Joint Model Of Longitudinal Item Response Measurements And Survival Times With A Cure Fraction

In clinical medicine,joint models of longitudinal biomarkers and survival data have been paid more and more attention.In this paper,we propose a new joint semi-parametric model for joint analysis of multiple-item ordered longitudinal data and cured survival data.It combines the propotional odds model with the promotion time cure model by a shared frailty.We use the Fisher-scoring algorithm and the adaptive Gauss-Hermite quadrature to acheive the maximum likelihood inference of the joint model.In the simulation study,the estimation result obtained from the proposed joint model was compared with separate analysis,which showed the advantages of the joint model.In the example,we used the quality of life(QOL)data of epileptic patients.The joint model was used to analyze the different effects between carbamazepine(CBZ)and lamotrigine(LTG)by combining QOL data with survival data of patients.By calculating the area under the QOL curve(AUC),we can complete the comprehensive assessing of drug efficacy.