0
$\begingroup$

Joint Models are mainly viewed in the literature within the context of longitudinal and time-to-event data. For this reason R packages as JMBayes were built to fit these kind of models. Nonetheless I would like to fit Joint Model for longitudinal and binary data, where the binary stands for the outcomes (e.g. death/alive), but I found no available package in R to do this. Does someone know some package or routine to overcome to this problem and fit such a model in R? In particular,for those who knows well the argument, I'd like to fit a Joint latent class mixture model with logistic regression submodel for outcome.

$\endgroup$
4
  • $\begingroup$ Sometimes it is more natural to fit a single longitudinal ordinal model, which is certainly more interpretable than having joing models with shared parameters. See fharrell.com/post/talk/cmstat $\endgroup$ Apr 22 at 11:38
  • $\begingroup$ Dear Professor, do you mean to fit a generalised longitudinal mixed effects model? If so, I don’t see how to involve the longitudinal process (eg the biomarker) as a covariate (it should not be possible at all in this model). Probably I did not get your point. Could tou kindly be more precise, if possible? (The link you shared, points at different arguments you treated in your books, but it is quite general in this case). Thank you! $\endgroup$
    – MPep
    Apr 23 at 13:32
  • $\begingroup$ I add the fact that I am interested in considering the trajectory depicted by the biomarker evolving through the time, taking into account also possible different latent classes of time evolution $\endgroup$
    – MPep
    Apr 23 at 13:37
  • $\begingroup$ Joint models do not give you marginal results. For example the part of the joint model devoted to Y=biomarker when adjusted for the shared cross-outcomes parameter will estimate a different thing than a Y=biomarker model that omits this shared parameter. And it's unclear what 'latent classes of time evolution' means. I hope it doesn't mean discrete classes, which would be artificial impositions on the data. $\endgroup$ Apr 23 at 14:36

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy