If they cannot detect a difference in behavior, is there a need to detect what isnt causing it?

I.e. there are so many different psychoactive substances, that effect, rather than presence, seems to be the only 'fair' way to discriminate in a world where most people take at least 1 medication a week or more.