Bayesian analysis and model selection for interval-censored survival data

Interval-censored data occur in survival analysis when the survival time of each patient is only known to be within an interval and these censoring in tervals differ from patient to patient. For such data, we present some Baye sian discretized semiparametric models, incorporating proportional and nonp roportional hazards structures, along with associated statistical analyses and tools for model selection using sampling-based methods. The scope of th ese methodologies is illustrated through a reanalysis of a breast cancer da ta set (Finkelstein, 1986, Biometrics 42, 845-854) to test whether the effe ct of covariate on survival changes over time.