Did they collect data to allow them to split patients into subgroups by case definition? I've been wondering why studies haven't done this as a matter of course, for years - research could have been a lot farther ahead if they had.
Study protocols should include what analyses are planned, especially big, publicly funded ones like this - some study protocols are published and registered with clinical trials registries (partly so that researchers can't change their story later about what analyses they intended). I would think the PACE planned analyses would be on public record somewhere.