I think early alcohol consumption is also associated with alcoholism. Seems like a bit of gamble to not strongly discourage it in adolescence.
(https://pubs.niaaa.nih.gov/publications/AA67/AA67.htm#:~:tex....)
These surveys are also done on American adolescents, who are more likely to obtain and consume alcohol illicitly than their European counterparts. It would be interesting to see comparative statistics on the two.
Finally: this bulletin mixes up different demographics: you have (1) COAs who are more likely to become alcoholics themselves, (2) adolescents who engage in binge drinking, which is generally correlated with alcoholism in adulthood, and (3) an unmeasured population of drinking-but-not-binging adolescents.