I have a longitudinal dataset that is hierarchical in nature with repeated measures of outcomes taken on individuals that are nested within groups. The only variables with missing are two dependent variables which are continuous. I have been using mixed effects models to make statistical inferences about my treatment of interest, but have been wondering whether there are multi-level multiple imputation techniques that I can use to estimate the missing data points.

I came across this previously: https://www.stata.com/support/faqs/s...and-mi-impute/, but it tackles missing data in the independent variables in the context of a two-level design. Is anyone familiar with any expansions of this for a three-level design?

Stata Version 15.1.