Where do they get the data?

So health organizations around the world seem to recommend fairly restrictive drinking limits like maybe 2-5 drinks a week. At the same time studies of people living over 100 show most of them drink significantly more than that. Doctors pull health data from patient survey and corelate it with disease those patients have. They don’t do a double blind study where they slowly kill one group with alcohol. People often lie to their doctor about their drinking especially if they have health problems due to it. A history of alcohol abuse can do things like put you lower on the transplant list so the incentive to lie is there. Does anyone think that as a result health organizations are getting data that overstates the danger of alcohol due to patients with disease understating their consumption?