MODEL EFEK RANDOM UNTUK PEMODELAN BERSAMA DATA TIME-TO-EVENT DAN LONGITUDINAL; RANDOM-EFFECTS MODELS FOR JOINT MODELLING OF LONGITUDINAL AND TIME-TO-EVENT OUTCOMES

AMELIA RAHMA MASNIARI, Herni Utami

2015 | Skripsi | FMIPA

The term of joint modeling refer to model-based methods for the analysis of data from a longitudinal study in which each subject produces outcome data of two kinds: a repeated of measurements at pre-specified follow-up times, and a point process of events in time. These models attempt to explain the relationship between the measurement repeated data Y and the event time data T through their shared dependentce on unobserved random effects. Joint model use Gaussian linear mixed-effects model for the longitudinal measurements and a Cox proportional hazards model for the event times conditional on the random effects. Specifically, we assumed a random-intercept-and-slope model, and developed an expectation–maximization (EM) algorithm for maximum likelihood estimation. Standard errors and optional confidence interval calculated via bootstrap for a joint model fit. Joint model is applied to real data that obtained from recall and medical records patient of Hypoalbuminaemia Hospitalization Patients in Dr. Sardjito Yogyakarta

Kata Kunci : joint model; time-dependentt covariat; frailty; longitudinal data; survival data; biomarker