stcrreg competing risk model with random intercept

[Moses Ngari]

Dear Stata community,

I have run the Fine-Gray competing risk model using stcrreg and worked perfectly well. However, my data is from several sites which are significantly different. I would like to account for this heterogeneity explicitly using a random effect multilevel model. I am aware the stcrreg support an option for robust SE but unfortunately this method does not explicitly account for the heterogeneity.

Here is sample model

stcrreg i.age_cat,comp(cevent) vce (cluster site) //this runs

The multilevel model I need would look like this

stcrreg i.age_cat || site:,comp(cevent)

However the above model does not work. Does in anyone know how I can control for the competing event as well as the heterogeneity in the model?

Thanks
Best regards Moses Ngari