I have a dataset of 30,000 patients who are matched by surgical procedure. My main exposure of interest is if the patient got regional anesthesia during surgery or did not. I have 27 different surgical procedure types, so instead of including 26 dummy variables in my Cox regression model, I stratified on them. In addition to the main exposure of interest in the model (regional anesthesia/no regional anesthesia), I have 3 other covariates I'm controlling for...Is there a simple way to calculate median survival for the patients who got regional anesthesia and for those that did not? Or would I have to calculate median survival for these two groups for all 27 strata since they all have different baseline hazards?