openpharma / brms.mmrm

R package to run Bayesian MMRMs using {brms}
https://openpharma.github.io/brms.mmrm/
Other
18 stars 2 forks source link

The `role` argument of brm_data() is unnecessary #119

Closed wlandau closed 2 months ago

wlandau commented 2 months ago

The purpose of the role argument of brm_data() is to declare whether the outcome variable is raw response (AVAL) or change from baseline (CHG). This matters because brm_marginal_draws() needs to know whether it can compute change from baseline. However, it is possible (and sometimes desirable) to analyze AVAL for datasets that do not even have a baseline value in the discrete time variable. So the question is not whether the outcome is AVAL or CHG, the question is whether the user specifies a baseline value in the reference_time argument. reference_time alone should be enough to decide. We can deprecate the role argument entirely.