Influencing factors and predictive model construction of recurrence within one year after discontinuation of medication in children with cough variant asthma after regular treatment

Objective To analyze the influencing factors of recurrence within one year after discontinuation of regular treatment of children's cough variant asthma (CVA) and to construct a predictive model, in order to provide guidance for doctors to evaluate the risk of recurrence after CVA treatment in children and develop prevention strategies.

Methods A total of 215 children with CVA were selected as the study objects. According to whether recurrence occurred within one year after discontinuation of regular treatment, the patients were divided into recurrence group (n=89) and non-recurrence group (n=126). Data of two groups were collected for univariate and multivariate Logistic regression analysis to screen out the influencing factors of recurrence in children with CVA. The GBM prediction model was constructed by using R software to run the gradient boosting model (GBM) algorithm. Receiver operating characteristic (ROC) curve was used to analyze and compare the predictive efficiency of GBM model and Logistic regression model.

Results The recurrence rate of 215 children with CVA within one year after regular treatment was 41.40%(89/215). Allergy history (OR=2.870, 95%CI, 1.442 to 5.714), repeated respiratory infection history (OR=4.132, 95%CI, 2.045 to 8.348), parents' cognitive level of CVA prevention and control (OR=1.063, 95%CI, 1.011 to 1.118), eosinophils (OR=0.801, 95%CI, 0.674 to 0.952), forced expiratory flow at 50% of vital capacity (FEF₅₀) (OR=1.055, 95%CI, 1.012 to 1.099), maximal midexpiratory flow rate 75/25 (MMEF_75/25) (OR=1.054, 95%CI, 1.009 to 1.102), and fractional exhaled nitric oxide (FeNO) (OR=0.883, 95%CI, 0.834 to 0.935) were influencing factors for recurrence within one year after regular treatment of children's CVA (P < 0.05). According to ROC curve analysis, the area under the curve of GBM model was 0.872, which was larger than 0.827 of Logistic regression model.

Conclusion GBM prediction model is established based on the influencing factors of recurrence within one year after regular treatment of CVA in children, which can be used as a reference for clinical measures to prevent recurrence in children after treatment of CVA.