A Joint Model Of Longitudinal And Survival Data Based On Parametrized Method With Its Application To Clinical Trial Data Analysis

Primary biliary cirrhosis(PBC)is a chronic,low-incidence,but fatal liver disease characterized by inflammatory destruction of the small bile ducts in the liver,which ultimately leads to cirrhosis.PBC can be slowed by treatment with ursodeoxycholic acid,and liver transplantation is an effective treatment for advanced patients.In the study of primary biliary cirrhosis,in order to evaluate the efficacy of treatment,usually involves two kinds of associated data,one is the progress of some primary biliary cirrhosis index data(longitudinal observation data),the other is a survival data,and clinical doctors are concerned about the efficacy of these treatments and can prolong the survival time of patients,and based on these data to evaluate curative effect is the best way is to establish a joint model of longitudinal and survival data.In this paper,a joint model with time-varying slope and lag effect is established for PBC data,and compared with the classical joint model,the maximum likelihood estimation is used to estimate the parameters of the joint model,and the Log.Lik,AIC and BIC information criteria are used to judge the advantages and disadvantages of the model.At the same time,dynamic prediction of clinical efficacy index is made,so as to provide an effective and high-precision model for individual assessment of PBCThe conclusions are as follows: the difference of mean square error of the three joint models is small and all of them are small,among which the AIC and BIC of the joint model with lag effect are small,and the joint model with lag effect has a small AIC and BIC,which has a good dynamic prediction effect on the therapeutic index.In the prediction of individual survival probability and longitudinal outcomes,a better prediction effect can be obtained by parameterizing the survival sub-model.From the perspective of parameter estimation time,a relatively simple joint model is chosen for dynamic prediction on the premise of ensuring the accuracy of prediction.The purpose of this study is to provide clinicians with reference recommendations for early intervention in patients with PBC,and to provide alternative methods for joint modeling of other initial and survival data.All data analysis in this paper is implemented by R software.