Identifying risk of death in children hospitalized with community-acquired pneumonia

Abstract

To externally validate a tool developed by the Pneumonia Research Partnership to Assess WHO Recommendations study group for identification of the risk of death in children hospitalized with community-acquired pneumonia, the PREPARE tool. We did a secondary analysis of data collected during hospital-based surveillance of children with community-acquired pneumonia in northern India from January 2015 to February 2022. We included children aged 2-59 months with pulse oximetry assessment. We used multivariable backward stepwise logistic regression analysis to assess the strength of association of the PREPARE variables (except hypothermia) with pneumonia-related death. We estimated sensitivity, specificity, and positive and negative likelihood ratios of the PREPARE score at cut-off scores ≥ 3, ≥ 4 and ≥ 5. Of 10 943 children screened, 6745 (61.6%) were included in our analysis, of whom 93 (1.4%) died. Age of < 1 year, female sex, weight-for-age < -3 standard deviations, respiratory rate of ≥ 20 breaths/min higher than the age-specific cut-off, and lethargy, convulsions, cyanosis and blood oxygen saturation < 90% were associated with death. In the validation, the PREPARE score had the highest sensitivity (79.6%) with concurrent highest specificity (72.5%) to identify hospitalized children at risk of death from community-acquired pneumonia at a cut-off score of ≥ 5. Area under curve was 0.82 (95% confidence interval: 0.77-0.86). The PREPARE tool with pulse oximetry showed good discriminatory ability on external validation in northern India. The tool can be used to assess risk of death of hospitalized children aged 2-59 months with community-acquired pneumonia for early referral to higher-level facilities.