I don’t think any of this is a mystery to people in psychology or substance abuse. Also, this is data at a high aggregate level and, as you mention, so much of this problem is personal (or interpersonal). It’s hard to assign cause and effect to environmental causes when person-level factors weigh so heavily.