I am deeply sceptical about counselling and "talk therapy". Partly, it dates backs to having read Jeffrey Masson's
Against Therapy years ago. Partly, it is based on observing counselling that seems to have either being extremely ineffectual, or made things worse.
My scepticism is based on two problems, one epistemic and the other evidentiary. The scientific/evidentiary basis of a lot of the theories used in therapy over the decades is -- to put it mildly -- thin. Rather worse than that used in medicine, which is not without its own problems.
The other problem is, if anything, more fundamental. The therapist is reliant on what the client tells them. They normally have no independent way of checking its accuracy or truthfulness. Forensic psychologist Nigel Latta, in his splendid
Into the Darklands: Unveiling the Predators Amongst Us, makes the point he does not bother reading the client's file -- that is just a record of the lies they have told previous clinicians. Instead, he reads the transcripts of their trial, since that tells him what actually happened.
But Nigel Latta is dealing with convicted criminals, so such evidence is available to him. Most counsellors/therapists do not have such an information source available to them. So they are in a very weak position to assess the accuracy or truthfulness of what a client is telling them and, in particular, in a weak position to assess the social dynamics the client lives in.
If they are dealing with a perceptive client with a low capacity for self-delusion, or misreading of emotional and other cues and dynamics, not such a problem. But the more the client has a problem with either, the less the therapist/counsellor has to go on and the more dominated by errors of omission or commission what the client/patient tells them will be.
In other words, the epistemic problem will tend to get worse the more the client/patient has inherent difficulties or problems with self-delusion, misreading or failing to notice. If they are sufficiently poor at it, the therapist/counsellor may find it easier to pick up contradictory cues but, if the client/patient is relatively high functioning, not so much.
Hence, for example, counselling tends to make narcissists worse, since their self-editing of reality is both so unconscious, and so complete, that they are likely to give very few cues for the therapist/counsellor to pick up while the process of therapy tends to validate their emotions, thus feeding the process of self-validation being their reality principle (which is the essence of narcissism).
Why couple's counselling is likely to be better
All of which puts couple's counselling in a bit of a different category. There is still the issue of the evidentiary base of whatever approach the therapist/counsellor is using. But there is much less of a fundamental epistemic problem, since the counsellor/therapist has two points of view to engage with.
Indeed, the evidentiary bases is likely to be less of a problem anyway, given dealing with blocked communication, problems of framing, etc are probably not so dependent on background theory.
So, I tend to be deeply dubious about counselling/talk therapy but rather more supportive of couple's counselling.
UPDATE I used to suffer from anxiety attacks and depression and have surmounted both without counselling or drugs. I did, however, do a lot of reading, thinking and talking with perceptive friends.