The Cox Model With Longitudinal Covariates And It’s Application In Assessing Effectiveness Of Treatment

Cox model has become more and more popular in survival analysis. How to build a good enough survival model is always an important challenge in survival analysis. With the deepening of the study, researchers find that the survival analysis method only based on the classical cox model is not enough to interpret the clinical manifestation of some disease. So the classical cox model need to be made supplement and amendment for interpreting the clinical data and ultimately guiding the clinical medication and treatment. In this paper,we build the so-called joint model with longitudinal data as the covariates that using linear mixed model fitted and Cox’s regression model. To amend deviation due to fitting cox model, the longitudinal time- dependent data analysis method is use, thus the analysis accuracy is improved.. We combine the longitudinal submodel and the survival submodel to build a joint model using two-step method. We fit the joint model with the clinical data for the survival analysis. Because the data is not suitable for the proportional hazards assumption, there are more deviation if we using Cox model hardly, special, using accelerated failure time model to build joint model, and compared with the results of the Cox model.