If things work you can prove they work.
I'm not so sure I agree with that. In general it's a pretty good rule of thumb for simple linear causality - in other cases, I would say that if something works, if you are testing the correct root cause, then you can prove it...but proving causality sometimes comes down to being lucky, if you are talking about a complicated system with multiple root causes or emergent properties (complex adaptive systems being the new buzzword), then it's not such a slam dunk.
Sometimes there's a correlation and not causality, sometimes it's a coincidence that you don't see because of your preconceptions (remember the movie Medicine Man?)
The history of science is full of times when we thought we knew why something happened, and it turned out we were wrong.
edited to add: I suppose I should mention that my post is talking about 'science' and 'the scientific' in general, and was critiquing the general (modernist, positivist) notion that "it can be proved!"
(critiques of science are apparently a by-product of having to write an anthro paper on Western "science-based" notions like what is nature, wilderness, and biodiversity...but I digress)...I wasn't specifically referring to "quack" medicine...although I will mention that most anthropologists who study Western medicine (one of my profs included) will point out that a lot of Western medicine has no more basis in "fact" than the stuff we call quackery...but that's a big debate I don't have the time or knowledge to summarize now!!!