Almost all studies use untrained vs trained athletes. They rarely qualified what "trained" means and could be anything from a lifetime athlete to someone who self-reported following a 6 week plan prior to the study and now we consider this person "trained". Most athletic studies are used to compare a sedentary person to someone who workouts casually (less than about 10 hours per week). Once you pass this point, you're in diminishing returns territory and you've got to actively sacrifice some parts of your overall health to make gains in a single activity (sprinting, jumping, etc.).
So will running a few times a week help fight mild arthritis? Probably. Will training like an Olympic athlete? No. Same thing applies to the OP article. Will sledding down a hill give you brain damage? No. Will constantly sledding at Olympic athlete speeds and g-forces? Seems like it.