Coffee Art 14

Women, Depression, & Caffeine: Turning a Critical Eye on Research Reports

A recent study published in the Archives of Internal Medicine that describes the relationship between women’s caffeine consumption (primarily through coffee) and risk for developing depression has been all the buzz (pun intended) over the past few days. The story was picked up by numerous media outlets, including the New York Times, NPR, and Time, with catchy titles like “Caffeinated Women May Be Fighting Depression with Every Cup.”

The articles highlight the study’s findings that women who drank two to three cups of coffee a day reduced their risk of developing depression by 15 percent as compared with women who only consumed a cup or less a week. Those who drank four cups or more per day reduced their risk of depression by 20 percent.

If you’re a woman and a coffee drinker, upon hearing of the research results, you’re likely to feel good about your daily habit. If you don’t currently consume coffee, it’s unlikely that you’re going to rush out to purchase an espresso maker, but maybe you’ll file the information away as interesting.

What you probably won’t do is take a look at the study in question to determine what it’s really saying. And of course, you’re not alone; it’s impossible to background check every story published in the popular media. What’s unfortunate is that reporters who pick up these kinds of stories often do readers the disservice of presenting research results in a way that they are not meant to be presented.

Unfortunately, I can’t access the study’s full text, but reading the abstract and doing some background reading about the study brings to light some questions about the results that were overlooked in all of the various journalists’ and bloggers’ takes on the study I’ve encountered. Let’s take a look at some of the context for the purported link between caffeine consumption and reduced risk of depression for women.

The Archives of Internal Medicine study drew from a larger, ongoing study, The Nurses’ Health Study, which was started in 1976. There were 50,739 women (nurses), whose mean age was 63, enrolled in the coffee study. Right off the bat, I notice that this seems like a skewed sample and I wonder about the wisdom in generalizing the results to younger women.

At the outset of the study, all of the women were “free of depressive symptoms at baseline.” In other words, none of them reported experiencing depression in 1996, when the caffeine study started. There was no information in the abstract about the women’s prior history of depression, but I would be curious about that as well. Over 10 years of follow-up, 2607 of the women were identified as having developed depression, which was measured by the nurses’ self-report of having been diagnosed as depressed by a physician and using antidepressant medication.

Every study needs to determine how to measure the things they are trying to research. I would argue that defining depression as occurring only when women were diagnosed and taking antidepressants is rather limiting and likely underestimates the true incidence of depression among participants. Not everyone who experiences depression seeks treatment, and not everyone who seeks treatment receives a prescription for medication.

All in all, it’s not a terrible definition, though; it makes tracking the incidence of depression an unambiguous process. What stands out to me, however, is that only about 5% of the women developed depression during the time period of the study. Even taking the probable underestimation of depression incidence based on the way the researchers defined it, this is a surprisingly low incidence of depression.

Most studies estimate the lifetime prevalence of depression in women to be around 20%. It’s not valid to compare lifetime prevalence with the percentage of women in the study who were diagnosed with depression, but I would guess that the general 10-year prevalence rate would still be greater than 5%. I’m curious why, caffeine consumption notwithstanding, this particular cohort of women experienced such a low rate of depression.

That brings me to the fundamental misstep made by journalists who claim that drinking a lot of coffee may reduce the risk of depression for women. Those kinds of general statements, which are peppered throughout the articles I read, do not take into account the particulars of the study, like the mean age of the participants and the extremely low prevalence of depression in the study. They also place an implied causal spin on correlational survey research.

What does that mean? Just because there is an association between two things, like coffee consumption and reduced risk for depression, it doesn’t necessarily follow that it was the coffee consumption that directly caused the lowered risk.

Does drinking coffee prevent depression in women? I don’t know. And neither do the researchers. All they know is that, in the sample they studied, there was a relationship between those two variables. The reason behind that association is anybody’s guess.

I wish that news outlets would look a little more critically at the research results they popularize. Also, I’d like a pony. Not going to happen, you say? Bummer.

(1) Comment

  1. Sue
    SueSep 29, 2011

    I agree with you. The journalists did not look at the study critically.

Leave a Reply