| The joint analysis of longitudinal measurements and survival data have been widely used in clinical trials and medical research.In this paper,a new joint model based on B-spline is considered,where the longitudinal measurements come from a linear mixed tt model,and we investigate the survival time with a mixed cure model.The two sub-models are combined by a latent random variable which follow gamma distribution.What is more,survival model is a semi-parametric model with unknown cumulative baseline hazard function,and we estimate parameters by using EM algorithm and B-spline approximation.From the simulation study,the proposed method is efficient,and we apply the method in the longitudinal CD4 and the survival time of patients infected with HIV to compare the effect of two antiretroviral drug therapy ddI(didanosine)and ddC(zalcitabine).The introduction of the article present our research background,research situation and literature review.The first part reviews the basic knowledge of the models and methods we studied.The second part we detail the survival submodel and the longitudinal submodel of the joint model,as well as the statistical inference to obtain the maximum likelihood estimation(MLE)by EM algorithm.In the third part,the simulation was carried out,and the HIV data was analyzed in the fourth part. |
PDF Full Text Request