This comment is so strange. You've clearly described a group prone to confounding factors, identified a viable case study outlier, and then just concluded that the dependent variable is causally impactful anyway.

Don't you think it's more likely that the cannabis use in this group is highly correlated with other risk factors?