文章摘要
基于中医体质构建髋部骨折患者术后谵妄的预测模型
Construction of a postoperative delirium prediction model in elderly patients with hip fracture based on traditional Chinese medicine constitution
  
DOI：10.12089/jca.2026.08.003
中文关键词: 老年  髋部骨折  术后谵妄  中医体质  预测模型  列线图
英文关键词: Aged  Hip fracture  Postoperative delirium  Traditional Chinese medicine constitution  Prediction model  Nomogram
基金项目:甘肃省联合科研基金重大项目（25JRRA1217）；兰州市科技计划项目（2024-3-36）
作者单位E-mail
郭玲玲 730000,兰州市,甘肃省中医院麻醉医学科  
冀轲婷 甘肃中医药大学第一临床医学院  
陈博 甘肃中医药大学第一临床医学院  
徐紫清 730000,兰州市,甘肃省中医院麻醉医学科  
侯怀晶 730000,兰州市,甘肃省中医院麻醉医学科  
邓寅芸 甘肃中医药大学第一临床医学院  
王东红 730000,兰州市,甘肃省中医院麻醉医学科  
薛建军 730000,兰州市,甘肃省中医院麻醉医学科 xjjfei419@163.com 
摘要点击次数: 24
全文下载次数: 0
中文摘要:
      
目的： 探讨基于中医体质的老年髋部骨折患者在行手术治疗后术后谵妄（POD）发生的影响因素，构建并验证临床预测模型，为POD早期防控提供循证依据。
方法： 结合Logistic回归预测模型样本量标准完成样本量估算，纳入2023年11月至2024年9月老年髋部骨折患者，年龄≥65岁，ASA Ⅰ—Ⅲ级。采用简单随机抽样按7∶3划分为训练集与验证集；依据西医《意识模糊评估量表（CAM）》诊断POD，参照《中医体质分类与判定》量表辨识中医体质。收集患者一般信息、手术信息及生化指标，通过LASSO回归筛选潜在预测因子，将LASSO回归筛选出的变量进行多因素Logistic回归分析。采用R Studio 4.4.2软件构建预测模型并绘制列线图。采用受试者工作特征曲线（ROC）、Hosmer-Lemeshow检验及决策曲线分析（DCA）分别评价模型的区分度、校准度及临床价值。同步对比未纳入中医体质变量时模型的预测效能，明确中医体质对模型的影响。
结果： 共纳入老年患者515例，训练集360例，验证集155例。LASSO回归筛选出20个潜在预测因子。多因素Logistic回归显示，股骨颈骨折、气虚质、痰湿质、阳虚质、阴虚质、糖尿病、术后入ICU、术中低血压、年龄偏大、麻醉时间过长是老年髋部骨折患者发生POD的独立危险因素（P<0.05）。纳入中医体质可提升模型区分能力，不同偏颇体质亚组POD发生风险存在明显差异。基于上述因素构建列线图模型，训练集ROC曲线下面积（AUC）为0.778（95%CI 0.730～0.826），验证集AUC为0.754（95%CI 0.672～0.837）；Hosmer-Lemeshow检验提示模型校准度良好，DCA显示模型在一定风险阈值内具有较高临床应用价值。
结论： 基于中医体质构建的髋部骨折患者术后谵妄的预测模型具有较好的区分度、校准度和临床有效性，具有一定的理论意义及现实意义，可为医务人员快速评估患者POD风险提供便捷工具。
英文摘要:
      
Objective: To explore the influencing factors of postoperative delirium (POD) in elderly patients with hip fracture after surgical treatment based on traditional Chinese medicine (TCM) constitution, construct and validate a clinical prediction model, so as to provide evidence-based references for the early prevention and intervention of POD.
Methods: The sample size was estimated in accordance with the sample size criteria for Logistic regression prediction models. Elderly patients aged ≥ 65 years, ASA physical status Ⅰ-Ⅲ, with hip fracture from November 2023 to September 2024 were enrolled. All patients were divided into a training set and a validation set at a ratio of 7∶3 by simple random sampling. POD was diagnosed using the confusion assessment method (CAM), and TCM constitution was identified by referring to the classification and determination of traditional Chinese medicine constitution scale. General information, surgical data, and biochemical indicators of patients were collected. LASSO regression was applied to screen potential predictive factors, and the selected variables were further analyzed by multivariate logistic regression. R Studio 4.4.2 software was used to establish the prediction model and draw a nomogram. The receiver operating characteristic (ROC) curve, Hosmer-Lemeshow test and decision curve analysis (DCA) were adopted to evaluate the discrimination, calibration, and clinical value of the model, respectively. Meanwhile, the predictive efficacy of the model without TCM constitution variables was compared to clarify the impact of TCM constitution on the model.
Results: A total of 515 elderly patients were included, 360 patients in the training set and 155 patients in the validation set. Twenty potential predictive factors were screened out by LASSO regression. Multivariate logistic regression revealed that femoral neck fracture, qi deficiency constitution, phlegm-dampness constitution, yang deficiency constitution, yin deficiency constitution, diabetes mellitus, postoperative admission to intensive care unit (ICU), intraoperative hypotension, advanced age and prolonged anesthesia duration were independent risk factors for POD in elderly patients with hip fracture (P < 0.05). Incorporation of TCM constitution improved the discriminative ability of the model, and the risk of POD varied significantly among subgroups with different biased TCM constitutions. A nomogram model was established based on the above factors, the area under the ROC curve (AUC) was 0.778 (95% CI 0.730-0.826) for the training set and 0.754 (95% CI 0.672-0.837) for the validation set. The Hosmer-Lemeshow test indicated favorable model calibration, and DCA verified that the postoperative delirium prediction model in elderly patients with hip fracture based on traditional Chinese medicine constitution possessed high clinical application value within a certain risk threshold.
Conclusion: The postoperative delirium prediction model in elderly patients with hip fracture based on traditional Chinese medicine constitution has good discrimination, calibration, and clinical effectiveness, with certain theoretical and practical significance. It can provide a convenient tool for medical staff to quickly assess the risk of POD in patients.
查看全文   查看/发表评论  下载PDF阅读器
关闭