Bayesian Interim Analysis of Weibull Regression Models with Gamma Frailty
This paper considers the problem of planning prospective clinical studies where the primary endpoint is a terminal event and the response variable is a survival time. It is assumed that the lifetimes of the individuals in the study display extra-Weibull variability that causes the usual proportional hazards assumption to fail. The introduction of a Gamma-distributed frailty term to accommodate the between-subject heterogeneity leads to a logarithmic F accelerated failure time model to which the second-order expansions of Papandonatos & Geisser  can be applied. The predictive simulation approach of Papandonatos & Geisser  can then be used to evaluate the length of the study period needed for a Bayesian hypothesis testing procedure to achieve a conclusive result.
Key wordsBayesian Inference Stochastic Curtailment Weibull Frailty
Unable to display preview. Download preview PDF.
- G.D. Papandonatos And S. Geisser, Bayesian Interim Analysis of Lifetime Data, Canad. J. Statist., to appear.Google Scholar
- P. Armitage, Discussion of the paper by Jennison and Turnbull, J. R. Statist. Soc. B, 51 (1989), pp. 333–335.Google Scholar
- D.R. Cox And D. Oakes, Analysis of Survival Data, Chapman & Hall, London (1984).Google Scholar
- T. Lancaster, The econometric analysis of transition data, Econometric Society Monographs, Cambridge University Press, Cambridge (1992).Google Scholar
- D.J. Spiegelhalter And L.S. Freedman, Bayesian approaches to Clinical Trials, Bayesian Statistics 3, J.M. Bernardo, M.H. DeGroot, D.V. Lindley and A.F.M. Smith (eds.), Clarendon Press, Oxford (1988), pp. 453–477.Google Scholar