External Validation of Pediatric Pneumonia and Bronchiolitis Risk Scores to Predict Mortality in Children Hospitalized in Kenya: A Retrospective Cohort Study.

Abstract

Acute lower respiratory tract infections (ALRIs) are a leading cause of pediatric mortality in low- and middle-income countries. In recent years, substantial research has been done to enhance risk stratification of children presenting with ALRIs, in a bid to improve health outcomes in resource-limited settings. We sought to analyze the performance of several pediatric ALRI risk scores in the prediction of mortality among children hospitalized with ALRIs in Kenya. We retrospectively analyzed the data of 2182 children aged 2-24 months who were admitted to Kilifi County Referral Hospital, Kenya with severe ALRIs between January 2015 and December 2024. We evaluated the performance of 6 ALRI risk scores (RISC [HIV-negative], mRISC, RISC-Malawi, PERCH, PREPARE, and ReSVinet) in this population. Additionally, we created and evaluated a modified version of the ReSVinet score by including nutrition status. Discrimination was assessed using the area under the receiver operating characteristic curve (AUROC). The mid-upper arm circumference (MUAC) version of the RISC-Malawi score showed the highest discrimination for the outcome of in-hospital mortality (AUROC, 0.83; 95% confidence interval, .79-.86), whilst all other scores showed acceptable discrimination (AUROC, 0.70-0.79). The modification of ReSVinet to include nutrition status significantly improved its AUROC from 0.72 to 0.79. All risk scores showed at least fair performance in the prediction of in-hospital mortality within our dataset. The RISC-Malawi (MUAC) score appears to be the most promising candidate for future implementation; however, further research is needed to evaluate the calibration, feasibility, and clinical utility of these scores.