Yeah, this article seems to be measuring detentions, including short term holds (different than longterm commitments), but not unique by person. So it's detentions per population vs unique people detained per population. I assume there is a high recurrence rate.