patients (aged 0-18 years) presenting to PEDs, where mortality predictors were used to assess for in-ED and short-term post-ED mortality. We employed a bivariate random-effects model for data synthesis and analysis to calculate pooled sensitivity, specificity, diagnostic odds ratio (DOR), and area under the curve (AUC) values.
Results: 329 Pediatric Early Warning Score (PEWS) thresholds were analyzed, with the model-derived optimal cutoff 2.189 (AUC) = 0.70; 95% CI: 0.63 to 0.76), high pooled sensitivity (0.95, 95% CI: 0.72 to 0.80) and specificity (0.93, 95% CI: 0.62 to 0.80). In addition, A strong negative predictive value (NPV = 0.0006) and modest positive predictive value (PPV = 0.0003) were noted. Heterogeneity was significant (I²>99%), driven by PEWS versions and clinical settings. PEWS implementation reduced mortality without increasing ICU admissions in resource-limited settings.
Conclusions: PEWS is a good exclusion tool for those at low mortality risk. However, a comprehensive approach with clinical judgment is needed for the risk assessment of high-risk pediatrics.
Keywords: Clinical predictors; Diagnostic accuracy; Early warning scores; Emergency department; Pediatric mortality predictors; Risk stratification.
© 2025. The Author(s).