# Sensitivity analysis after multiple imputation

After reviewing the literature there seems to be little consensus regarding the best way of performing sensitivity analysis following multiple imputation for missing values. However, the growing consensus is that some sort of sensitivity analysis needs to be performed in relation to possibility data is MNAR rather than the traditional assumption that data is MAR (example 1, example 2). I am using the mice package to perform the MI, but there doesn't seem to be clear support for sensitivity analysis and I can not find a current package that provides support for this (I did find SensMice and a slightly updated derivative), however neither of these seem to work any longer. I was wondering if there's a current package or simpler approach that is able to complete this - or a tutorial for completing this in R. In particular, the data I have has one missing categorical and one missing continuous variable.