Validation of Pediatric ALRI Risk Scores for Mortality Prediction in Kenyan Children

Overview

This retrospective cohort study evaluated six pediatric acute lower respiratory tract infection (ALRI) risk scores in predicting in-hospital mortality among 2182 children aged 2–24 months hospitalized in Kenya. The RISC-Malawi score incorporating mid-upper arm circumference (MUAC) demonstrated the highest discrimination (AUROC 0.83), while modification of the ReSVinet score to include nutritional status improved its performance significantly.

Background

Acute lower respiratory tract infections, including pneumonia and bronchiolitis, remain a leading cause of pediatric mortality in low- and middle-income countries, accounting for over 700,000 deaths globally in children under five years. In Kenya, ALRIs cause approximately 5,000 deaths annually in this age group. Accurate risk stratification tools are essential to identify children at high risk of mortality to optimize clinical management and resource allocation. Several risk scores have been developed and validated in LMICs, but comparative performance data within the same population are limited.

Data Highlights

Key Findings

• The RISC-Malawi score incorporating MUAC had the highest discrimination for predicting in-hospital mortality (AUROC 0.83).
• All other evaluated risk scores demonstrated acceptable discrimination with AUROC values ranging from 0.70 to 0.79.
• Modification of the ReSVinet bronchiolitis severity score to include nutritional status significantly improved its predictive performance (AUROC increased from 0.72 to 0.79).
• These risk scores were validated in a large Kenyan pediatric cohort hospitalized with severe ALRIs, enhancing generalizability in LMIC settings.
• LIBSS score was excluded due to substantial missing data in the dataset.

Clinical Implications

The RISC-Malawi (MUAC) score shows promise as a practical tool for mortality risk stratification in children hospitalized with ALRIs in resource-limited settings. Incorporating nutritional status into existing severity scores, such as ReSVinet, can enhance predictive accuracy. Clinicians should consider these validated scores to identify high-risk patients early and optimize resource allocation, though further research on calibration and feasibility is warranted before widespread implementation.

Conclusion

All evaluated pediatric ALRI risk scores demonstrated fair to good discrimination for in-hospital mortality in Kenyan children, with the RISC-Malawi (MUAC) score performing best. Incorporating nutritional status into severity scores improves prediction, supporting their potential utility in LMIC clinical settings.

References

1. Authors et al. 2024 -- Validation of Pediatric Risk Scores for Pneumonia and Bronchiolitis in Predicting Mortality Among Hospitalized Children in Kenya