The Joint Model Of Longitudinal And Survival Data And It’s Application In AIDS Clinical Comparative Effectiveness Research

In the clinical trials, medical and reliability studies, we often collect both repeat measurements of longitudinal data and survival data associated with survival time. In some practical problems, we often need to observe a response variable; and need to record the survival time of individual, it may have the censored data in the process. A lot of effective methods exist for analyzing longitudinal data and survival data independently, but it is not suitable when the longitudinal data is associated with the survival data. In order to solve this problem, the joint modeling of longitudinal and survival data is raised. Joint modeling of longitudinal and survival data method is often used to solve such problems, so it has some actual application value.In the past twenty years, joint models of the longitudinal and the survival data have been widely concerned. These models usually include the following conditions: time-dependent covariates of survival models have missing data or measurement errors; and longitudinal models have information dropouts;and a longitudinal process and a survival process are correlated with latent variables. In these cases, if we use the separated longitudinal model or survival model, it may result biased or the poor effect.This paper mainly introduces the research status of joint models of the longitudinal and the survival data, separately introduces the longitudinal model and survival model and their estimation method, and then discusses the two-Stage method, one of the joint models of the longitudinal and the survival data. Then we analysis AIDS data with this method,and conduct joint modeling. We also analysis the actual data with R statistical software,and establish a suitable model, and simulation the results, finally come to the conclusion.