Alcohol and health risk: what a 2026 modelling study found

A modelling study published in June 2026 found no protective health effect of alcohol at any level of drinking, and estimated a mortality risk of 1 in 25 at 14 drinks a week. The same study found that drinking up to 7 drinks a week was associated with minimally elevated risk for most conditions. It is an estimate built from published evidence, not a trial of drinkers.

What did the 2026 alcohol study find?

The paper, in the Journal of Studies on Alcohol and Drugs (9 June 2026), was led by Kevin Shield of the University of Toronto and Katherine Keyes of Columbia University, with co-authors. As Columbia’s summary describes it, the team reviewed more than 7,200 scientific articles on alcohol-related diseases, then applied the risk levels identified to national US health datasets and used statistical modelling to estimate long-term health outcomes at different drinking levels.

Headline results:

• Mortality risk of 1 in 25 for people drinking 14 drinks a week.
• Up to 7 drinks a week was associated with minimally elevated risks for most conditions.
• Increased risk above 1 drink a day for both sexes, and no protective effect at any level of consumption.
• Conditions covered included cancers (oesophageal, oral, breast), cardiovascular disease, liver disease and injury.

How should the results be read?

With three points of context. First, this is modelling: the estimates depend on the assumptions and datasets used, and researchers disagree about light drinking. Second, the authors say individual risk depends on other factors, including lifestyle, genetics and drinking patterns, and that emerging research on additional conditions such as pancreatic cancer needs further investigation. Third, the study was funded by Synergy Enterprises Inc. and the US Substance Abuse and Mental Health Services Administration.

The US Dietary Guidelines for 2025 to 2030, released in January 2026, say only to “consume less alcohol for better health”, without giving specific limits, according to Harvard’s Nutrition Source. The study’s own press release, headlined results not included in dietary guidelines, points to the same gap.

Denominators matter: “1 in 25” is 4 in 100, for one specific level of weekly drinking. Read as a personal forecast, it overstates what a population model can say.

The bottom line

• A June 2026 modelling study of 7,200+ articles found no protective effect of alcohol at any level and a mortality risk of 1 in 25 at 14 drinks a week.
• Up to 7 drinks a week was associated with minimally elevated risk for most conditions; risk rose above 1 drink a day for both sexes.
• The results are modelled estimates, and the authors stress that individual risk depends on lifestyle, genetics and drinking patterns.