Paediatric pneumonia: deriving a model to identify severe disease

Area of Science:

• Pediatric Pulmonology
• Infectious Diseases
• Clinical Epidemiology

Background:

• Community-acquired pneumonia (CAP) is a significant cause of pediatric hospitalizations.
• Limited data and tools exist for predicting severe CAP in children.
• There is a need for pediatric-specific predictive tools for CAP severity.

Purpose of the Study:

• To identify clinical factors predicting severe outcomes in hospitalized children with CAP.
• To develop and validate a clinical risk prediction tool for severe pediatric CAP.
• To stratify children at risk for invasive or intensive care interventions.

Main Methods:

• Retrospective cohort study of 3330 hospitalized CAP cases (2011-2016).
• Analysis of clinical variables at hospital presentation against stratified outcomes (mild, moderate, severe).
• Development and validation of a clinical risk stratification tool using logistic regression and ROC analysis.

Main Results:

• Hypoxia, increased work of breathing, comorbidities, and indigenous status were associated with moderate-severe CAP outcomes.
• Febrile children had a lower likelihood of moderate-severe outcomes.
• The derived clinical tool achieved an AUC of 0.72 for predicting severe disease, with high scores indicating significantly increased odds of moderate-severe outcomes.

Conclusions:

• A clinical risk prediction tool is essential for managing pediatric CAP.
• Identified risk factors and a novel clinical tool can stratify children's risk of severe CAP.
• The developed tool demonstrates comparable performance to adult pneumonia prediction tools.