Hi

I'm currently trying to use multiple imputation for a register based cohort analysis on maternal smoking during pregnancy and the risk of cryptorchidism and hypospadias.
My exposure smoke_cat i categorized as 0-6 + missings (4%). Category 6 is "Smoker amount unknown".
I wish to predict the missings as category 0-5 (0: Never, 1: Former, 2:<5cig/day, 3: 6-10, 4: 11-20, 5: >21). I also want to treat category 6 "Smoker, amount unknown" as missing, but to only predict it into category 2-5. How is that possible.

My current syntax is the following:

mi impute chained (ologit) smoke_cat education (regress) gestage_week birthweight (logit) parity maritalstatus (mlogit) nationality = boy_no cryptorchidism hypospadias maternalage birthyear, add(100)