A glass of wine a day is within epsilon of the most healthy possible option. You're making this out as if this is a big shift, but it isn't. There are just huge error bars on the measurements relative to the effect of the intervention.