Integrated likelihoods in parametric survival models for highly clustered censored data
Tóm tắt
In studies that involve censored time-to-event data, stratification is frequently encountered due to different reasons, such as stratified sampling or model adjustment due to violation of model assumptions. Often, the main interest is not in the clustering variables, and the cluster-related parameters are treated as nuisance. When inference is about a parameter of interest in presence of many nuisance parameters, standard likelihood methods often perform very poorly and may lead to severe bias. This problem is particularly evident in models for clustered data with cluster-specific nuisance parameters, when the number of clusters is relatively high with respect to the within-cluster size. However, it is still unclear how the presence of censoring would affect this issue. We consider clustered failure time data with independent censoring, and propose frequentist inference based on an integrated likelihood. We then apply the proposed approach to a stratified Weibull model. Simulation studies show that appropriately defined integrated likelihoods provide very accurate inferential results in all circumstances, such as for highly clustered data or heavy censoring, even in extreme settings where standard likelihood procedures lead to strongly misleading results. We show that the proposed method performs generally as well as the frailty model, but it is superior when the frailty distribution is seriously misspecified. An application, which concerns treatments for a frequent disease in late-stage HIV-infected people, illustrates the proposed inferential method in Weibull regression models, and compares different inferential conclusions from alternative methods.
Tài liệu tham khảo
citation_title=Survival and event history analysis: a process point of view; citation_publication_date=2008; citation_id=CR1; citation_author=OO Aalen; citation_author=Ø Borgan; citation_author=HK Gjessing; citation_publisher=Springer
citation_title=Statistical models based on counting processes; citation_publication_date=1993; citation_id=CR2; citation_author=PK Andersen; citation_author=Ø Borgan; citation_author=RD Gill; citation_author=N Keiding; citation_publisher=Springer
citation_title=Inference and asymptotics; citation_publication_date=1994; citation_id=CR3; citation_author=OE Barndorff-Nielsen; citation_author=DR Cox; citation_publisher=Chapman and Hall
Bartolucci F, Bellio R, Salvan A, Sartori N (2015) Modified profile likelihood for fixed-effects panel data models. Econ Rev (to appear). doi:
10.1080/07474938.2014.975642
Bellio R, Guolo A (2015) Integrated likelihood inference in small sample meta-analysis for continuous outcomes. Scand J Stat (to appear). doi:
10.1111/sjos.12172
citation_journal_title=Biom J; citation_title=Practical use of modified maximum likelihoods for stratified data; citation_author=R Bellio, N Sartori; citation_volume=48; citation_publication_date=2006; citation_pages=876-886; citation_doi=10.1002/bimj.200510221; citation_id=CR6
citation_journal_title=Biometrics; citation_title=Hierarchical proportional hazards regression models for highly stratified data; citation_author=PB Carlin, JS Hodges; citation_volume=55; citation_publication_date=1999; citation_pages=1162-1170; citation_doi=10.1111/j.0006-341X.1999.01162.x; citation_id=CR7
Cohn DL, Fisher EJ, Peng JS, Hodges JS, Chesnut J, Child CC, Franchino B, Gibert CL, El-Sadr W, Hafner R, Korvick J, Ropka M, Heifets L, Clotfelter J, Munroe D, Horsburgh CR, and Terry Beirn Community Programs for Clinical Research on AIDS (1999) A prospective randomized trial of four three-drug regimens in the treatment of disseminated mycobacterium avium complex disease in AIDS patients: excess mortality associated with high-dose clarithromycin. Clin Infect Dis 29:125–133
citation_journal_title=J R Stat Soc Ser B Met; citation_title=Parameter orthogonality and approximate conditional inference (with discussion); citation_author=DR Cox, N Reid; citation_volume=49; citation_publication_date=1987; citation_pages=1-39; citation_id=CR9
De Bin R (2012) Integrated likelihood for the treatment of nuisance parameters. Ph.D. Thesis, Department of Statistical Sciences, University of Padua.
http://paduaresearch.cab.unipd.it/4872
citation_journal_title=Electron J Stat; citation_title=Integrated likelihoods in models with stratum nuisance parameters; citation_author=R Bin, N Sartori, TA Severini; citation_volume=9; citation_publication_date=2015; citation_pages=1474-1491; citation_doi=10.1214/15-EJS1045; citation_id=CR11
citation_title=The statistical analysis of failure times data; citation_publication_date=2002; citation_id=CR12; citation_author=JD Kalbfleish; citation_author=RL Prentice; citation_publisher=Wiley
citation_journal_title=J R Stat Soc Ser B Met; citation_title=A simple method for the adjustment of profile likelihoods; citation_author=P McCullagh, R Tibshirani; citation_volume=52; citation_publication_date=1990; citation_pages=325-344; citation_id=CR13
Munda M, Rotolo F, Legrand C (2012) Parfm: parametric frailty models in R. J Stat Softw 51(11)
citation_journal_title=Biometrika; citation_title=Effects of the reference set on frequentist inferences; citation_author=DA Pierce, R Bellio; citation_volume=93; citation_publication_date=2006; citation_pages=425-438; citation_doi=10.1093/biomet/93.2.425; citation_id=CR15
R Core Team (2013) R: A language and environment for statistical computing. R Foundation for Statistical Computing , Vienna. URL
http://www.R-project.org/
citation_journal_title=Biometrika; citation_title=Modified profile likelihoods in models with stratum nuisance parameters; citation_author=N Sartori; citation_volume=90; citation_publication_date=2003; citation_pages=533-549; citation_doi=10.1093/biomet/90.3.533; citation_id=CR17
citation_title=Likelihood methods in statistics; citation_publication_date=2000; citation_id=CR18; citation_author=TA Severini; citation_publisher=Oxford University Press
citation_journal_title=Biometrika; citation_title=Integrated likelihood functions for non-Bayesian inference; citation_author=TA Severini; citation_volume=94; citation_publication_date=2007; citation_pages=529-542; citation_doi=10.1093/biomet/asm040; citation_id=CR19
citation_journal_title=Biometrika; citation_title=Likelihood ratio statistics based on an integrated likelihood; citation_author=TA Severini; citation_volume=97; citation_publication_date=2010; citation_pages=481-496; citation_doi=10.1093/biomet/asq015; citation_id=CR20
citation_journal_title=Stat Sin; citation_title=Frequency properties on inferences based on an integrated likelihood function; citation_author=TA Severini; citation_volume=21; citation_publication_date=2011; citation_pages=433-447; citation_id=CR21