Is Type III Analysis Effect in Proc-Lifereg reliable?

[Tcook]

Hi All,

Please I am fitting an AFT model using Proc Lifereg.

I have several levels for my categorical variables and in the type III effect the covariates are significant but in the model output most of the levels of the variables are not significant.

Please is this a problem, if yes how can I fix it?

Moreover, it seems Proc Lifereg does not give robust estimates for recurrent event time, is the type III effect result still correct given that I have recurrent events per subject?

Thanks.

Cook.