When calculating fitted values ($\hat{y}$), we currently don't provide them for the participants with a value below the LOD. We are going to change that behavior to use the mean of the sampled values for each covariate subject to an LOD, for each participant.
This has been implemented in latest Github push. I also added the option to use the lower limit of detection (though I set the mean to the default option).
When calculating fitted values ($\hat{y}$), we currently don't provide them for the participants with a value below the LOD. We are going to change that behavior to use the mean of the sampled values for each covariate subject to an LOD, for each participant.