Estimating Covariances and Goodness of Fit Plots for Accelerated Failure Time Models

Let the response variable Y be the time until an event such as death. Assume that there are p predictors x1,…,xp and that the response variable is right censored. Several survival regression models, including accelerated failure time models, have the form Z=log(Y)=αZ+xiTβZ+e. This paper gives a simple method for estimating the covariances Cov(xi,Z) for some of these models. Plots are given for checking the goodness of fit of accelerated failure time models. Plots for checking the proportional hazards regression model are often harder to use.