I have an epidemiological data set and am interested in the association of exposure to cigarette smoke on child blood pressure. I have run complete case analysis, and as a comparison would like to impute the missing data. Data is missing 5-10% of participants across covariates and is MAR. I am following the stata help guide to MI but am slightly confused about a couple of things.

So far I have run;

mi set wide

## listed the covariates I would like impute
mi register impute maternal_alcohol family_income child_IQ child_height

## here I have listed my outcome, exposure and vars that are complete but are related to my covariates I'd like to impute- is it appropriate to include the exposure and outcome?
mi register regular child_BP maternal_smoking maternal_BP paternal_BP

## my imputation command as follows- again after the "=" should I be including all complete vars that are related?

mi impute chained (regress) childIQ child_height (mlogit) family_income maternal_alcohol = child_BP i.maternal_smoking maternal_BP paternal_BP, add(20)

This results in 20 sets of data for each covar, and I am confused as how to merge these into a functional dataset? If anyone could advise on this (and of course if there are any errors in the way I have gone about this). If this is an appropriate way to go about imputing my covariates, what checks should I be doing before I continue with my main analyses?

Many thanks for any help and advice on this.