Dear All,

While adapting multiple imputation in longitudinal datasets to handle missing data in multiple covariates. Is it recommended to carry out multiple imputation model for each analysis model separately? For eg: Model 1 is for crossectional study that is considering only the baseline values, Model 2 is for longitudinal study, Model 3 is would include different covariates. So, in this case, we would have separate multiple imputation model for each analyses model separately. Any recommendations, suggestions are much appreciated,,

Thank you so much!!