Brian Reed: Eliza, hey — I have to ask, do you drink?
Eliza Ward: That is a very specific opening. Why?
Brian Reed: Because I've been doing that thing where you read something and then suddenly every choice feels annotated. There's a study — the Alcohol Intake and Health Study, out October 25th, 2026, in the Journal of Studies on Alcohol and Drugs — and ScienceDaily's headline for it is, let me see if I can get this right: 'Just two drinks a day may be far riskier than you think.'
Eliza Ward: And the riskier-than-you-think part is a one-in-twenty-five lifetime death risk. At two drinks a day. Fourteen a week.
Brian Reed: Which is — hang on — that is exactly what the U.S. Dietary Guidelines currently call moderate. That's not heavy drinking. The 2020–2025 guidelines say two drinks per day for men is the moderate threshold, and nothing has been revised.
Eliza Ward: Right, and here's the part that doesn't just surprise me — it actually kind of baffles me: the federal government commissioned this research specifically to inform updated dietary recommendations. The study is complete, it's published, and the guidelines have not moved.
Brian Reed: So the question I want to actually work through — is this a science problem, a communication failure, or something structural about how dietary guidance gets updated at all?
Eliza Ward: All three, probably — but let's start with what one-in-twenty-five actually means for a real person before we get to the policy lag.
Brian Reed: Twenty-five people at a dinner party — walk me through that.
Eliza Ward: Wait, actually let me back up. Lifetime attributable risk is the concept. And what it's saying is: take twenty-five people, all drinking two glasses of wine every night, hold that across their entire lives. The math says one of them dies from an alcohol-related cause that they wouldn't have otherwise. That's the number. One person at that table.
Brian Reed: But it's not — hang on — it's not saying *which* person.
Eliza Ward: Exactly. It's not a personal guarantee. Rutgers and University of Toronto researchers — this is described as the most comprehensive U.S. estimates to date, those exact words from a Columbia Mailman School of Public Health release in June 2026 — they're modeling population-level probability. The 4% isn't written on your liver. It's written across a group.
Brian Reed: So the same study found seven drinks a week was — what, basically fine? Is that the quiet threshold here?
Eliza Ward: Small increases in risk for most conditions — that's the language. Not zero, but small. Which is, I mean, that internal contrast is doing a lot of work. Fourteen drinks a week is one-in-twenty-five. Seven drinks a week is a different calculation entirely. The study holds both of those at the same time.
Brian Reed: The part I don't get — Katherine Keyes at Columbia says no protective effect at *any* level. So where does the seven-drink threshold even come from if the direction of risk is always upward?
Eliza Ward: Right — so what Keyes is saying, and she's specific about this, is no protective effect observed at any consumption level. That's a statement about net outcome. The seven-drink finding isn't 'safe,' it's 'the increases are small enough that they didn't register as significant across most conditions.' Those aren't the same claim. One person at the dinner party still fits the math — the table just got a little bigger.
Brian Reed: But that's the part that keeps snagging me — because Kevin Shield, the lead researcher, he actually notes that low-level drinking may be associated with reduced ischemic heart disease and stroke risk. That's not a fringe claim. That's in the same study.
Eliza Ward: Right — and that's real. He's not dismissing it.
Brian Reed: So the net-harm framing isn't 'the heart benefit was a myth.' It's — let me see if I've got this — it's that cancer and other chronic disease risk on the other side of the ledger just outweighs it. Even at seven drinks a week.
Eliza Ward: That's the accounting. And there's independent weight behind it — the Nature journal Burden of Proof study, which pulled from 843 cohort and case-control studies, 16 systematic reviews, 20 health outcomes. That's a separate meta-analysis landing in roughly the same place: when you add up all the outcomes together, net harm.
Eliza Ward: Across 20 outcomes. So this isn't one paper flipping the table — it's a reweighting. The cardiovascular signal was always real. The question was whether it dominated the full picture, and now there's enough data to say: it doesn't.
Brian Reed: Which means — okay, think about the person who switched to a nightly glass of red wine in 2010 because their cardiologist sort of nodded at the research. Friday evening, glass of Malbec, felt virtuous about it. They weren't acting on a myth. The science just got — I mean, the ledger got rebalanced under them.
Eliza Ward: They weren't wrong then. The cardiovascular benefit was a genuine pillar of moderate-drinking guidance. What shifted is that cancer risk evidence accumulated enough to tip the net-harm calculation. The historical guidance wasn't fabricated — it just wasn't complete.
Brian Reed: And the part that comes later — why a study the federal government literally commissioned to update dietary recommendations hasn't actually moved those recommendations — things get pretty strange there.
Eliza Ward: Commissioned the research, got the answer, and now — the mechanism to update the guidelines doesn't auto-fire. That's the actual structure here. The U.S. Dietary Guidelines run on a five-year cycle. This study hit ScienceDaily on October 25th, 2026. The cycle doesn't care.
Brian Reed: So there's no — hang on — there's no provision where federally commissioned research that contradicts the standing guidelines triggers a mandatory review?
Eliza Ward: Not automatic. Commissioning a study signals intent. It doesn't create an obligation. Those are regulatory and political steps, not scientific ones. The intent and the mechanism are completely separate tracks.
Brian Reed: Which means someone who asks their doctor about this right now is navigating — what, exactly? The 2020–2025 guidelines still say two drinks a day is moderate, the study says that's a one-in-twenty-five lifetime death risk, and nothing official has closed that gap.
Eliza Ward: And the WHO closed it three years ago. January 2023 — no level of alcohol consumption is safe. That's not this study. That's already sitting there, international, and the U.S. didn't move then either.
Brian Reed: Wait — January 2023. That's nearly four years before this study published.
Eliza Ward: Four years. And in that gap, I mean — this is where Gallup comes in, actually. The polling from October 25th shows 52% of Catholics, 56% of religiously unaffiliated adults, 48% of Protestants — majorities across all three groups now say moderate drinking is bad for health. That shift happened without a revised dietary guideline. Culture moved on its own.
Brian Reed: The Catholic number is the one that stops me — 79% drinking in 2002, down to 55% now. That's a 24-point drop. Though the researchers flagged that subgroup survey data can get unreliable. So we're — wait, are we seeing actual behavior change or just attitude change?
Eliza Ward: That's the caveat worth holding. Attitude and behavior aren't the same measurement. But here's what's confirmed: public perception is moving independently of any official revision. The dietary guidelines may no longer be the primary signal people are using. They're reading ScienceDaily, or something downstream of it, and doing the math themselves.
Brian Reed: That's the thing that actually settled for me somewhere in all of this — it's not that the guidelines are wrong, exactly. It's that they've stopped being the thing people consult. A federally commissioned study in the Journal of Studies on Alcohol and Drugs says one-in-twenty-five. The WHO said no safe level in January 2023. Gallup shows the public already moved. And the 2020–2025 guidelines just... sit there. People aren't waiting anymore.
Eliza Ward: Right — and I think that's the part I keep landing on too. The gap between what Katherine Keyes found and what official guidance says is wide enough now that people are just — they're doing their own arithmetic. That's not ideal. That's not how this is supposed to work. But it's what's happening.
Brian Reed: Yeah. I mean — that's an uncomfortable place to stop, but it might be the honest one. The mechanism didn't keep up.
Eliza Ward: It didn't. Thanks for working through it.
