> How do you know what actually works for you, and what simply coincided with a change that happened for some other reason? It is literally impossible to tease

> How do you know what actually works for you, and what simply coincided with a change that happened for some other reason? It is literally impossible to tease out these questions without some kind of control. It depends on the question. For lots of medical issues you could serve as your own control. If your medical issue is a rash or acne or trouble sleeping well or high blood pressure or any other condition that persists over time and is measurable, you can just start tracking it then try various changes and see how the metric matches with the changes. Do two weeks on and two weeks off a few times - either there's a break in the chart - a big noticeable improvement - at the appropriate times or there isn't. I agree this wouldn't work for cancer, but it would work for a lot of things. If the effect magnitude is large enough you don't need a big clinical trial to see it. And if the treatment is cheap and available enough there's no chance in hell that anybody is going to fund a big clinical trial - trials are really only done for drugs that can be patented, so whoever sponsors the tests has a way to get their money back. So long as it takes hundreds of millions of dollars to get a treatment past the FDA's tests, there are inevitably going to be a huge number of perfectly valid treatment options that nobody has financial incentive to pay for testing in that particular way.