Likelihood ratio tests for a changepoint with survival data
生存数据中变点的似然比检验
We consider the likelihood ratio test for a changepoint or lag of the effect of some covariates, e.g. treatment, in regression modelling of survival data subject to right censoring. Asymptotic distributions under null hypotheses are derived. Numerical examples are used to illustrate the implementation of the procedure. Finally we apply the theory to data from a large randomised cancer prevention trial.