I enjoyed reading your systematic consideration of race/ethnicity in evaluating oropharyngeal cancer outcomes. I did not consider using the Barron-Kenny analysis but appreciate how you deconstructed how race can be approached from so many ways. From what I am reading, race/ethnicity can be associated with worse outcomes but whether this is due to variables that mediate this relationship or whether there is an intrinsic link between race/ethnicity and outcomes after correction for these variables is interesting. What I struggle with is how to combat these mediators. We know a confounder can be adjusted for and an effect modifier should be noted. I am not sure if we have determined a way to eliminate a mediator as their on the causal pathway so assume we would just note this as well? Fascinating.