Diet soda and stroke & dementia news coverage: 3 key points weren’t always reported

Michael Joyce produces multimedia for and tweets as @mlmjoyce.

The Framingham Heart Study group will turn 70 next year. That’s given them enough time to start studying the offspring of their original study cohort — which they have done since 1971 —  as well as know how to cautiously present results from their bread-and-butter: long-term observational studies.

soda dementia stroke

Which is what they do with a study released this week in the American Heart Association’s journal, Stroke.

The findings? Drinking at least one artificially sweetened beverage daily was associated with an increased risk of developing stroke or dementia compared to those who drank these beverages less than once a week.

The news release headline is accurate and cautious:

“Diet drinks and possible association with stroke and dementia; current science suggests need for more research”

And right out of the gates, in the second paragraph, they quickly highlight:

“The authors caution that the long-term observational study was not designed or able to prove cause and effect, and only shows a trend among one group of people.”

I asked Preeti Malani, MD, an associate editor at JAMA, and one of our editorial contributors, what she thinks of the press release.

“In general , I agree that the release is quite good and does try to highlight limitations of this observational study,” says Malani, a former journalist herself. “One potential issue is the presentation of relative risk (‘three times the risk’) at the top. Even if a risk is ‘three times higher’ the absolute risk is likely quite low. That’s one thing all press releases should include if those numbers are available. A quote from one of the authors does this, but you have to read down toward the end of the release.”

And here is that quote from the news release by co-author Matthew Pase, MD:

“In our study three percent of the people had a new stroke and five percent developed dementia, so we’re still talking about a small number of people developing either stroke or dementia.”

The news release also does well to highlight some of the limitations of the study, appropriate considerations when comparing sugar-sweetened and artificially sweetened drinks, as well as important related concerns with regard to obesity and diabetes.

How did reporters handle this?

It was inevitable that this study would get heavy news coverage, especially in the United States where we drink over 40 gallons of soda per person, per year. That makes covering this story carefully all the more vital.

But that’s not what happened.

Other than reporting the basic design and findings of this study, adequate reporting of the Framingham findings would need to make three caveats quite clear. First, that association does not imply causality. Second, absolute risk numbers need to be included to place the relative risk in context. Finally, the limitations of the study are not trivial and should be explained.

Most coverage made it clear that the association between artificially sweetened drinks and stroke and dementia did not imply causality. That this was well reported may be attributable to not just the carefully written press release, but also to the authors who clearly made an effort of bringing this up in most interviews.

But very few stories – the USA Today and WebMD are two examples — provided the absolute risk numbers. It may be that relative risk makes for better headlines while absolute numbers have less punch.

Likewise, the limitations of this study — which were very clearly laid out out in both the published study and the news release — were rarely mentioned. It wouldn’t have taken much extra effort for reporters to simply mention that most Framingham studies are hard to generalize to the public at large because, not only are they observational, but because they under-represent ethnic minorities. Also, this study relied upon a food questionnaire that required participants to look back quite a few years. Recall bias is a major concern.

I only found one reporter who covered all three caveats. That was Alexandria Bachert in her article in MedPage Today. This is hardly a mainstream news outlet … but I wish it were.

You might also like

Comments (4)

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

Gary Schwitzer

April 21, 2017 at 11:42 am

Perhaps the worst story we saw on this was from Yahoo News, headlined: “Just one diet fizzy drink a day can triple your chances of a stroke or dementia, researchers suggest.”

Gary Schwitzer

Jim Lukaszewski

April 24, 2017 at 11:36 am

My, my such a gentile reasoned and toothless commentary. The Minneapolis Tribune trumpeted an erroneous front-page headline: Diet soda linked to dementia. Like so many even major news mediums there is little effort or attempt to read and truly understand the release. This is an important and tragic error by this newspaper that cannot be easily retracted or just walked back.

But the real question is why was this story even a story? It has more caveats than any of President Trump’s prenuptial agreements. All those caveats say this is not news. What was the motivation for putting this reputation puffer out? It’s about the inconclusive results of 70 years of observation followed by some torturous data manipulation. Is that science? Seems to me if they follow their own advice the science may be about to begin. Probably should have waited with the confusing and vacuous announcement until something substantive could be reported.

Andrew DePristo

April 25, 2017 at 7:18 am

I have read the original article, which makes rather cautious claims. However, I see no evidence that a major confounding variable was treated adequately. The effect of blood pressure variation on stroke has a hazard ration of 5 to >30 depending upon exactly how high the BP ( The correction for this would have to be done for each individual, not for a group as was done in their Model 3. And, the correction might have to be done as a function of time since BP varies. I don’t believe that anyone knows how to do the latter. And, groups with the same average BP can have very different distributions. IMHO, this article presents unsubstantiated evidence and makes no sense biologically.

Paul Chacho Jr.

May 20, 2017 at 9:25 pm

My wife told me about this study. I Googled it and right away became suspicious. All the outlets leading the parade of this misleading study were the MSM Fake News Kings. Thank you for the honesty found on this page.