基于倾向性评分匹配和亚群治疗效果模式图队列研究暴露与结局关联的异质性分析
Heterogeneous analysis on association between exposure and outcome in cohort studies:based on propensity score matching and subpopulation treatment effect pattern plot
王威 1于茜 2胡馨茹 3王晓骁4
作者信息
- 1. 南京中医药大学附属医院(江苏省中医院)GCP中心,南京 210029
- 2. 南京中医药大学附属医院(江苏省中医院)科技处,南京 210029
- 3. 南京中医药大学医学院·整合医学学院公共卫生学系,南京 210046
- 4. 南京中医药大学附属医院(江苏省中医院)临床研究院,南京 210029
- 折叠
摘要
介绍亚群治疗效果模式图(STEPP)的基本原理、应用场景与主要步骤,梳理将其与倾向性评分匹配(PSM)联合应用在队列研究数据中的理论基础,旨在将STEPP应用到暴露与结局风险关联的异质性分析中.通过模拟案例"不同输液导管类型的肿瘤患者发生静脉血栓栓塞(VTE)风险的队列研究"展示如何将PSM与STEPP在队列研究数据中联合应用,得出亚群中经外周置入中心静脉导管组和中心静脉导管组发生VTE的绝对风险值和相对风险值,以及亚群间的异质性分析结果,并对PSM与STEPP联合应用的局限性和研究前景进行讨论,为实际应用提供参考.
Abstract
This paper briefly introduces the basic principle,application scenarios and main steps of subpopulation treatment effect pattern plot(STEPP),and summarizes the theoretical basis of its combined application with propensity score matching(PSM)in cohort study data for the heterogeneity analysis on the association between exposure and outcome risk.In this paper,the combined application of PSM and STEPP in cohort study data is summarized by simulation of"cohort study of the risk of venous thromboembolism(VTE)in cancer patients with different infusion catheter types".The absolute risk value and relative risk value of VTE in the peripherally inserted central catheter group and the central venous catheter group in subpopulations are introduced,as well as the results of heterogeneity analysis among subpopulations.The limitations and research prospects of the combined application of PSM and STEPP are discussed to provide reference for researchers in practical application.
关键词
倾向性评分匹配/亚群治疗效果模式图/队列研究/暴露/关联Key words
Propensity score matching/Subpopulation treatment effect pattern plot/Cohort study/Exposure/Association引用本文复制引用
基金项目
江苏省卫生健康委医学科研项目(M2021001)
南京中医药大学自然科学研究项目(XZR2021005)
科主任学术提升专项(2022)(Y2022ZR14)
预防医学研究课题(2023)(Ym2023103)
出版年
2024
NETL
NSTL
万方数据