This week The BMJ sent journalists a news release, “Regular consumption of spicy foods linked to lower risk of death.” The second paragraph – the third sentence overall – of the news release read: “This is an observational study so no definitive conclusions can be drawn about cause and effect, but the authors call for more research that may “lead to updated dietary recommendations and development of functional foods.”
If you go to the journal article on which the news release is based, you see that the seeds of appropriate explanation were planted further upstream.
In the conclusion paragraph of the published study manuscript, the researchers wrote:
“given the observational nature of this study, it is not possible to make a causal inference.”
Did that clarity – that emphasis on the fact that association ≠ causation – make a difference in subsequent news stories based on the study or on the news release? It appears that may be the case in this instance.
- TIME.com included this: “More research is needed to make any causal case for the protective effects of chili—this does not prove that the spicy foods were the reason for the health outcomes.”
- The New York Times Well blog had a line: “The authors drew no conclusions about cause and effect.”
- With even greater emphasis, the Los Angeles Times reported: “Although the study included nearly half a million volunteers who were tracked for a total of 3.5 million person-years, the researchers emphasized that they couldn’t show a causal relationship between eating spicy foods and living longer.”
- In The Washington Post: “The researchers said that while it isn’t possible to draw any conclusions about whether eating spicy foods causes you live longer from their work that more studies are needed to look at this link in more depth.”
- From HealthDay: “However, the study authors cautioned that their investigation was not able to draw a direct cause-and-effect link between the consumption of spicy foods and lower mortality. They could only find an association between these factors.”
- CBSNews.com stated: “The authors emphasize that this is an observational study so no definitive cause and effect relationship can be drawn.”
We’ve had a long-running challenge to news release writers for The BMJ and for news releases for others of the ~50 journals that BMJ publishes, to consistently state the limitations of observational studies that they write about. And we’ve brought this up with other journals as well.
</});
In this latest chapter, kudos to The BMJ. The words matter. What the researcher-authors submit matters. The journal’s editorial scrutiny matters. The accuracy of the news releases matters as well.