INTERVAL CENSORED SURVIVAL-DATA - A GENERALIZED LINEAR MODELING APPROACH

A method is described for weak parametric modelling of arbitrarily int erval censored survival data using generalized linear models. The meth od makes use of an associated Bernoulli model, with standard errors ba sed on the observed information matrix. Three types of models are disc ussed: additive and multiplicative hazard models with piecewise consta nt baseline hazard, and a proportional hazards model with discrete bas eline survivor function. These models may be fitted in the statistical package GLIM.