Objective. In this study, a prognostic model for the respiratory function was constructed based on the treatment methods of patients with nonsevere pulmonary infection, aiming to provide a reference for clinical decision-making. Method. A total of 500 patients with nonsevere pulmonary infection were included in this study. The patients were randomized into training set () and validation set (), and the baseline characteristics were collected. All patients received breathing exercises or breathing exercises combined with acupuncture therapy for 3 months, and then the forced expiratory volume in one second (FEV1)/forced vital capacity (FVC) was assessed. Next, an ordinal multinomial logistic regression model was used to analyze prognostic factors affecting respiratory function of patients with nonsevere pulmonary infection. The Test of Parallel Lines was used to determine the accuracy (ACC) of the model and screen the influencing factors. The confusion matrix was drawn, and the ACC and harmonic mean (F1 score) were calculated to evaluate the feasibility of the model results. Results. Results of the ordinal multinomial logistic regression model showed that age (), treatment method (), underlying diseases (), and sex () were independent factors affecting the respiratory function of patients in the training set. The ACC value of the training set was 88.86%, and that of the validation set was 91.33%, indicating a high accuracy and favorable predictive ability of the model. Besides, the F1 score was 62.38%, indicating a high reliability of the model. Conclusion. The prognostic model for respiratory function of patients with nonsevere pulmonary infection constructed in this study had favorable predictive performance, which is of great significance in the clinical nursing and treatment of patients with pulmonary infection.

中文翻译：

---

基于呼吸练习和针灸治疗的非重症肺部感染患者呼吸功能预后模型：开发和验证

客观。本研究基于非重症肺部感染患者的治疗方法构建呼吸功能预后模型，旨在为临床决策提供参考。方法。本研究共纳入 500 名非严重肺部感染患者。患者被随机分配到训练集（)和验证集 ()，并收集基线特征。所有患者均接受呼吸运动或呼吸运动结合针灸治疗3个月，然后评估一秒用力呼气量（FEV1）/用力肺活量（FVC）。其次，采用有序多项逻辑回归模型分析影响非重症肺部感染患者呼吸功能的预后因素。平行线检验用于确定模型的准确度（ACC）并筛选影响因素。绘制混淆矩阵，计算ACC和调和均值（F1得分），评价模型结果的可行性。结果。序数多项逻辑回归模型的结果表明，年龄 ()、处理方法(),基础疾病 ()和性别 ()是影响训练集中患者呼吸功能的独立因素。训练集的ACC值为88.86%，验证集的ACC值为91.33%，表明该模型具有较高的准确率和良好的预测能力。此外，F1 得分为 62.38%，表明该模型具有较高的可靠性。结论。本研究构建的非重症肺部感染患者呼吸功能预后模型具有良好的预测性能，对肺部感染患者的临床护理和治疗具有重要意义。