生存分析的应用误区
|
摘要
随着调查数据的日渐丰富和统计软件易用性的不断提高,生存分析在人口学和社会学研究领域中得到广泛的应用。但对这两个研究领域的SSCI和CSSCI来源期刊中应用生存分析的文献进行考察后发现,目前在生存分析应用中存在一些误区:(1)仅考虑生存数据的右删截特征,忽略截平等特征,导致错误的似然函数、参数误估和模型拟合差等问题。右删截虽然是生存数据最常见的特征,但不是唯一的特征。判断生存数据的特征要综合考虑事件的起点时间、历险时间及观测时间;(2)忽视Cox比例风险模型的等比例假设检验;(3)把时变变量误用为非时变变量而造成估计偏差;(4)死亡风险分析中,错将观测期当作历险时间,实际上,死亡风险的历险时间应该是年龄。此外,中文文献中对生存分析的专有名词翻译不一致。导致这些问题的主要原因是使用者没有透彻理解该方法。生存分析的规范和正确应用要求使用者准确理解和把握重要的理论和方法细节。
With the increasingly rich data and powerful statistical software, survival analysis has been widely applied in the research field of demography and sociology. With an examination of the journal articles from Web of Sciences SSCI and Chinese Social Sciences Citation Index(CSSCI), however, there are some problems and mistakes in the application of survival analysis.(1) While right censoring is usually concerned, other kinds of censoring and truncation are often ignored. This could lead to incorrect likelihood function, error in parameter estimation and poor model fit. Although right censoring is common in survival analysis, it is not the only issue. It is vital to assess other key features such as the start time, time to event and observation time of one event.(2) Tests of the proportionality assumption in Coxproportional hazard models are frequently ignored.(3)Incorrect use oftime-dependent as time-invariant may lead to estimation bias.(4) Time to event is sometimes mis-specified in mortality analysis, which should be age, not the observation time. In addition, the translation of terminology is rather confusing in the Chinese literature. The main reason is that the users do not fully understand the method. A correct application requires the user to accurately understand and grasp important theoretical and methodological details.
With the increasingly rich data and powerful statistical software, survival analysis has been widely applied in the research field of demography and sociology. With an examination of the journal articles from Web of Sciences SSCI and Chinese Social Sciences Citation Index(CSSCI), however, there are some problems and mistakes in the application of survival analysis.(1) While right censoring is usually concerned, other kinds of censoring and truncation are often ignored. This could lead to incorrect likelihood function, error in parameter estimation and poor model fit. Although right censoring is common in survival analysis, it is not the only issue. It is vital to assess other key features such as the start time, time to event and observation time of one event.(2) Tests of the proportionality assumption in Coxproportional hazard models are frequently ignored.(3)Incorrect use oftime-dependent as time-invariant may lead to estimation bias.(4) Time to event is sometimes mis-specified in mortality analysis, which should be age, not the observation time. In addition, the translation of terminology is rather confusing in the Chinese literature. The main reason is that the users do not fully understand the method. A correct application requires the user to accurately understand and grasp important theoretical and methodological details.
引文
1.常宝宁(2010):《免费政策实施后儿童辍学问题实证研究---基于COX比例风险模型的分析》,《青年研究》,第6期。
2.郭志刚(2001):《历时研究与事件史分析》,《中国人口科学》,第1期。
3.李强、张震(2009):《生存分析中时间变量的选择》,《中国人口科学》,第6期。
4.梁同贵(2016):《乡城流动人口的生育间隔及其影响因素---以上海市为例》,《人口与经济》,第5期。
5.潘光辉(2017):《“撤点并校”、家庭背景与入学机会》,《社会》,第3期。
6.彭非、王伟(2004):《生存分析》,中国人民大学出版社。
7.位秀平、吴瑞君(2015):《中国老年人的躯体功能对死亡风险的影响》,《人口与经济》,第2期。
8.Allison,P.D.(1982),Discrete-time Methods for the Analysis of Event Histories.Sociological Methodology.13:61-98.
9.Austin,P.C.,Mamdani,M.M.,Van Walraven,C.,&Tu,J.V.(2006),Quantifying the Impact of Survivor Treatment Bias in Observational Studies.Journal of Evaluation in Clinical Practice.12(6):601-612.
10.Cox,D.R.(1972),Regression Models and Life Tables.Journal of Royal Statistical Society.34(2):187-202.
11.Gampe,J.(2010),Human Mortality Beyond Age 110.In Maier,H.et al.(eds)Supercentenarians.Springer,Berlin,Heidelberg.
12.Guo,G.(1993),Event-history Analysis for Left-truncated Data.Sociological Methodology.23:217-243.
13.Guo,S.(2010),Survival Analysis.Oxford University Press.
14.Klein,P.J.,&Moeschberger,M.L.(2003),Survival Analysis:Techniques for Censored and Truncated Data(second edition).Springer-Verlag New York.Inc.
15.Korn,E.L.,Graubard,B.I.,&Midthune,D.(1997),Time-to-event Analysis of Longitudinal Follow-up of a Survey:Choice of the Time-scale.American Journal of Epidemiology.145(1):72-80.
16.Long,J.S.,Allison,P.D.,&Mc Ginnis,R.(1993),Rank Advancement in Academic Careers:Sex Differences and the Effects of Productivity.American Sociological Review.58(5):703-722.
17.Lopoo,L.M.,&Western,B.(2005),Incarceration and the Formation and Stability of Marital Unions.Journal of Marriage and Family.67(3):721-734.
18.Massey,D.S.,&Espinosa,K.E.(1997),What's Driving Mexico-US Migration?A Theoretical,Empirical,and Policy Analysis.American Journal of Sociology.102(4):939-999.
19.Molitoris,J.(2017),Disparities in Death:Inequality in Cause-specific Infant and Child Mortality in Stockholm,1878-1926.Demographic Research.36:455-500.
20.Oppenheimer,V.K.(2003),Cohabiting and Marriage During Young Men’s Career-development Process.Demography.40(1):127-149.
21.Sweeney,M.M.(2002),Two Decades of Family Change:The Shifting Economic Foundations of Marriage.American Sociological Review.67(1):132-147.
22.Thiébaut,A.C.,&Bénichou,J.(2004),Choice of Time-scale in Cox’s Model Analysis of Epidemiologic Cohort Data:A Simulation Study.Statistics in Medicine.23(24):3803-3820.
23.van Walraven,C.,Davis,D.,Forster,A.J.,&Wells,G.A.(2004),Time-Dependent Bias was Common in Survival Analyses Published in Leading Clinical Journals.Journal of Clinical Epidemiology.57(7):672-682.
24.Whitbeck,L.B.,Yoder,K.A.,Hoyt,D.R.,&Conger,R.D.(1999),Early Adolescent Sexual Activity:A Developmental Study.Journal of Marriage and Family.61(4):934-946.
25.Xue,Y.,&Schifano,E.D.(2017),Diagnostics for the Cox Model.Communications for Statistical Applications and Methods.24(6):583-604.
(1)本研究没有包含生命表分析,主要关注现代统计分析中的生存分析方法。
2.郭志刚(2001):《历时研究与事件史分析》,《中国人口科学》,第1期。
3.李强、张震(2009):《生存分析中时间变量的选择》,《中国人口科学》,第6期。
4.梁同贵(2016):《乡城流动人口的生育间隔及其影响因素---以上海市为例》,《人口与经济》,第5期。
5.潘光辉(2017):《“撤点并校”、家庭背景与入学机会》,《社会》,第3期。
6.彭非、王伟(2004):《生存分析》,中国人民大学出版社。
7.位秀平、吴瑞君(2015):《中国老年人的躯体功能对死亡风险的影响》,《人口与经济》,第2期。
8.Allison,P.D.(1982),Discrete-time Methods for the Analysis of Event Histories.Sociological Methodology.13:61-98.
9.Austin,P.C.,Mamdani,M.M.,Van Walraven,C.,&Tu,J.V.(2006),Quantifying the Impact of Survivor Treatment Bias in Observational Studies.Journal of Evaluation in Clinical Practice.12(6):601-612.
10.Cox,D.R.(1972),Regression Models and Life Tables.Journal of Royal Statistical Society.34(2):187-202.
11.Gampe,J.(2010),Human Mortality Beyond Age 110.In Maier,H.et al.(eds)Supercentenarians.Springer,Berlin,Heidelberg.
12.Guo,G.(1993),Event-history Analysis for Left-truncated Data.Sociological Methodology.23:217-243.
13.Guo,S.(2010),Survival Analysis.Oxford University Press.
14.Klein,P.J.,&Moeschberger,M.L.(2003),Survival Analysis:Techniques for Censored and Truncated Data(second edition).Springer-Verlag New York.Inc.
15.Korn,E.L.,Graubard,B.I.,&Midthune,D.(1997),Time-to-event Analysis of Longitudinal Follow-up of a Survey:Choice of the Time-scale.American Journal of Epidemiology.145(1):72-80.
16.Long,J.S.,Allison,P.D.,&Mc Ginnis,R.(1993),Rank Advancement in Academic Careers:Sex Differences and the Effects of Productivity.American Sociological Review.58(5):703-722.
17.Lopoo,L.M.,&Western,B.(2005),Incarceration and the Formation and Stability of Marital Unions.Journal of Marriage and Family.67(3):721-734.
18.Massey,D.S.,&Espinosa,K.E.(1997),What's Driving Mexico-US Migration?A Theoretical,Empirical,and Policy Analysis.American Journal of Sociology.102(4):939-999.
19.Molitoris,J.(2017),Disparities in Death:Inequality in Cause-specific Infant and Child Mortality in Stockholm,1878-1926.Demographic Research.36:455-500.
20.Oppenheimer,V.K.(2003),Cohabiting and Marriage During Young Men’s Career-development Process.Demography.40(1):127-149.
21.Sweeney,M.M.(2002),Two Decades of Family Change:The Shifting Economic Foundations of Marriage.American Sociological Review.67(1):132-147.
22.Thiébaut,A.C.,&Bénichou,J.(2004),Choice of Time-scale in Cox’s Model Analysis of Epidemiologic Cohort Data:A Simulation Study.Statistics in Medicine.23(24):3803-3820.
23.van Walraven,C.,Davis,D.,Forster,A.J.,&Wells,G.A.(2004),Time-Dependent Bias was Common in Survival Analyses Published in Leading Clinical Journals.Journal of Clinical Epidemiology.57(7):672-682.
24.Whitbeck,L.B.,Yoder,K.A.,Hoyt,D.R.,&Conger,R.D.(1999),Early Adolescent Sexual Activity:A Developmental Study.Journal of Marriage and Family.61(4):934-946.
25.Xue,Y.,&Schifano,E.D.(2017),Diagnostics for the Cox Model.Communications for Statistical Applications and Methods.24(6):583-604.
(1)本研究没有包含生命表分析,主要关注现代统计分析中的生存分析方法。
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |