Bias correction with count data

Hi, 

I was wondering if there is any specificity that need to be input in bias.correct.control=list(sd=FALSE, split=NULL, nsplit=1, vars_to_correct="Index_cyl") when dealing with count data models. 
I seem to get a good convergence but the bias correction seems to struggle to back calculate my indices (Getting all NAs in parameter_estimates$SD$unbiased$value). 

Thank you for your help. 
