Yang Shen1, Yao Zhou1, Cuian Ma1, Yuqiao Liu1, Botao Wei1. 1. Department of Infectious Disease, Tianjin Children's Hospital/Tianjin University Children's Hospital, Tianjin, China.
Abstract
BACKGROUND: Severe adenovirus pneumonia (SAP) of children is prone to multi-system complications, has the high mortality rate and high incidence of sequelae. Severity prediction can facilitate an adequate individualized treatment plan. Our study try to develop and evaluate a predictive nomogram for children with SAP. METHODS: An observational study was designed and performed retrospectively. The data were categorized as training and validation datasets using the method of credible random split-sample (split ratio =0.7:0.3). The predictors were selected using Lasso (least absolute shrinkage and selection operator) logistic regression and the nomogram was developed. Nomogram discrimination was assessed using the receiver operating characteristic (ROC) curve, and the prediction accuracy was evaluated using a calibration curve. The nomogram was also evaluated for clinical effectiveness by the decision curve analysis (DCA). A P value of <0.05 was deemed statistically significant. RESULTS: The identified predictors were fever duration, and interleukin-6 and CD4+ T cells and were assembled into the nomogram. The nomogram exhibited good discrimination with area under ROC curve in training dataset (0.79, 95% CI: 0.60-0.92) and test dataset (0.76, 95% CI: 0.63-0.87). The nomogram seems to be useful clinically as per DCA. CONCLUSIONS: A nomogram with a potentially effective application was developed to facilitate individualized prediction for SAP in children.
BACKGROUND: Severe adenovirus pneumonia (SAP) of children is prone to multi-system complications, has the high mortality rate and high incidence of sequelae. Severity prediction can facilitate an adequate individualized treatment plan. Our study try to develop and evaluate a predictive nomogram for children with SAP. METHODS: An observational study was designed and performed retrospectively. The data were categorized as training and validation datasets using the method of credible random split-sample (split ratio =0.7:0.3). The predictors were selected using Lasso (least absolute shrinkage and selection operator) logistic regression and the nomogram was developed. Nomogram discrimination was assessed using the receiver operating characteristic (ROC) curve, and the prediction accuracy was evaluated using a calibration curve. The nomogram was also evaluated for clinical effectiveness by the decision curve analysis (DCA). A P value of <0.05 was deemed statistically significant. RESULTS: The identified predictors were fever duration, and interleukin-6 and CD4+ T cells and were assembled into the nomogram. The nomogram exhibited good discrimination with area under ROC curve in training dataset (0.79, 95% CI: 0.60-0.92) and test dataset (0.76, 95% CI: 0.63-0.87). The nomogram seems to be useful clinically as per DCA. CONCLUSIONS: A nomogram with a potentially effective application was developed to facilitate individualized prediction for SAP in children.