I am conducting multiple chained imputation (MICE) and I have continuous variables such as BMI and age plus derived categorical variables of these ones. Should I add both in the imputation model or just the categorical variables. Just to note: they are regular variables