药物经济学评价中生存曲线的拟合外推方法和对比研究  

Fitting and Extrapolation Methods of Survival Curve in Pharmacoeconomic Evaluation and the Comparative Study

作　　者：杨盟雨 代展菁 卢钰琼 杨晶 崔翔宇 常峰[1] 路云[1] YANG Meng-yu;DAI Zhan-jing;LU Yu-qiong;YANG Jing;CUI Xiang-yu;CHANG Feng;LU Yun(School of International Pharmaceutical Business,China Pharmaceutical University,Nanjing 211198,China;不详)

机构地区：[1]中国药科大学国际医药商学院,南京211198

出　　处：《中国卫生经济》2023年第3期50-58,共9页Chinese Health Economics

基　　金：中国价格协会委托项目。

摘　　要：目的:探讨生存曲线拟合外推的新方法,以应对标准参数模型无法捕捉复杂风险函数的问题,为我国抗肿瘤药物经济学评价中生存曲线的拟合外推提供方法学参考。方法:基于国内外文献总结治愈模型、混合模型、分段模型、限制性三次样条模型、界标模型5种方法的基本原理和应用案例。结果:通过对比分析,发现5种新方法都灵活运用6种参数分布、采用对生存时间或人群进行划分的方式来捕捉和拟合复杂的风险函数,为生存曲线的更优拟合提供了可能,但目前还存在模型选择和拟合结果的不确定性。结论:应深入研究以风险函数估计为基础的拟合外推理论,规范模型选择的路径方法,并验证模型选择的合理性,不断提高我国药物经济学评价中生存分析方面的证据质量。Objective:Exploring and discussing new fitting and extrapolation methods of survival curve to cope with the inability of standard parametric model to capture complex hazard functions,and to provide methodological references for relevant researches in pharmacoeconomic evaluation of antitumor drugs in China.Methods:The fundamental principles and application cases of cure model,mixture model,piecewise model,restricted cubic spline model and landmark model are summarized based on domestic and foreign literatures.Results:Through comparative analysis,it is found that all five new methods can capture and fit complex hazard functions by flexibly using six types of parametric distributions and employing segmentation of survival time or populations,which provides the possibility of better fitting of survival curves,but there is currently uncertainty in model selection and fitting results.Conclusion:It is necessary to conduct in-depth study of fitting and extrapolation theory based on hazard function estimation,standardize the path of model selection,and verify the rationality of model selection to continuously improve the quality of evidence in survival analysis of pharmacoeconomic evaluation in China.

关 键 词：药物经济学 生存曲线 拟合外推 风险函数 生存函数 

分 类 号：R1-9[医药卫生—公共卫生与预防医学] R956