Endometrial cancer joins the “coffee club” in which association ≠ causation

Did the latest round of causal claims about a coffee observational study stem from a news release? I’m betting so, because not many journalists I know regularly read the journal Cancer Epidemiology, Biomakers & Prevention, which is where the latest coffee study appeared.

Indeed, that’s a journal published by the American Association for Cancer Research.  AACR also published a news release, “Coffee Intake May Lower Endometrial Cancer Risk.”  It wasn’t a bad news release, but it never mentioned anything about the limitations of drawing conclusions from observational studies.  It didn’t use any inappropriate causal language, instead referring to observations and associations.  But it also didn’t educate readers as well as it might have. News releases from journals and professional associations could do so much to help educate on such matters.  When they don’t, it’s lost opportunity.

It’s important research – don’t get me wrong.  But people are so consistently confused by media messages about observational data, we expect journals and professional associations to make an extra effort to educate on limitations.

And it makes a difference, when the news releases get picked up by journalists.

FoxNews.com posted a story (based on a Medical News Today story) that appears to be lifted directly from that AACR news release, quotes and all.  But it opens with the line:

“Researchers say that women who drink multiple cups of coffee per day may reduce their risk….”

Yes, they may.  But an observational study, based on women keeping a dietary questionnaire (which introduces all sorts of evidentiary questions), cannot prove that they may or may not.

A HealthDay story (and, therefore, a story by CBSNews.com and others who subscribe to HealthDay) had the subhead, “Women benefited form drinking about 4 cups daily.”  Benefit implies that cause-and-effect had been established.  It had not been. And it leads with one of those throwaway lines so often seen in stories about observational studies:

“Ladies, a heavy coffee habit might do more than perk you up. New research suggests it may also reduce your risk of endometrial cancer.”

To its credit, though, the HealthDay story later stated:

“The researchers found a link, but not a cause-and-effect relationship, between coffee drinking and lower risk of endometrial cancer.”

That’s what we’re looking for.  The first 12 words of that sentence are the money quote.  This is the second story in 2 weeks by HealthDay that included such a caveat. Kudos to them for this improvement.  And this one quoted another researching saying, “some studies published recently have not found a link, so he believes that ‘the jury is still out.’ ”

At mid-day today, I got about 100 returns on a search result looking for news of coffee and endometrial cancer risk.  Almost none that I saw made any mention of observational studies and their limits.

Coffee, as we’ve written before, is the poster child for abused stories on observational data.  Some of our past posts:

As always, we remind our readers that we offer a primer,

“Does The Language Fit The Evidence? – Association Versus Causation”

It’s been on our site for 8 years.  Some have learned from it.  Others have not.  We’ll keep trying to reach them.


Follow us on Twitter:



and on Facebook.

You might also like


Please note, comments are no longer published through this website. All previously made comments are still archived and available for viewing through select posts.

Comments are closed.