The assumption here is that only people with means got care and were surveyed. I am not sure that this is the case. Moreover, you can correct for those factors, and, I assume, any statistician worth their salt are.
Given the reproducibility crisis, particularly in the social sciences, I wouldn’t put too much weight into the skill or honesty of the people doing that work (and statisticians they are not - more like people with a humanities background who take some statistics courses and then do numerology)
