Construction and validation of a prediction model for small airway dysfunction in children with bronchial asthma

LIU Yanli; XU Wenfu; FANG Hongjuan

doi:10.11958/20241912

Abstract：Objective To analyze influencing factors of small airway dysfunction(SAD)in children with bronchial asthma,and construct a risk prediction model for SAD.Methods A total of 221 bronchial asthma children with normal pulmonary ventilation function were retrospectively selected as the modeling group.According to the presence of SAD,children were divided into the SAD group(43 cases)and the non-SAD group(178 cases).The related factors affecting the occurrence of SAD were analyzed by univariate analysis and multivariate Logistic regression analysis,and risk prediction model of SAD was constructed.In addition,74 bronchial asthma children with normal pulmonary ventilation function were collected and used as the validation group.Hosmer-Lemeshow test was adopted to evaluate the calibration of the model.Receiver operating characteristic(ROC)curve was drawn and area under the curve(AUC)was calculated to evaluate the predictive efficiency of the prediction model.Results Multivariate Logistic regression analysis showed that poor asthma control(OR=10.722,95%CI:4.044-28.428),allergic rhinitis in the attack stage(OR=4.816,95%CI:1.929-12.023)and long use of inhaled corticosteroids(ICS)or ICS/long-acting β2 receptor agonist(LABA,OR=2.903,95%CI:1.269-6.643)were independent risk factors for SAD in children(P＜0.05),and older age at first onset of suspected asthma symptoms(OR=0.599,95%CI:0.395-0.907)was a protective factor.Hosmer-Lemeshow test showed good calibration(χ2=8.301,P=0.307).The AUC,sensitivity and specificity of the model in the modeling group for predicting the occurrence of SAD in children with bronchial asthma were 0.820(95%CI:0.749-0.892),74.42%and 78.65%,and the AUC,sensitivity,specificity and accuracy of the model in the validation group were 0.849(95%CI:0.718-0.981),75.00%,94.83%and 90.54%(67/74).Conclusion Children with bronchial asthma who are younger at the time of their first suspected asthma symptoms,poor asthma control,allergic rhinitis in the exacerbation phase,and longer duration of use of ICS or ICS+LABA/LTRA are at higher risk of SAD.Therefore,SAD risk prediction model constructed on this basis has good predictive ability.