An Interval-Censored Proportional Hazards Model

We fit a Cox proportional hazards (PH) model to interval-censored survival data by first subdividing each individual's failure interval into nonoverlapping sub-intervals. Using the set of all endpoints in set, those that fall are then used as cut points for Each sub-interval has an accompanying weight calculated from parametric Weibull based on current parameter estimates. A weighted PH is with multiple lines observations corresponding sub-intervals individual, where lower end observed time weights incorporated. Right-censored handled usual manner. iterate between estimating baseline distribution and fitting until regression parameters interest converge. The estimates fixed offset when we update recalculate weights. Our approach similar Satten et al.'s (1998) method analysis imputed times generated model. Simulation results demonstrate apparently unbiased estimation correctly specified little no bias mis-specified log-logistic Breast cosmetic deterioration ICU hyperlactemia analyzed.