Weekend Reads: Is Drinking Coffee Good or Bad For You?

by Kevin Schofield

This weekend's read is a research study on the effects of drinking coffee — potentially a lot of coffee — on your health. Specifically it looks at whether people who drink coffee were more likely to die over a ten-year period from any cause (called a "mortality hazard ratio").

The researchers from the National Cancer Institute and Northwestern University studied a sample of 490,000 persons from the United Kingdom whose medical records from 2006 to 2016 are in a research "biobank" along with data about their demographics and lifestyle. Biobanks like this one are essential resources for looking for connections with health and medical outcomes.

The results they found were surprising. According to the researchers, drinking coffee is inversely associated with mortality: that is, people who drank coffee were less likely to die over the ten-year study period. Even more surprising: even people who drank six or more cups of coffee per day — which, let's be honest, sounds scary — had a lower mortality rate than those who drank fewer cups.

The results were the same regardless of whether you drink instant coffee or ground coffee. They were also the same for people who drank decaf coffee; in fact, the researchers specifically looked at those people whose genetics showed that they were likely to metabolize caffeine faster, and that made no difference either. So while they weren't able to isolate what it is about coffee-drinking that might lower mortality, it clearly wasn't the caffeine.

The association between coffee-drinking and lower mortality was independent of sex, age, body-mass index (BMI), history of diabetes, and a previous diagnosis of cancer, heart attack, or stroke. The researchers looked at several other factors in the study participants' medical histories and seemed to find a few correlations, but the number of people in the study group with those conditions were so low that the correlation couldn't be trusted. That's measured with what statisticians called a "p-value": the probability that the data might look that way even if (in this case) there is no correlation between coffee drinking and mortality for that group. The lower the p-value, the more you can trust the result to be real. For example, the data shows that people who drink more coffee have a lower risk of dying from cardiovascular disease, with a p-value of .05: that means that there is a 5% chance that the same number of people would have died from cardiovascular disease if there really wasn't a correlation with coffee-drinking. On the other hand, the data also shows that coffee drinkers have a lower risk of dying from pancreatic cancer, but the p-value is .50 — there is a 50/50 chance that there really is no underlying correlation.

The researchers are very careful not to conclude that coffee is good for you or makes you live longer — or that you should drink more of it. They also didn't prove that drinking coffee causes any particular outcome, good or bad; they simply found a correlation, which could also be caused by some other factor such as lifestyle differences or socioeconomic status. Do people who drink more coffee have different jobs than those who don't (and perhaps are drinking it at work), or make more money to allow them to buy that much coffee? It may be that only certain people have the means and access to that much coffee — and the circumstances that grant them that means and access perhaps also give them better access to healthcare or a better overall diet.

But at the same time, this study is important for what it doesn't show: drinking coffee wasn't associated with a shortened life. Or as the researchers so carefully put it, "Coffee drinking can be part of a healthy diet." This study won't put to bed questions about whether coffee is ultimately good or bad for you, but it suggests that it won't be the thing that kills you.

Association of Coffee Drinking With Mortality by Genetic Variation in Caffeine Metabolism

📸 Featured Image: Photo by amenic181/Shutterstock.com