Hello all,

I'm trying to fit a Poisson model to the rate of hospitalizations for children under 5 years.

This is my code: glm adm_rate time month, family(poisson) link(nbinomial) offset(oset)

I have added time and month as explanatory variables.

The offset is the log of the number of children at each time point.

I would like to specify the residual correlation structure as autoregressive order 1 (AR(1)). I'm struggling with how to implement this using the glm command. Is this possible? If so, can someone help me with the best way to do this?

Thank you

David