They said that they collected prior clinical symptoms in the survey, but have not mentioned that they adjusted for those.
Given that, I think it's fair to say that they have not adjusted for the selection bias of people with prior clinical symptoms being more likely to click the ad.