A Scoring System to Predict Severe Acute Lower Respiratory Infection in Children Caused by Respiratory Syncytial Virus

Abstract

There were several factors associated with respiratory syncytial virus (RSV) severe acute lower respiratory infection (RSV-sALRI) in infants and young children. It is vital to develop a convenient scoring system to predict RSV-sALRI in children. Pediatric patients with RSV-ALRI from January 2009 to December 2021 were recruited retrospectively. Two-third of them were randomly grouped into the development set and one-third to the validation set. In the development set, risk factors for RSV-sALRI were transferred into the logistic regression analysis, then their receiver operating characteristic (ROC) curves were built to obtain the area under the ROC curve (AUC), and regression coefficients for each predictor were converted to points. Finally, the value of the scoring system was evaluated in the validation set. A total of 1 066 children with RSV-ALRI were recruited, including 710 in the development set and 356 in the validation set. By logistic regression analysis, six factors (younger than 2 years, gestational age <37 weeks, have siblings, birth weight ≤2500 g, artificial/mix feeding, CHD) showed statistical difference and then were scored with points according to the coefficient value (OR) in the development set. In the validation set, the sensitivity of the scoring system was 70.25%, the specificity 85.53%, the positive predictive value 71.43%, the negative predictive value 84.81%, and coincidence rate 0.80. The Kolmogorov–Smirnov test showed the distribution of AUC 0.765 (SE = 0.027; 95% CI = 0.713–0.818; p < 0.001). A simplified scoring system was developed in the study with high prediction value for RSV-sALRI in children.