The first I would paraphrase as: "The patient refuses to accept that his condition is psychological, and we do him a disservice by validating the idea that it might not be."
The second I would paraphrase as: "While there is always the possibility that this is a new condition we don't yet understand, statistically speaking it is far more likely to be psychological. Given finite resources, diagnosis and treatment for depression/anxiety is by far the best bet for giving the patient real relief."
To me the certainty of exclusion in the first message just immediately sets off alarm bells.
I think part of the miscommunication may lie in professional idiom. In medicine, all diagnosis is provisional and statistical - which both elevates "probably" to "certainly" in casual conversation, and renders most absolute statements hollow. It's just an automatic translation that has to happen: when you tell someone with a three-day sniffle that they have a cold, it's an absolute statement aloud, but in your head you have to honestly say "99.999% cold; 0.0009% HIV; 0.0001% cancer." And you have to treat them accordingly, until proven otherwise, because if you try to rule out cancer on everyone with a sniffle you're going to actually kill more people with complications than you'll ever save.
"This person has X" most of the time means "X has a strong statistical advantage in explaining this patient's symptoms, so we're going to pursue treatment for X until accruing evidence suggests we shouldn't." I guess that's not always true, given a few things that exist by definition, but it's mostly true.
And that's generally good enough, since... well, statistical masses tend to obey statistical properties.
