Nomogram prediction model for antimicrobial-associated diarrhea in children with severe pneumonia
ZHANG Ke
LI Ximei
SU Yanyan
TANG Yu
Abstract：Objective To investigate the risk factors for antibiotic-associated diarrhea(AAD)in children with severe pneumonia,construct a Nomogram prediction model,and explore corresponding prevention strategies.Methods A total of 293 children with severe pneumonia admitted to Henan Children's Hospital Zhengzhou Children's Hospital from January 2022 to December 2024 were retrospectively enrolled and divided into an AAD group and a non-AAD group based on the occurrence of AAD.Univariate and multivariate Logistic regression analyses(stepwise backward method)were performed to identify the inde-pendent predictors.R software packages were used to develop the Nomogram model.The model's predictive performance,cali-bration,internal validation,and clinical benefit were assessed using the receiver operating characteristic(ROC)curve,calibration curve,Hosmer-Lemeshow test,Bootstrap resampling(n=1 000),and decision curve analysis.Risk stratification was performed to validate the model's effectiveness.Results There were 89 cases(30.38%)in the AAD group and 204 cases(69.62%)in the non-AAD group.Multivariate Logistic regression analysis revealed that erythromycin and mechanical ventilation were excluded due to statistical insignificance.Age(≤2 years),C-reactive protein(CRP)level,use of cefoperazone-sulbactam,use of azithromycin,medication duration(≥5 days),and concurrent use of multiple antibiotics(≥3 types)were independent predictors of AAD in children with severe pneumonia(OR=6.836,1.703,3.850,3.196,2.061,and 6.082,respectively,all P<0.05).The Nomogram model built with these factors had an area under the ROC curve(AUC)of 0.812.The predicted probabilities were relatively close to the actual probabilities,indicating good model calibration(χ2=8.975,P=0.342).The corrected concordance index(C-index)was 0.770,and the model demonstrated good net clinical benefit.Risk stratification validation showed that the actual AAD incidence rates in the high-risk group(n=93,prediction probability>50.00%),medium-risk group(n=120,prediction probability 20.00%-50.00%),and low-risk group(n=80,prediction probability<20.00%)was 55.91%,23.33%,and 11.25%,respectively.Conclusion Age(≤2 years),CRP lev-el,use of cefoperazone-sulbactam,use of azithromycin,duration of antibiotic therapy(≥5 days),and concurrent use of multiple antibi-otics(≥3 types)are independent predictors of AAD in children with severe pneumonia.The constructed Nomogram prediction model demonstrates good predictive performance and can facilitate early identification of high-risk pediatric patients.
Keywords：severe pneumoniachildrenantibiotic-associated diarrheaNomogram prediction modelprevention strategies
Publication Date：2026-03-25
Online Publishing Date：2026-04-07(First online date of this platform, not the publication date of the document)
Pages：5( 305-309 )