Ah, so you first check for a candidate and then check a (somewhat random) group of people infected with that candidate for severity. That sounds sensible. Thanks for updating a stranger from the internet!