Clustered survival data arise when groups of failure times share a common ingredient; typically, they refer to the same individual or individuals with a common factor. When the association between failure times within the same cluster is of interest, statistical methods called frailty models have been used. The frailty is an unobserved random component which affects the risk level, changing from cluster to cluster but shared by all observations within the same cluster. Various probability distributions have been proposed for the frailty term, with special emphasis on the gamma and log-normal distribution. Since adequate modelling of the frailty distribution is essential to properly investigate the dependence structure, we introduce a new class of frailty models with a flexible distribution form. Specifically, we adopt the skew-normal distribution for the log-transformed frailty, leading to an extension of the log-normal model. After presenting the methodology connected to this choice, we illustrate it with a case study of multiple myeloma patients with autologous stem cells transplantation.
AalenOO (1992) Modelling heterogeneity in survival analysis by the compound Poisson distribution. Annals of Applied Probability, 4, 951–72.
2.
AbrahantesJCBurzykowskiT (2005) A version of the EM algorithm for proportional hazard model with random effects. Biometrical Journal, 47, 847–62.
3.
AzzaliniA (1985) A class of distributions which includes the normal ones. Scandinavian Journal of Statistics, 12, 171–8.
4.
AzzaliniACapitanioA (1999) Statistical applications of the multivariate skew normal distribution. Journal of the Royal Statistical Society, Series B, 61, 579–602. Full version of the paper at arXiv.org:0911.2093.
5.
AzzaliniADalla ValleA (1996) The multivariate skew-normal distribution. Biometrika, 83, 715–26.
ClaytonDG (1978) A model for association in bivariate life tables and its application in epidemiologic studies of familial tendency in chronic disease incidence. Biometrika, 65, 141–51.
8.
CoxDR (1972) Regression models and life tables. Journal of the Royal Statistical Society, Series B, 34, 187–220.
9.
DempsterALairdNRubinD (1977) Maximum likelihood from incomplete data via the EM algorithm. Journal of the Royal Statistical Society, Series B, 39, 1–38.
10.
DuchateauLJanssenP (2008) The frailty model. New York: Springer.
11.
FirthD (1993) Bias reduction of maximum likelihood estimates. Biometrika, 80, 27–38.
12.
HougaardP (1986a) A class of multivariate failure time distributions. Biometrika, 73, 671–78.
13.
HougaardP (1986b) Survival models for heterogeneous populations derived from stable distributions. Biometrika, 73, 387–96.
14.
HougaardP (1995) Frailty models for survival data. Lifetime Data Analysis, 1, 255–73.
15.
HougaardP (2000) Analysis of multivariate survival data. New York: Springer.
16.
KimSDeyDK (2008) Modeling multilevel survival data using frailty models. Communications in Statistics – Theory and Methods, 37, 1734–41.
17.
KleinJ (1992) Semiparametric estimation of random effects using the Cox model based on the EM algorithm. Biometrics, 48, 795–806.
18.
KomárekALesaffreE (2008) Bayesian accelerated failure time model with multivariate doubly-interval-censored data and flexible distributional assumptions. Journal of the American Statistical Association, 103, 523–33.
19.
KomárekALesaffreE (2009) The regression analysis of correlated interval-censored data: Illustration using accelerated failure time models with flexible distributional assumptions. Statistical Modelling, 9, 299–319.
20.
KomárekALesaffreEHärkänenTDeclerckDVirtanenJI (2005) A Bayesian analysis of multivariate doubly-interval-censored dental data. Biostatistics, 6, 145–55.
21.
LamKFKukAYC (1997) A marginal likelihood approach to estimation in frailty models. Journal of the American Statistical Association, 92, 985–90.
22.
LouisTA (1982) Finding the observed information matrix when using the EM algorithm. Journal of the Royal Statistical Society, Series B, 44, 226–33.
23.
McGilchristCAAisbettCW (1991) Regression with frailty in survival analysis. Biometrics, 47, 461–66.
24.
OakesD (1994) Use of frailty models for multivariate survival data. Proceedings of the XVII th International Biometrics Conference, Hamilton, Ontario, Canada, pp. 275–86.
25.
RipattiSPalmgrenJ (2000) Estimation of multivariate frailty models using penalized partial likelihood. Biometrics, 56, 1016–22.
26.
SahuSKDeyDK (2004) On a Bayesian multivariate survival models with a skewed frailty. In GentonMG (ed.), Skew-elliptical distributions and their applications: a journey beyond normality, 321–38. Boca Raton, FL: CRC/Chapman & Hall.
27.
SartoriN (2006) Bias prevention of maximum likelihood estimates: skew normal and skew-t distributions. Journal of Statistical Planning and Inference, 136, 4259–75.
28.
Therneau T and original R port by Thomas Lumley (2009) Survival: Survival analysis, including penalised likelihood. R package version 2.35–8.
29.
TherneauTMGrambschPMPankratzVS (2003) Penalized survival models and frailty. Journal of Computational and Graphical Statistics, 12, 156–75.
30.
VaidaFXuR (2000) Proportional hazards models with random effects. Statistics in Medicine, 19, 3309–24.
31.
VaupelJWMantonKGStallardE (1979) The impact of heterogeneity in individual frailty on the dynamics of mortality. Demography, 16, 439–54.
32.
WienkeALichtensteinPYashinAI (2003) A bivariate frailty model with a cure fraction for modeling familial correlations in diseases. Biometrics, 59, 1178–83.
33.
XueXBrookmeyerR (1996) Bivariate frailty model for the analysis of multivariate survival time. Lifetime Data Analysis, 2, 277–89.
34.
YashinAIVaupelJWIachineIA (1995) Correlated individual frailty: an advantageous approach to survival analysis of bivariate data. Mathematical Population Studies, 5, 145–59.
35.
YauKKWMcGilchristCA (1997) Correlated individual frailty: use of generalised linear mixed models for the analysis of clustered survival data. Biometrical Journal, 39, 3–11.
