RMS Semiparametric Ordinal Longitudinal Model

While I also do prefer to use Bayes, I do not always have to luxury to do so. I also have my doubts about the practicality of the ATE, but currently that’s what everybody wants… We can of course use Bayes to target the ATE as well, but that also takes ~30 minutes, so requires quite a lot of compute for larger simulations.