I spent a long, long time trying to answer a question that I really care about. Is moderate drinking—say, 1 glass of wine a night (and, let's be honest, sometimes 2!)—"okay"? Here's the article: theatlantic.com/ideas/archiv… 4 things to emphasize: 1. The science is a total mess The surgeon general wants new cancer warning labels on booze based on large meta-analyses. But wait: The National Academies also did a large meta-analysis and found moderate drinking is associated with longer lives. But wait: Many popular podcasts emphasize the risk of drinking *any* alcohol on cancer and brain matter shrinkage. But wait: Many other popular scientists point out that those studies are deeply flawed. Ahhh!! 2. Beware over-corrections in science The 1990s conventional wisdom that moderate amounts of red wine were very good for your heart were (almost certainly) based on over-confident conclusions drawn from bad studies with improper comparisons between moderate drinkers (who are often healthy for reasons that have nothing to do with alcohol) and non-drinkers (who are often less healthy for reasons that have nothing to do with abstinence). We're smarter now. We've fixed this error. But now we're talking about slapping warnings on wine bottles about mouth and throat and breast cancer. And I'm worried we're throwing the same bad "fruit" into the "juicer" of health meta-analyses and calling it a perfect smoothie (to paragraphse @VPrasadMDMPH). 3. When evaluating health warnings, consider the difference between relative and absolute risk. Even if you choose to take the new cautious CW seriously, the upshot is that moderate drinking only slightly increases the risk of some cancers from an low absolute number to another low absolute number. What might appear to be a large increase in relative risk—say, a 40% increase in oral cancers—is a very small increase in absolute risk, because oral cancer is so rare to begin with. 4. "Every drink of alcohol decreases your expected lifespan by five minutes" *If* you choose to believe the new observational study meta-analyses (an important if! see point 1 above!) this is the all-cause mortality upshot. That sounds kinda bad. But as @euanashley and his team at Stanford have also calculated from observational studies, every minute of exercise *increases* your expected lifespan by five minutes! So, there's your new longevity math: HAVE ONE DRINK; JOG FOR ONE MINUTE. TLDR: I'm gonna keep enjoying my evening glass of pinot. Because, at the end of the day, I'm much more confident that I love wine and that it makes me happy that I'm confident that messy observational study meta-analyses should rule my life.