| 本文已被：浏览 187次 下载 79次 |
码上扫一扫！ |
|
|
| 天津地区社区人群脑卒中早期筛查模型的构建与评价 |
|
杨继1,2, 张垚1,2, 张秋月3, 赵英强4
|
|
1.天津中医药大学第一附属医院南院急症部, 天津 300381;2.国家中医针灸临床医学研究中心, 天津 300381;3.天津中医药大学第二附属医院老年病科, 天津 300250;4.天津中医药大学第二附属医院心血管科, 天津 300250
|
|
| 摘要: |
| [目的] 探讨天津地区脑卒中患病影响因素，构建社区人群脑卒中中西医结合早期筛查模型。[方法] 利用社区居民年度体检数据，采用倾向性评分匹配法（PSM）按年龄和性别进行匹配，建立规范数据库，并按8∶2比例随机分为训练集和测试集。通过单因素及多因素Logistic回归分析筛选预测因子并建立模型。内部验证采用Bootstrap法评估模型的拟合度、灵敏度和特异度，外部验证评估模型的区分度和校准度，并与单纯西医指标模型进行比较。[结果] 从24 792例档案资料完整的社区人群中匹配到脑卒中患者和非脑卒中患者各1 376例，其中训练集2 202例，测试集550例。模型AUC为0.891，灵敏度和特异度分别为70.57%和94.10%。校准曲线及临床决策分析显示模型预测效能良好，加入中医体质变量后，模型预测效能显著提高。[结论] 基于社区体检数据的脑卒中中西医结合早期筛查模型具有良好的临床预测效能，可有效评估社区居民脑卒中患病风险，对早期诊断具有重要价值。 |
| 关键词: 脑卒中 中医体质 中西医结合 筛查模型 |
| DOI：10.11656/j.issn.1673-9043.2025.11.01 |
| 分类号:R255.2;R743.3 |
| 基金项目:国家重点研发计划项目（2019YFC1710005）。 |
|
| Construction and evaluation of an early screening model for stroke in community populations in Tianjin area |
|
YANG Ji1,2, ZHANG Yao1,2, ZHANG Qiuyue3, ZHAO Yingqiang4
|
|
1.Department of Emergency, South Campus, First Teaching Hospital of Tianjin University of Traditional Chinese Medicine, Tianjin 300381, China;2.National Clinical Research Center for Acupuncture and Moxibustion, Tianjin 300381, China;3.Department of Geriatrics, Second Affiliated Hospital of Tianjin University of Traditional Chinese Medicine, Tianjin 300250, China;4.Department of Cardiovascular Medicine, Second Affiliated Hospital of Tianjin University of Traditional Chinese Medicine, Tianjin 300250, China
|
| Abstract: |
| [Objective] To investigate the influencing factors of stroke prevalence in Tianjin area and construct an integrated Chinese and Western medicine early screening model for stroke in community population.[Methods] Using annual physical examination data of community residents,propensity score matching(PSM) was employed to match subjects by age and gender,establishing a standardized database that was randomly divided into training and test sets at an 8:2 ratio. Predictive factors were screened through univariate and multivariate logistic regression analyses to build the model. Internal validation used the Bootstrap method to evaluate model fit,sensitivity,and specificity,while external validation assessed the model's discrimination and calibration,with comparison to a Western medicine indicators-only model.[Results] From 24 792 community residents with complete archival data,1 376 stroke patients and 1 376 non-stroke patients were matched,including 2 202 cases in the training set and 550 cases in the test set. The model achieved an AUC of 0.891,with sensitivity and specificity of 70.57% and 94.10% respectively. Calibration curve and clinical decision analysis showed good predictive performance of the model,which was significantly improved after incorporating Traditional Chinese Medicine constitution variables.[Conclusion] The integrated Chinese and Western medicine early screening model for stroke based on community physical examination data demonstrates good clinical predictive performance,can effectively assess stroke risk in community residents,and has important value for early diagnosis. |
| Key words: cerebral stroke constitution of traditional Chinese medicine integration of traditional Chinese and Western medicine screening model |
