[frereubu]

There are a lot of people who "feel" that data should be able to be used in situations that are much more sketchy than you're suggesting. If you don't abide by the initial permissions that people give, how do you decide whose feelings win out? The best you can hope for is to ask people explicitly in the original study whether they're happy to be contacted for a follow-up - it's not clear to me whether that was the case here.

[foolmeonce]

This is all quite true, the loophole seems to be that you can take these rules seriously but have much less obligation to keep what you publish within the limits your choices within the rules create.

The original publisher could only claim to know it had an unconfirmable suppressive affect on drinking for the short period they were allowed to contact the subjects. The limits were clearly used to imply more and alternative researchers that responsibly asked for sufficient permissions at the start would have been at a double disadvantage. (Harder to recruit subjects and results that were less likely to be remarkable.)