| 邵淑玲,谢雁鸣,黎元元,等.基于倾向评分法探讨真实世界注射用灯盏花素治疗冠心病的临床实效研究[J].浙江中医药大学学报,2022,46(3):300-308. |
| 基于倾向评分法探讨真实世界注射用灯盏花素治疗冠心病的临床实效研究 |
| Propensity Score-based Study on Effectiveness of Clinical Drugs for Treating Coronary Heart Disease with Breviscapine Injection in the Real World |
| DOI：10.16466/j.issn1005-5509.2022.03.014 |
| 中文关键词: 注射用灯盏花素 冠心病 HIS 真实世界 倾向性评分 Logistic回归分析 协变量 疗效 |
| 英文关键词: Breviscapine injection coronary heart disease HIS real world propensity score Logistic regression analysis covariates curative efficacy |
| 基金项目:国家重点研发计划(2018YFC1707400) |
|
| 摘要点击次数: 1628 |
| 全文下载次数: 613 |
| 中文摘要: |
| [目的] 了解注射用灯盏花素在真实世界中治疗冠状动脉粥样硬化性心脏病(冠心病)的实际疗效。[方法] 收集2002年1月至2015年12月全国37家三级甲等医院信息系统(hospital information system，HIS)中所有联合使用注射用灯盏花素和未联合使用注射用灯盏花素的冠心病病例，根据性别、年龄、入院病情进行1:1匹配后，灯盏花素组和对照组各3 837例，采用广义增强模型(generalized boosted models，GBM)倾向评分加权法及3种Logistic回归方法(单因素Logistic回归、不带协变量的倾向性评分加权Logistic回归、带协变量的倾向性评分加权Logistic回归)对两组数据进行分析，比较两组间的疗效差异。 [结果] 单因素Logistic回归表明两组治疗结果差异有统计学意义(r=-1.1766，P＜0.001)；使用GBM倾向评分对两组个体进行加权处理后，进行Logistic回归分析，表明两组治疗结果差异有统计学意义(r=-0.8398，P＜0.001)；将对处理分配影响程度不同的各协变量纳入到模型中，进行倾向评分加权后再次进行Logistic回归分析，提示两组治疗结果差异有统计学意义(r=-1.0833，P＜0.001)。[结论] 基于数据统计分析发现，注射用灯盏花素在真实世界中治疗冠心病真实有效。 |
| 英文摘要: |
| [Objective] To evaluate the practical effect of Breviscapine (BRE) injection on the treatment of coronary heart disease (CHD) in the real world. [Methods] Through collecting all cases from 37 Three-A hospitals information system (HIS) from January 2002 to December 2015， which diagnosed as CHD with and without using BRE injection. The patients using BRE injection were used as observation group， and control group was matched according to gender， age and hospitalization condition (without BRE injection). After matching， there were 3 837 cases in BRE injection group and control group respectively. The generalized boosted models (GBM) propensity score weighting method and three Logistic regression methods were used for data analysis. [Results] The single factor‘s Logistic regression analysis indicated the difference between the two groups was statistically significant (r=-1.1766， P＜0.001); after individuals in control group were weighted by propensity score weighting， the result of the single factor‘s Logistic regression analysis showed there was a significant difference between the two groups (r=-0.8398， P＜0.001). The third Logistic regression analysis， which the covariates with different effects on grouping random allocation were included in the model， still showed statistical significance (r=-1.0833， P＜0.001). [Conclusion] The BRE injection was effective on the treatment of CHD in the real world. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |