That's nearly a contradiction in terms. Mainstream medicine, as a practice, refuses empirical experimentation on particular patients. Here's the disease. Here is the cookbook cure that has been sanctioned by some professional board, with backing by a billion dollar study that shows some statistical advantage over a whole population. If there is no such study, then the prescription is a shrug and a pat on the head.

This is actually not quite true. There's been a surprising amount of self-experimentation in medicine and experimentation on individual patients. Lawrence Altman's "Who Goes First?" is a somewhat dry book on this subject that looks at self-experimentation by doctors and scientists in history. Most examples are things where they do deliberate harm to themselves (e.g. trying to figure out how a specific disease is transmitted by trying all sorts of different methods of infection, or deliberately reducing their intake of some specific vitamin and measuring the effects) but others involve actively trying to cure diseases that they happen to have. (Edit: For one recent unsuccessful example see here).

If it works for you, then it works for you. No billion dollar study required, and it doesn't matter what the statistical averages are when you determine it works for you.

It is often difficult to tell if actually works for you or not. Self-evaluation is really tough, and when your sample size is one, the noise level can easily outweigh the signal. Still, there are actual steps that self-experimentation can include that help a lot. One can without too much difficulty double-blind self-tests with a help of a friend.