I think you're mixing up a couple things. One is scientific theory. And indeed psychiatry is not at all in good shape as pertains to scientific theory. Neither is neurology.

The other is effective treatment. You can have effective treatments without good theories. The human body is incredibly complicated, and scientists don't understand the details very well at all. However, if you give some people a chemical and some symptom they have goes away and their life is better (whereas placebo controls don't) then it's an effective medicine and well worthwhile. (Yes, corruption from big pharma in the testing is a problem, but not an insurmountable one).

The story for the development of Viagra is fascinating. It was in a trial for some other condition, and it didn't work so well. As part of the study protocol, the people agree to give back unused medicine when it's over. In this case, the return rate was unusually low, and someone took the initiative to ask them why. The effect on erectile function was a complete, unexpected accident. They hadn't the foggiest idea it would have that effect. They may have figured it out after the fact, sort of.

You are committing a very common mistake. You are assuming that because some things, like hallucinations and some cases of depression, as in your examples ... are valid organic diseases which have an empirically testable profile, that then all things that psychiatrists describe as diseases truly are diseases

I don't think I ever said that. I know I've never believed it.

Things don't have to have a known organic basis to be subject of valid scientific inquiry. If you look at something like human visual perception and things that come readily to mind like optical illusions, they can be studied very reliably without having any idea what brain processes are involved. All they are measuring is what people say they see.