The development of methodology and software for fitting joint models to multivariate longitudinal and time-to-event data.

Description As of version 1.2.0, joineR fits the joint model proposed by Williamson et al. (2008) for joint models of longitudinal data and competing risks. Here, the longitudinal data submodel remains as per that analyzed in Henderson et al. (2000), namely
[Y_i(t) = X_i(t)^\top \beta + Z_i(t)^\top U_i + \epsilon_i(t),]
where (Y_i(t)) is a repeated measurement on subject (i) at time (t), (U_i) is a latent vector that follows a zero-mean multivariate normal distribution, (X_i(t)) and (Z_i(t)) are vectors of explanatory variables that may be time-constant or time-varying, and the (\epsilon_i(t)) are mutually independent errors, (\epsilon_i(t) \sim N(0, \tau^2)).

The joineRML package implements methods for analyzing data from multiple longitudinal studies in which the responses from each subject consists of time-sequences of repeated measurements and a possibly censored time-to-event outcome. The modelling framework for the repeated measurements is the multivariate linear mixed effects model. The model for the time-to-event outcome is a Cox proportional hazards model with log-Gaussian frailty. Stochastic dependence is captured by allowing the Gaussian random effects of the linear model to be correlated with the frailty term of the Cox proportional hazards model.

An R package to fit joint models to a single repeated measure and a time-to-event outcome (single or competing risks) using an EM algorithm.

An R package to fit joint models to multivariate repeated measures data and a time-to-event outcome using an MCEM algorithm.

© 2017 Graeme L. Hickey · Powered by the Academic theme for Hugo.