In a model we are developping for a project, we use a Cox Proportional Hazard model.

We compared the table from the summary

and the tab_model() of the model.

What we have observed is the discrepancy between the 2 outputs: if you look to the variable AGE_J3_c you would find that the 95% CI of the hazard ratio are slightly different. In the summary from R the 95%CI is 0.91095-0.9984 (P<.05) but the tab_model output is slightly different. P-value is the same but 95% CI is now 0.89-1.02. Does anyone know where it comes from? We are trying to look for the explanation for this difference but were not successful until now. Do you know where it comes from?