And the reason for this is surprisingly technical: Medical journals from all statistics and mathematics available in the world only accept randomised double-blind trials.

Now try and find a reference population of people with "skill and desire to train breathing techniques to improve their health". Administer half of those people a "placebo" technique, where both the trainer and the subject are convinced they are working on the "correct breathing technique". Measure the improved health condition. Deliver "statistically significant results" showing how the placebo group improved less than the actual test group.