倾向性评分法在脑血管病临床研究中的应用 Application of Propensity Score in Clinical Research on Cerebrovascular Disease

Abstract

真实世界研究起源于实用性临床试验，在真实临床实践中评估干预的有效性和安全性，对随机对照试验以外临床疗效评价进行补充。然而，此类研究缺乏随机化，因此对混杂偏倚的控制尤为重要。倾向性评分法对多个混杂因素/协变量进行校正，是真实世界研究控制混杂因素的重要分析方法。本文介绍匹配法、分层法、协变量校正法和逆概率加权法这4种倾向性评分法，并阐述各自的优缺点。随后结合脑血管病临床研究特点，通过研究案例重点介绍倾向性评分法在脑血管病真实世界研究中的应用。 Abstract: Real-world studies, originating from pragmatic clinical trials, evaluate the effectiveness and safety of interventions in real clinical practice, as a supplement for randomized controlled trials to evaluate the effectiveness of intervention measures in real clinical environment. However, the control of confounding factors is particularly important as this type of study lacks randomization. The propensity score method, which can adjust for multiple confounders/covariates, is an important analysis method for controlling confounders in real world studies. This article introduced four methods of propensity score and discussed their advantages and disadvantages; and introduced the application of propensity score method in the real-world research on cerebrovascular disease through case studies, combined with the characteristics of clinical research on cerebrovascular disease.

Keywords

• 倾向性评分法
• 临床研究
• 脑血管病
• 真实世界研究
• propensity score
• clinical research
• cerebrovascular disease
• real-world study