I think you've (unsurprisingly, given it comes from SAMHSA) hit upon an important concept -- recovery can happen (almost?) independently of clinical outcomes -- which really challenges clinical service researchers, and those reviewing their research, to include recovery outcomes and not just clinical outcomes in summative evaluations of treatment program "success". This disturbs me when someone pointed out what therapy approaches may be recommended, as was done recently.
For example, in reviews of Acceptance and Commitment Therapy, if one looks at helping people with chronic pain (and lots of people with ACEs have such problems), including "pain reduction" as an outcome would be grand, but given that CBT/ACT can bring about, even in those with the most "disabling" pain, meaningful, valued, lives, pain reduction almost becomes irrelevant (look at Lance McCracken's work, for example). A similar conclusion can be drawn for those with psychotic symptoms -- almost like the idyllic outcome as seen in "A Beautiful Mind".
So one should be careful in reviewing "randomised clinical trials" if a narrow range of outcomes are included -- should one be happy if a patient's psychotic symptoms are reduced if they also suffer disabling side effects from the medication, and thus their "quality of life" is still pretty poor -- the way things often were in the past, when Haloperidol would tan your hide, and Melleril would turn you blind, and ... (etc).
Unfortunately, because your post uses fairly global terms, it leaves me unsettled and questioning -- is this true for everybody, in every situation, with every "problem" ?, although no exceptions spring easily to mind.