Hello,

I am analysing a longitudinal dataset and I would like to analyse the effect of several factors on my outcome measure (surgical revision hazard) with competing risks regression, where mortality is the competing risk.

The data has missing data in some of the independent variables, which we believe are missing at random. I would like to multiply impute (MI) missing data using MICE. I use Stata/IC 16.1, and Stata command to perform MI (mi impute) does not include crreg as one of the possible methods do impute the data. However, the command to fit a model to a multiply imputed dataset (mi estimate) does include crreg as a possible estimation command.

My question is: what method should be used to impute the data, or how should I impute the data, so that I can fit a crreg model to the multiply imputed data?

Many thanks,

Rocio