基于倾向性评分简单加权估计的慢性乙型肝炎抗病毒治疗的成本-效果分析  

Cost-effectiveness analysis of antiviral therapy for chronic hepatitis B based on simple weighted estimation of propensity score

作　　者：陈舸 杨朔 周峰 李丽霞 周舒冬 贾卫东[2] 郜艳晖 CHEN Ge;YANG Shuo;ZHOU Feng;LI Lixia;ZHOU Shudong;JIAWeidong;GAO Yanhui(School of Public Health,Guangdong Pharmaceutical University,Guangzhou 510310,China;Guangzhou Eighth People's Hospital,Guangzhou 510080,China;School of Medicine,Jinan University,Guangzhou 510270,China)

机构地区：[1]广东药科大学公共卫生学院,广东广州510310 [2]广州市第八人民医院信息科,广东广州510080 [3]暨南大学基础医学与公共卫生学院,广东广州510270

出　　处：《广东药科大学学报》2022年第1期29-36,共8页Journal of Guangdong Pharmaceutical University

基　　金：国家自然科学基金资助项目(71573059)。

摘　　要：目的介绍倾向性评分简单加权估计方法的原理,并将该方法运用于“真实世界”纵向队列数据的药物成本-效果评价。方法以慢性乙型肝炎(chronic hepatitis B,CHB)抗病毒初治患者的电子病例数据为例,采用简单加权方法,对存在删失信息时的各组患者疗效和医疗费用进行估计,并运用倾向性评分的逆概率加权方法对观测数据中的潜在混杂进行调整,进一步评估各抗病毒药物的成本-效果。结果本研究的方法能使混杂因素在各个治疗组间的分布达到平衡,并对各组的费用和疗效进行合理的估计。研究结果发现干扰素治疗方案在短期治疗(3年和5年)上不具有成本-效果,但在长期治疗(8年)上是成本-效果最优的治疗方案。在核苷(酸)类似物治疗方案中,LAM+ADV初始联合治疗方案在3年、5年和8年的治疗中最具成本-效果。结论倾向性评分简单加权在基于假设的条件下,能够有效处理组间混杂,并对删失疗效和医疗费用进行无偏估计,为“真实世界”复杂医疗数据的长期疗效比较和成本-效果的实证分析提供了方法选择。Objective To introduce the principle of simple weighted estimation method of propensity score,and apply this method to drug cost-effectiveness evaluation of"real world"longitudinal cohort data.Methods Taking the electronic case data of newly treated patients with chronic hepatitis B(CHB)virus as an example,the simple weighted method was used to estimate the efficacy and medical expenses of each group of patients with deleted information,and the inverse probability weighting method of propensity score was used to adjust the potential confounding in the observation data,so as to further evaluate the cost-effectiveness of each antiviral drug.Results The method of this study balanced the distribution of confounding factors among the treatment groups,and reasonably estimated the cost and curative effect of each group.The results showed that interferon treatment regimen had no cost-effectiveness in short-term treatment(3 years and 5 years),but it was the most cost-effective treatment regimen in long-term treatment(8 years).Among the nucleoside(acid)analogues,LAM+ADV initial combined treatment regimen had the most cost-effectiveness in 3 years,5 years and 8 years.Conclusion Simple weighting of propensity score can effectively deal with intergroup confounding under the condition of hypothesis,and unbiased estimate of the deleted efficacy and medical cost,which provides a method choice for the long-term efficacy comparison and cost-effectiveness empirical analysis of"real-world"complex medical data.

关 键 词：倾向性评分 简单加权 成本-效果分析 慢性乙型肝炎 

分 类 号：R956[医药卫生—药学]