Correlated survival times can be modelled by introducing a random effect, or frailty component, into the hazard function. For multivariate survival data, we extend a non-proportional hazards (PH) model, the generalized time-dependent logistic survival model, to include random effects. The hierarchical likelihood procedure, which obviates the need for marginalization over the random effect distribution, is derived for this extended model and its properties are discussed. The extended model leads to a robust estimation result for the regression parameters against the misspecification of the form of the basic hazard function or frailty distribution compared to PH-based alternatives. The proposed method is illustrated by two practical examples and a simulation study which demonstrate the advantages of the new model.
Agresti A , Caffo B and Ohman-Strickland P (2004) Examples in which misspecification of a random effect distribution reduces efficiency, and possible remedies . Computational Statistics and Data Analysis, 47, 639–53 .
2.
Blagojevic M and MacKenzie G (2007) A comparison of PH and non-PH frailty models. In Gregori D , MacKenzie G , Friedl H and Corradetti R (eds). Correlated data modelling 2004. Milano, Italy: Franco Angelli , 239–49.
3.
Cox DR (1970) The analysis of binary data. London: Methuen .
4.
Cox DR (1972) Regression models and life-tables (with discussion) . Journal of the Royal Statistical Society, Series B, 34, 187–220 .
5.
Ferreira A and Garcia NL (2001) Simulation study for misspecifications on a frailty model . Brazilian Journal of Probability and Statistics, 15, 121–34 .
6.
Fleming TR and Harrington DP (1991) Counting processes and survival analysis. New York: Wiley .
7.
Gillon N and MacKenzie G (2004) The North Staffordshire Breast Cancer Study. Internal Technical Report, Series No. 2, 1–80. Centre for Medical Statistics, Keele University, England .
8.
Gray RJ (1994) A Bayesian analysis of institutional effects in multicenter cancer clinical trial . Biometrics, 50, 244–53 .
9.
Ha ID and Lee Y (2003) Estimating frailty models via Poisson hierarchical generalized linear models . Journal of Computational and Graphical Statistics, 12, 663–81 .
10.
Ha ID and Lee Y (2005) Comparison of hierarchical likelihood versus orthodox best linear unbiased predictor approaches for frailty models . Biometrika, 92, 717–23 .
11.
Ha ID , Lee Y and MacKenzie G (2007) Model selection for multi-component frailty models . Statistics in Medicine, 26, 4790–807 .
12.
Ha ID , Lee Y and Song J-K (2001) Hierarchical likelihood approach for frailty models . Biometrika, 88, 233–43 .
13.
Ha ID , Lee Y and Song J-K (2002) Hierarchical likelihood approach for mixed linear models with censored data . Lifetime Data Analysis, 8, 163–76 .
14.
Hougaard P (1991) Modelling heterogeneity in survival data . Journal of Applied Probability, 28, 695–701 .
15.
Hougaard P (2000) Analysis of multivariate survival data. New York: Springer-Verlag .
16.
Hsu L , Gorfine M and Malone K (2007) On robustness of marginal regression coefficient estimates and hazard functions in multivariate survival analysis of family data when the frailty distribution is mis-specified . Statistics in Medicine, 26, 4657–78 .
17.
Kuk AYC and Chen C-H (1992) A mixture model combining logistic regression with proportional hazards regression . Biometrika, 79, 531–41 .
18.
Laird N (1978) Non-parametric maximum likelihood estimation of mixing distribution . Journal of the American Statistical Association, 73, 805–11 .
19.
Lee Y and Nelder JA (1996) Hierarchical generalized linear models (with discussion) . Journal of the Royal Statistical Society, Series B, 58, 619–78 .
20.
Lee Y and Nelder JA (2001) Hierarchical generalised linear models: a synthesis of generalised linear models, random-effect models and structured dispersions . Biometrika, 88, 987–1006 .
21.
Lee Y and Nelder JA (2006) Double hierarchical generalised linear models (with discussion) . Journal of the Royal Statistical Society, Series C, 55, 139–85 .
22.
Lee Y , Nelder JA and Pawitan Y (2006) Generalised linear models with random effects: unified analysis via h-likelihood. London: Chapman & Hall .
23.
Louzada-Neto F , Cremasco CP and MacKenzie G (2010) Bayesian modelling for the GTDL survival model . JSTA, 9 (June, in press).
24.
Lynch J and MacKenzie G (2007) The analysis of breast cancer survival in local health authorities . Proceedings of the 22nd International Workshop on Statistical Modelling (IWSM). Barcelona, Spain, July 2007, 408–11 .
25.
Lynch J and MacKenzie G (2008) Non-PH survival modelling with frailty and structured dispersion . Contributed Oral Presentation, 24th International Biometric Conference, Dublin, Ireland July. Abstract may be found at http://www.cpregistrations.com/ibc/2008/outbox/IBSCASI%5B1%5D20080204173357.pdf and presentation at www.ul.ie/biostatistics.
26.
MacKenzie G (1996) Regression models for survival data: the generalised time dependent logistic family . Journal of the Royal Statistical Socity, Series D, 45, 21–34 .
27.
MacKenzie G (1997) On a non-proportional hazards regression model for repeated medical random counts . Statistics in Medicine, 16, 1831–43 .
28.
MacKenzie G (2002) A logistic regression model for survival data . Proceedings of the 17th International Workshop on Statistical Modelling (IWSM). Chania, Crete, July 2002, 431–38 .
29.
Mantel N , Bohidar NR and Ciminera JL (1977) Mantel-Haenszel analysis of litter-matched time-to-response data, with modifications for recovery of interlitter information . Cancer Research, 37, 3863–68 .
30.
McGilchrist CA (1993) REML estimation for survival models with frailty . Biometrics, 49, 221–25 .
31.
McGilchrist CA and Aisbett CW (1991) Regression with frailty in survival analysis . Biometrics, 47, 461–66 .
32.
Ng ETM and Cook RJ (2000) A comparison of some random effect models for parameter estimation in recurrent events . Mathematical and Computer Modelling, 32, 11–26 .
33.
Noh M , Lee Y and Pawitan Y (2005) Robust ascertainment-adjusted parameter estimation . Genetic Epidemiology, 29, 68–75 .
34.
Pickles A and Crouchley R (1995). A comparison of frailty models for multivariate survival data . Statistics in Medicine, 14, 1447–61 .
35.
Ripatti S and Palmgren J (2000) Estimation of multivariate frailty models using penalized partial likelihood . Biometrics, 56, 1016–22 .
36.
Royston P and Parmar MKB (2002) Flexible parametric proportional-hazards and proportional-odds models for censored data, with application to prognostic modelling and estimation of treatment effects . Statistics in Medicine, 21, 2175–97 .
37.
Vaida F and Xu R (2000) Proportional hazards models with random effects . Statistics in Medicine, 19, 3309–24 .
38.
Yau KKW (2001) Multilevel models for survival analysis with random effects . Biometrics, 57, 96–102 .
39.
Yau KKW and McGilchrist CA (1998) ML and REML estimation in survival analysis with time dependent correlated frailty . Statistics in Medicine, 17, 1201–13 .
