1997 · 10 citations · 0 references
Public HealthBiostatisticsSurvival AnalysisChange-point DetectionHazard FunctionsKaplan-meier EstimationMaximum Likelihood EstimationParametric Survival ModelsSurvival ModelingChange-point Confidence IntervalsChange-point EstimationLikelihood Ratio TestingLikelihood Ratio TestsLikelihood RatiosLikelihood-ratio-based InferenceParametric Change-point Detection
this paper, we consider a modification of the proportional hazards failure time regression model, also used by Liang et al. (1990), which includes a threshold lag effect for some of the covariates. The likelihood ratio test is proposed for testing the hypothesis that the changepoint is equal to some specified value. Asymptotic properties of the test statistic are derived. Examples and an application are given. For simplicity of exposition, we start with the two-sample problem where there is a single binary covariate, assigned treatment say. The more general case is described in Section 5. Proofs appear in the Appendix.