ABSTRACT
Joint models for longitudinal data and censored time-to-events have been developed since the late 1990s; however, they were little used until the late 2000s because of their complexity and the lack of software. The situation changed with the release of several R packages dedicated to these models (JM, JMbayes, joineR and lcmm), and also of some functions in general statistical software (such as stjm in Stata or JMFit in SAS) and the publication of a monograph on some of these models (Rizopoulos, 2012b).
