Coffee and tea sit in a strange spot in the wellness world. They’re among the first things people are told to cut when they want to live more cleanly, and they’re often treated as small vices to be managed rather than habits to be kept.

A large study published in JAMA in February 2026 complicates that picture. Researchers from Harvard T.H. Chan School of Public Health, Mass General Brigham and the Broad Institute found that people who regularly drank moderate amounts of caffeinated coffee or tea were less likely to develop dementia. Decaf showed no such link, and that difference is what makes the finding stand out.

This isn’t a case for drinking more. It’s a closer look at two everyday habits and at how the researchers reached their conclusion.

Vice 1: A Moderate Coffee Habit, Caffeine Included

The study drew on two of the longest-running health studies in the United States. The first was the Nurses’ Health Study, with 86,606 women followed from 1980. The second was the Health Professionals Follow-up Study, with 45,215 men followed from 1986. No participant had cancer, Parkinson’s disease or dementia at the start.

What makes this kind of research valuable is its patience. Picture a nurse in her mid-40s in 1980, filling out a detailed questionnaire about what she eats and drinks, then doing it again every two to four years for decades. Researchers call this a prospective cohort design. People report their habits before anyone knows who will get sick, so their answers can’t be colored by a later diagnosis.

Participants were followed for up to 43 years, and 11,033 developed dementia, identified through doctors’ diagnoses and death records. The researchers compared the quarter of people who drank the most caffeinated coffee with the quarter who drank the least. After accounting for other differences in health and lifestyle, the heavier coffee drinkers had about an 18% lower risk of dementia.

The pattern also showed up in how people felt about their own minds. Heavier coffee drinkers were slightly less likely to report feeling that their memory or thinking was slipping, which researchers call subjective cognitive decline. The figures were 7.8% for the heavier drinkers and 9.5% for the lighter ones. Heavier drinkers also scored slightly higher on a cognitive test given over the phone.

Vice 2: A Cup Or Two Of Tea A Day

Tea showed a similar pattern. The people who drank the most tea had about a 14% lower risk of dementia, and they also did slightly better on the thinking and memory measures.

The more interesting finding was about amount. More was not always better. The link was strongest at moderate levels: about 2 to 3 cups of caffeinated coffee or 1 to 2 cups of tea a day, which is roughly 300 milligrams of caffeine. Drinking more than that showed no clear extra advantage.

In practice, that points to an ordinary routine of the morning cup and the afternoon tea rather than a fourth espresso at 4 p.m . And because people reported their habits again and again over decades, the finding reflects long-term patterns, not what anyone drinks on a given day.

Why The ‘Vice’ Might Trump The Decaf Fix

The most revealing part of the research is what didn’t show a link. Most earlier studies on coffee and brain health didn’t separate regular coffee from decaf, so it was hard to tell what, if anything, was making a difference.

This team looked at them separately, and decaf drinkers showed no lower dementia risk and no better test scores. That makes decaf a useful comparison. Decaf keeps much of what coffee offers, including the ritual, the social setting and many of the same plant compounds, but not the caffeine. If the link disappears when the caffeine is removed, caffeine becomes the most likely explanation.

It’s like comparing two nearly identical recipes that differ by one ingredient. If only one turns out well, the missing ingredient is the first place to look. The comparison isn’t perfect, though. People who choose decaf may differ in other ways; some switch because of a health concern, for example.

What These 2 Vices Share, And What The Study Can’t Say

The likely common thread is caffeine, in moderate amounts, over many years. Participants were also, on average, about 46 (the women) and 54 (the men) when tracking began, so the habits studied were mostly midlife habits. That fits the idea that midlife may be an important time for long-term brain health, well before any symptoms appear.

Still, the researchers were careful, and you should be too. This type of study can show that two things tend to go together, but it can’t prove that one causes the other. Coffee drinkers may differ from non-drinkers in ways a questionnaire can’t capture.

The link may also run partly in reverse: someone in the very earliest, unnoticed stages of memory decline might drift away from their usual coffee routine, which would make non-drinkers look worse off than coffee alone would explain. The differences on the thinking tests were also small, and one overall measure of thinking ability fell just short of statistical significance.

What gives the finding weight is the size of the study, the decades of repeated check-ins and the telling contrast with decaf.

The practical takeaway is modest and reassuring. People who already enjoy coffee or tea and handle caffeine well have little reason to treat the habit as a vice they need to quit. The data points toward moderation. For anyone whose sleep, anxiety or health makes caffeine a poor fit, this finding doesn’t change that.

Your coffee ‘vice’ may be doing more for you than you realized. Find out how well the rest of your daily routine is taking care of you with this science-backed test: Self-Care Inventory