Model establishment and validation of individualized prediction of the number of bronchoalveolar lavage for Mycoplasma pneumoniae pneumonia in children
Objective:To develop a Nomogram model to predict the number of bronchoalveolar lavage(BAL)per-formed for Mycoplasma pneumoniae pneumoniae(MPP)in children and to analyze the clinical value of the model.Methods:Clinical data of 240 children with MPP who attended our hospital for BAL treatment from January 2023 to January 2024 were collected retrospectively.All the children were divided into training and validation sets(ratio 7∶3)according to the random number table method.The children in training set were divided into single and multiple groups according to the number of times they underwent BAL treatment.Independent risk factors were screened based on univariate and multivariate Logistic regression analyses,and Nomogram model was constructed on this ba-sis.The accuracy of the model was evaluated using calibration curve and receiver operating characteristic(ROC)curve,and the clinical applicability and safety were assessed using clinical decision curve analysis.Results:Uni-variate analysis showed a statistically significant difference when comparing the pre-BAL thermal course,combined pleural effusion,extrapulmonary complications,lactate dehydrogenase(LDH)and C-reactive protein(CRP)be-tween the single and multiple groups(P＜0.05).Multivariate analysis showed that pre-BAL thermal course ≥ 10 d,combined pleural effusion,presence of extrapulmonary complications,LDH ≥462 U·L-1 and CRP ≥31.4 mg·L-1 were independent risk factors for multiple BALs in children with MPP(P＜0.05).The ROC curve analysis of the Nomogram model constructed based on the above risk factors showed that the area under the curve(AUC)in the training set and validation set were 0.820 and 0.843(P＜0.05),respectively.Calibration curve analysis showed that the model fitted well in the training and validation sets with good agreement.Clinical decision curve analysis showed that the model had a wide range of clinical use and high safety in predicting the number of BALs performed in children with MPP.Conclusion:Multiple BALs in children with MPP were associated with pre-BAL thermal course,combined pleural effusion,extrapulmonary complications,LDH and CRP levels.The Nomogram prediction model constructed based on the above factors is highly accurate and has a wide range of clinical applicability,provi-ding theoretical support for clinicians to identify the risk of multiple BALs in children with MPP in advance.
万方数据
维普
