The way the question is posed ("What works? I want your anecdotes") unfortunately really limits the usefulness of the discussion. Due to a large number of well-known and extensively documented cognitive effects and biases (placebo, hindsight, selection, confirmation, expectation, bandwagon, survivorship, etc.), anecdotes and testimonials are one of the *worst* known ways of deciding complex health questions.
The best known way to figure out if a treatment actually works, which we humans have perfected over the past 100 years or so, in the randomized controlled trial. (Having a plausible underlying physical mechanism also is a big plus.) Take a large group of people, randomly assign them one of two alternative treatments (or real vs fake treatment) and measure the results, without participants or researchers knowing which group is which. For example, you truly believe ACV or herb X or amber necklaces worked for you. How do you know it was that one single thing and not something else happening in your life at the time (or placebo effect, or random remission / cyclic nature of the condition)? How did you measure "working"? E.g., did you and 49 friends get 25 real amber necklaces and 25 fake amber ones, randomly draw them without knowing which was which, and then record a daily diary of the child's teething discomfort levels? Without a quality trial, then due to all the cognitive biases and effects mentioned above, we really have no clue whether it was effective. It's true that a collection of anecdotes is a great way of identifying treatments for further clinical testing; it's just not a substitute for testing.
Don't get me wrong, I know that many natural cures do in fact work, and I'm not some pharma-for-everything type of person. I just want people to remember how we really learn which treatments work: science! Not testimonials. So, I encourage everybody to re-pose the question in their minds: "What non-mainstream treatments are supported by strong evidence from high-quality clinical trials?"
The best known way to figure out if a treatment actually works, which we humans have perfected over the past 100 years or so, in the randomized controlled trial. (Having a plausible underlying physical mechanism also is a big plus.) Take a large group of people, randomly assign them one of two alternative treatments (or real vs fake treatment) and measure the results, without participants or researchers knowing which group is which. For example, you truly believe ACV or herb X or amber necklaces worked for you. How do you know it was that one single thing and not something else happening in your life at the time (or placebo effect, or random remission / cyclic nature of the condition)? How did you measure "working"? E.g., did you and 49 friends get 25 real amber necklaces and 25 fake amber ones, randomly draw them without knowing which was which, and then record a daily diary of the child's teething discomfort levels? Without a quality trial, then due to all the cognitive biases and effects mentioned above, we really have no clue whether it was effective. It's true that a collection of anecdotes is a great way of identifying treatments for further clinical testing; it's just not a substitute for testing.
Don't get me wrong, I know that many natural cures do in fact work, and I'm not some pharma-for-everything type of person. I just want people to remember how we really learn which treatments work: science! Not testimonials. So, I encourage everybody to re-pose the question in their minds: "What non-mainstream treatments are supported by strong evidence from high-quality clinical trials?"
Here's a nice, user-friendly starter: Snake Oil? Scientific evidence for popular supplements. And, The Cochrane Collaboration.