Fig. 2From: Joint modeling of multivariate longitudinal data and survival data in several observational studies of Huntington’s diseaseDiagram for the proportional hazards model (left) and the joint model survival submodel (right). The down-stream variable is the log hazard, which is a weighted combination of the up-stream variables, with the weights being the arrow labels. The dotted lines indicate correlation among the covariates; t* is time on study (t* = 0 is study entry), and t is age (t = 0 is birth)Back to article page