I think there's a bias in the data because a) lots of people die from the respiratory symptoms before they get a stroke, and b) young people who get a stroke might not know how to recognize it/what to do about it, so they may just stay at home and die there and never get recorded as a COVID death.