Construction of risk prediction model for AECOPD patients with pulmonary infection based on decision tree method
10.3760/cma.j.cn115682-20250217-00711
- VernacularTitle:基于决策树法构建AECOPD患者并发肺部感染的风险预测模型
- Author:
Bing LI
1
;
Cailin LIU
;
Jie ZHANG
Author Information
1. 郑州大学第一附属医院呼吸与危重症医学科，郑州 450002
- Publication Type:Journal Article
- Keywords:
Chronic obstructive pulmonary diseases;
Acute exacerbation;
Decision trees;
Pulmonary infection;
Prediction model;
Nursing strategy
- From:
Chinese Journal of Modern Nursing
2025;31(31):4262-4268
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To construct a risk prediction model for pulmonary infection in patients with acute exacerbation of chronic obstructive pulmonary disease (AECOPD) based on decision tree method, and analyze nursing strategies.Methods:From March 2022 to June 2024, convenience sampling was used to select 162 patients with AECOPD at the First Affiliated Hospital of Zhengzhou University as study subjects. Patients with AECOPD complicated by pulmonary infection were designated as infection group ( n=82), while patients without pulmonary infection were designated as non-infection group ( n=80). Relevant clinical data were collected, and multiple Logistic regression was used to analyze the influencing factors of pulmonary infections in AECOPD patients, and a decision tree prediction model was established. Results:Multivariate Logistic regression analysis showed that age, glucocorticoid use time, hospital stay, mechanical ventilation history, diabetes, C-reactive protein and procalcitonin were the influencing factors of AECOPD patients with pulmonary infection ( P<0.05). The risk decision tree model of AECOPD patients with pulmonary infection based on the above factors showed that procalcitonin, C-reactive protein, mechanical ventilation history and diabetes were nodes, with four layers, six terminal nodes and four explanatory variables. The area under the receiver operating characteristic curve was 0.929 [95% CI (0.889, 0.968) ], the sensitivity was 0.854, the specificity was 0.875, and the predictive performance was good. Conclusions:Age, glucocorticoid use time, hospital stay, mechanical ventilation history, diabetes, C-reactive protein and procalcitonin are independent risk factors for AECOPD patients with pulmonary infection. The decision tree prediction model constructed has good predictive performance and can be used as a basis for selecting nursing strategies.
