0

Is there a way to present results from univariate analysis of multinomial gee model (perhaps using multgee package) using tbl_uvregression from daniel sjobergs gtsummary package?

I found a previous answer for a multinomial model: multinomial logistic regression results table in wide format using the gtsummary package

Would it be possible to do something like that for the multinomial gee model and while removing NAs from variables when running multiple univariate regressions with tbl_uvregression?

Code for running multinomial gee model from the package document for a nominal multinomial outcome:

library(multgee)
data(housing)   
fitmod <- nomLORgee(y ~ factor(time) * sec, data = housing, id = id, repeated = time)
s.g
  • 126
  • 7
  • 1
    That type of model may not be supported. Can you provide sample code that we can run on our machines to create a model? We can then assess 1. if the model is supported, and 2. if not supported, then we can assess if can be supported. (You're the first to mention a multinomial GEE model!) – Daniel D. Sjoberg Oct 08 '22 at 15:46
  • Thanks for the reply, I have added a code snippet taken from the package document for running the multinomial gee model for a nominal outcome (unordered) – s.g Oct 08 '22 at 16:18
  • 1
    Thanks for updating. The issue is that the package does not support standard modeling package functions like `model.matrix()`. Oddly, they do seem to export methods for `model.frame()` and `terms()` (which is great!). – Daniel D. Sjoberg Oct 08 '22 at 20:27
  • 1
    You can follow this GitHub Issue to track progress on its implementation into {broom.helpers} and {gtsummary} – Daniel D. Sjoberg Oct 08 '22 at 22:13

0 Answers0