| 李丹丹.儿童难治性肺炎支原体肺炎的危险因素分析及列线图模型构建[J].安徽医药,待发表. |
| 儿童难治性肺炎支原体肺炎的危险因素分析及列线图模型构建 |
|
| 投稿时间：2024-05-08 录用日期：2024-06-15 |
| DOI： |
| 中文关键词: 支原体肺炎 难治性肺炎 列线图 预测模型 |
| 英文关键词: |
| 基金项目:河北省医学科学研究课题（20211652） |
|
| 摘要点击次数: 394 |
| 全文下载次数: 0 |
| 中文摘要: |
| 目的 探索儿童难治性肺炎支原体肺炎（RMPP）的危险因素，构建RMPP列线图风险预测模型。方法 回顾性分析2021年1月至2024年1月在河北省儿童医院呼吸科收治的262例肺炎支原体肺炎（MPP）病例资料。将数据集按照7:3随机分为训练集（182例）和验证集（80例）。对训练集病例资料进行单因素、相关性、ROC分析和多因素logistic回归分析，筛选出RMPP的独立影响因素，采用R软件制作列线图，用验证集中样本进行验证，模型区分度评估采用一致性指数（C?index），校准度评估采用校准图。结果 262例MPP患儿中RMPP患儿65例（24.8%），单因素分析、相关性分析、ROC分析及多因素logistic回归分析结果显示中性粒细胞计数 (OR=3.026，95%CI:1.428～6.410)、红细胞沉降率 (OR=3.530，95%CI: 1.688～7.383)、 乳酸脱氢酶水平(OR=4.293，95%CI: 2.052～8.982)、C反应蛋白水平(OR=5.468，95%CI: 2.277～13.131)和降钙素原水平(OR=4.683，95%CI: 2.264～9.685)是MPP患儿中RMPP发生的最佳预测因子。预测模型的C-index为0.85(95%CI: 0.801～0.903)，特异度和灵敏度分别为65.4%和87.6%， 验证集的C-index为0.86，。校准图显示在训练集和验证集的校准曲线与理想曲线重合度较高。结论 本研究构建的难治性肺炎支原体肺炎的列线图风险预测模型区分度与校准度较好，可为临床提供参考价值。 |
| 英文摘要: |
| Objective To explore the risk factors of refractory Mycoplasma pneumoniae pneumonia (RMPP) in children and construct a risk prediction model for RMPP nomogram. Methods The data of 262 cases of Mycoplasma pneumoniae pneumonia (MPP) admitted to the Respiratory Department of Hebei Children's Hospital from January 2021 to January 2024 were retrospectively analyzed. The dataset was randomly divided into training set (182 cases) and validation set (80 cases) according to 7:3. Single factor, correlation, ROC analysis and multifactorial logistic regression analysis were performed on the data of cases in the training set to screen out the independent influencing factors of RMPP, and R software was used to produce nomograms, which were validated with the samples in the validation set, and the model differentiation was assessed using the consistency index (C?index), and calibration was assessed using the calibration graph. Results Among the 262 children with MPP, 65 (24.8%) had RMPP, and the results of univariate analysis, correlation analysis, ROC analysis, and multifactorial logistic regression analysis showed that the neutrophil count (OR=3.026, 95% CI: 1.428-6.410), erythrocyte sedimentation rate (OR=3.530, 95% CI: 1.688 to 7.383), lactate dehydrogenase level (OR=4.293, 95% CI: 2.052 to 8.982), C-reactive protein level (OR=5.468, 95% CI: 2.277 to 13.131) and calcitoninogen level (OR=4.683, 95% CI: 2.264 to 9.685) were the best Predictors. The C-index of the prediction model was 0.85 (95% CI: 0.801-0.903), with a specificity and sensitivity of 65.4% and 87.6%, respectively, and the C-index of the validation set was 0.86,. The calibration plots showed that the calibration curves in the training and validation sets overlapped well with the ideal curves. Conclusion The nomogram risk prediction model for refractory Mycoplasma pneumoniae constructed in this study has good differentiation and calibration, which can provide clinical reference value. |
|
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
