Wine And Chocolate Aren’t Health Foods
This is not a complicated fact
In the world of “things that we wish were true”, where unicorns frolic with antelopes and the globe is not warming towards an inevitable disaster, there is one thing that just keeps coming up.
We really, deeply want wine and chocolate to be healthy, everyday foods. Despite everything we know about nutrition. Despite everything we’ve learned about the human body, and alcohol.
I can see why: they are delicious. Wouldn’t it be so great if they were good for us too?
Recently, the world has been shaken by a study that media organisations say proves that these unhealthy foods are, in fact, good for your health. “Wine and Chocolate Can Help You Live Longer”, said the New York Post, while Yahoo! went with the even better headline: “Science says drinking wine and beer and eating chocolate could help you live longer”.
I mean, if science says that they could help you live longer, it’s gotta be true.
Sadly, these headlines are wrong.
Wine and chocolate are delicious, but remain steadfastly bad for your health.
What The Science Says
The new study that everyone is talking about was a large epidemiological study looking at how anti-inflammatory food eating might impact people’s risk of death from cancer or heart disease. It was quite an interesting piece of work, particularly because the researchers looked specifically at smoking and how these foods might impact smokers’ risks.
The study was extremely big — the researchers looked at over 60,000 people over a long period of time, collecting a total of more than 1 million person-years* worth of data. All of these people answered a food-frequency questionnaire, from which the researchers calculated a score called an Anti-Inflammatory Diet Index.
Overall, it appeared that people who ate diets high in anti-inflammatory foods — and low in pro-inflammatory foods — were less likely to die from either cancer or heart disease.
Which brings us to the first enormous problem: what are anti-inflammatory and pro-inflammatory foods?
The Boring Truth
Here is the list of food intakes that the researchers included in their calculation of anti-inflammatory diet:
Fruits and vegetables, wholegrain bread, breakfast cereal, unprocessed meat, processed meat, teas, coffee, soft drinks, low-fat cheese, offal, chips, chocolate, nuts, wine, beer.
(The bolded categories were pro-inflammatory, normal text was anti-inflammatory.)
It’s easy to see that all of the headlines were just boringly wrong. This was not a study on wine, or chocolate, or even anti-inflammatory foods. This study looked at whether diets that were anti-inflammatory as a whole had an impact on health.
Sadly, that’s not the only issue here.
The media stories all reported the relative risk difference between the group that had the worst diet against the group that had the best: people who had a higher anti-inflammatory score were 18% less likely to die for any reason.
However, relative risk is not always very useful. I’ve written about this before: if an event is uncommon, like death, then relative risk can be very misleading. This is because relative risk is the ratio of one risk to another — if something happens 0.003% of the time, and the relative risk is 300%, it’ll only happen 0.009% of the time instead. This can be important to the population, but rarely to the individual.
In this study, the relative risk worked out to 82% — i.e. 18% less likely. However, if you crunch the numbers, the absolute risk difference was much lower, about 0.01%. So while you could say that people who had higher anti-inflammatory scores were 18% less likely to die, you could also say that their absolute risk only decreased by 0.01%.
Not very impressive after all.
There were other issues with the media’s interpretation of the study. This research was observational, so it’s hard to tell whether the lower rate of death in the high anti-inflammatory group was caused by the food or some other factor (for example, those people being wealthier). The study also didn’t take into account supplementation, which is considered to be an important component of anti-inflammatory food intake and may have confounded the results. The population that the researchers looked at were mostly Swedish people over the age of 55, making the results hard to generalize to people not in that age/ethnic group. The differences that everyone reported were also between the highest and lowest groups — eating just a little more anti-inflammatory/less pro-inflammatory foods might make much less difference to health.
Which anyone who had bothered to read the study could’ve told you.
What Does This Mean To You?
Ultimately, these results mean very little to the individual. Not only did this study find only a tiny risk difference, anti-inflammatory dietary advice is mostly just commonsense recommendations: eat less processed meat, more vegetables, more wholegrains, and cut back on highly-processed foods. Chocolate and wine are, on the whole, a bit tangential to the story.
It turns out that large epidemiological trials, while fascinating to people like me, are almost entirely useless to the average person.
If you are looking at 20 million people, a risk difference of 0.01% means quite a lot of deaths, but for most individuals that’s a completely meaningless risk. If you’re interested in improving your diet, go and see a dietitian. If you are worried about inflammation, talk to a doctor.
Just don’t read epidemiological studies. And definitely don’t listen to the headlines.
They are probably wrong about what’s best for you.
*Note: a person year is essentially one year of study participation by one person. So, if you have 10 people who stay in the study for 5 years each, you have 50 person-years.