Note that [3] seems to be using historical data from patients correlating episodes with cannabis use (doesn't seem to include trials), so it's of course very su

Note that [3] seems to be using historical data from patients correlating episodes with cannabis use (doesn't seem to include trials), so it's of course very susceptible to "correlation vs causaution" problems.