교차위험 자료에 대한 가속화위험 모델링 접근법

Abstract

Proportional hazards model and accelerated failure time model have been widely used in survival analysis. However, the two models can not capture time-scaled effects such as crossing hazard or the gradual effect of treatment over time. For the purpose of this detection, accelerated hazards (AH) model (Chen and Wang, 2000a) has been studied. In this thesis, we propose the use of log-logistic distribution for the baseline hazard function of the AH model. In particular, the log-logistic distribution gives various forms (e.g. non-monotone) of hazard function and also an easy implementation for likelihood-based model fitting due to an explicit form of hazard and survival functions. The likelihood-based estimation procedures of the AH models are derived. The proposed method is extended to AH model with time-effect between groups and also to mixture cure model. Our method is demonstrated with two practical data sets with model cheking.