Having worked in that field, fwiw, it is shockingly difficult to attribute causality to any changes you see. Even if done in good faith you usually need really large trial cohorts to know if people are "faring better" because of an intervention.

Plus placebo and nocebo are really real..