In honor of graduating with my masters degree in public health today (yay!), I thought it would be appropriate to write about a related (and relevant) health topic.
The world these days is full of messages, what with all the venues for media and marketing. In terms of health, this means that messages being sent are, in effect, telling the public what they should and should not do. Saying, for example, that “smoking causes cancer” is basically telling us that we should not smoke, despite the fact that this is not explicitly written in the headline.
A lot of attention recently has been given to coffee. Here are a few of the messages that have appeared in the media within the past few years.
1.) Coffee, in moderation, is good for overall health.
2.) Coffee drinkers have a lower risk of dying compared to those who do not drink coffee.
3.) Coffee may be linked to increased risk for cancer.
4.) Coffee may be linked to decreased risk for cancer.
5.) Coffee is not associated with risk for cancer.
So WHAT DO WE DO??? There is no answer.
We can only turn to understanding why these messages conflict in the hopes of resolving the behavioral issue.
First, there isn’t only one cancer. There are tons of different cancers with different pathologies and different risks. You can see from the article links I provided that many of these associations specify a cancer type. What may be good in fighting against one cancer may be bad for another. The problem is not only that there are so many cancer types, but also that coffee contains over 100 different potentially carcinogenic (cancer causing) compounds.
Second, coffee intake is very difficult to measure. The best way to assess the health effects of coffee would be to have a large group of very compliant, diverse individuals, half of whom would drink “x” cups of coffee (precisely measured) daily, while the other half would drink no coffee at all over a given period of time. But not only is this logistically difficult, it also comes with ethical issues; if coffee really is “bad for us,” then it is unethical to ask a group of individuals to consume it.
Because a study like this isn’t feasible, research has resorted to studies that measure coffee intake by self report. That is, they ask their participants to tell them how much coffee they drink. This is frequently asked in two different ways.
1.) Over the past year(s), on average how many cups of coffee do you drink daily?
First, a “cup” may mean different things to different people. Technically, one “cup” as measured by a coffee pot doesn’t actually fill an entire coffee mug. Thus, if you are counting one mug as one cup, your measurement is different than mine.
Second, what if I drank 10 cups of coffee a day during a few stressful months of the year, while drinking only 1-2 the rest of the year? I’d average this out to, say, 5 per day, but this doesn’t accurately reflect my daily intake. There also might be important differences in “binge drinking” coffee over periods of time.
Third, it’s difficult to remember behaviors that go far back in time. Perhaps one year ago, I drank gallons of coffee- but I haven’t had any coffee in so long that one year ago seems like two years ago in my head.
2.) In the last week (or 10 days, or month), on average how many cups of coffee did you drink daily? Or, in the last week, how many total cups of coffee did you drink?
What if I was sick in the week they asked me about and barely drank much of anything? What’s to say that last week was the same as any other week in my life?
You see the problem?
But these inaccuracies in measurement are extremely important. This is not to discredit the studies that have examined the risks associated with coffee intake. It does, however, shed some light on why the conflict remains a conflict.
Here’s the thing: when we consider smoking, we understand the evidence to be overwhelming in terms of its link to many types of cancer. Every study demonstrated similar results with enormous differences between smokers and non-smokers. In the case of coffee, we are looking at much smaller differences, with some more variation in results between studies.
In addition to biological understanding and measurement errors, confounding factors also become extremely important. Confounders are things that might distort an association. For example, smoking (which is highly linked to increased risk of cancer) might distort the relationship between coffee and cancer if it is not accounted for in analysis. How? Let’s say for arguments sake that smokers were more likely to drink coffee. Then we might see that coffee drinkers are more likely to get cancer when in reality it’s that smokers are more likely to get cancer, but they just also happen to be more likely to drink coffee. Other potential (and common) confounders are race/ethnicity, BMI, and gender, to name a few. There is a way for studies to account for these other factors, but not all of them do.
What does all of this mean? You can’t believe everything you read! There are a lot of mixed messages out there, and it can be very difficult to tease them apart. The important thing is to think for yourself and always challenge the information you are given.
Happy coffee (non)drinking!