Study linking sleep time and heart disease risk shows correlation not causation

“Going to bed between 10 and 11pm every night cuts your risk of getting heart disease by up to 25%, study claims”

The Daily Mail has reported on a new study and in the headline claimed it shows going to bed between 10pm and 11pm can cut your risk of heart disease. It then repeated that claim in the first line of the story, saying that the study showed going to bed at that time “slashes” the risk.

The study was reported widely elsewhere as well, including by the BBC, ITV News and Sky News. But in their respective articles, they said going to bed at that time is “linked” to or “associated with” a lower risk of heart disease.

The distinction is important. 

The study found a correlation between going to bed between 10pm and 11pm and a lower risk of heart disease. But it couldn’t prove one caused the other. 

The Mail’s article does go on to use the term “link” in the subhead and second paragraph, but doesn’t clarify this link is non-causal (contrary to the claim in the headline) until the 20th paragraph, when the lead author is quoted as saying “the findings do not show causality”.

Unfortunately, this is a fairly common problem. We often see media reports claim studies have demonstrated causation when they only find correlation and it’s possible other factors have affected the findings. For example, last year we saw similar issues in the reporting of a study which found a link between the number of sexual partners women had and their cancer risk, but not causation. 

Going back to the study at hand, in order to isolate the possible effect of sleep time on cardiovascular disease risk, the researchers did adjust for several factors known to increase that risk including hypertension, body mass index, smoking and diabetes. Due to the participants selected for the study, the authors added that the findings “may not generalize well to other populations”.

The study says: “Although the findings of this article do not show causality, they mandate further research into sleep timing as an independent cardiac risk factor.”