GAD and PTSD and all that spectrum are coping mechanisms, they're just maladaptive. CBT teaches you how to cope productively.
(personal opinion incoming), I think a person can get to the point where they need an SSRI as a stepping stone to even have the energy/capability to benefit from CBT, then they learn adaptive behavior, then they can dump the SSRI and let their new skills handle day-to-day life.
The "underlying reasons" are only useful inasmuch as you know what to look for so you can proactively start coping/remove yourself from maladaptive situations, but IMO most people try way too hard with the "why" stuff. The only thing that matters is "what now?"
I followed your link to that paper -- sure, CBT, once unwind the acronym, sounds fine. That's what I thought would work on the patient in my family (by marriage, not by birth). But CBT didn't work. Not even a little, not a chance. And the patient was just awash in cognitive ability -- brilliant, actually. Or, maybe the brilliance caused seeing more threats and, thus, more anxiety.
The usual suspects, say, SSRIs, etc. didn't work, either. The result was, right, the worst possible, and of course there are hints that SSRIs can contribute to that result.
Much of the problem was social phobia, e.g., as in the link. Well, it was strongly in common to the mother and all three daughters. So, nature or nurture? If just nurture, maybe CBT, etc. should have worked. But, gotta tell you, at least on the social phobia part, nothing made a dent. All three of the daughters had at least talk therapy, and, no help at all. None.
If CBT can work, as in the link, fine, no, terrific. But, for anyone facing the problems mentioned in the link, need to keep in mind that for a specific case the averages don't have to matter and have to entertain that maybe CBT won't work.
Just why the Chief was so down on talk therapy, or CBT if that is close enough, I don't know. But, we're talking 20+ years ago.
It can be serious stuff, and darned tough to deal with.
Or, the Chief's summary remark went: "Get the patient all calmed down, stable, happy, and then suddenly there will some little event, say, a new file folder of work, and the patient will be all stressed out again."
Or, in my intuition, provide something like a padded cell life for the patient, and things could look fine. But, try to have the patient address the real world, and just some random, new event, say, where there might be some risk that they would have to think through and handle, and they could get all "stressed out" again -- sleepless nights, GI problems, tears, depression, clinical depression, etc.
I'm no expert, but, again, overall IMHO the OP is a bit simplistic.
What we do know is that CBT is exceptionally good at handling most cases. No single treatment (for any disorder) works equally well/at all for all patients, but of course medicine is concerned with helping the largest population efficiently.
I hope one day we understand the brain better, to the point that people like your relative don't have to suffer anymore.
CBT is based on challenging negative thoughts, assuming that those thoughts are manufactured or, at least, grossly exaggerated. That kinda puts you in a corner if those thoughts are factual.
ACT (Acceptance and Commitment Therapy) has more recent science behind it and gives you strategies on how to minimize the effect of unhelpful thoughts, factual or not.
https://en.wikipedia.org/wiki/Acceptance_and_commitment_ther...
Short term CBT style (which is recommended first line treatment) and long form therapies such as psychdynamic counselling. These long form talking therapies tend not to be useful for anxiety.
