Comment on 'Generating survival times to simulate Cox proportional hazards models with time-varying covariates'

摘要

Peter C. Austin's paper [1] gives a good description of generation of survival times with time-varying covariates. Nevertheless, we have found a serious mistake in the derivation of one simulation formula. In Section 3.2.2., Austin considers the Weibull distribution of survival times with continuous time-varying covariate z(t) = kt.