According to a new study from the HPLMGH Institute a five minute nap on Saturday afternoons increases your chances of breaking your toe while walking to work next week.
Sorry, this was a lie. But it is not much different from many other news items millions of people have to scroll through on their social media platforms. Who has not experienced the smug, slightly sarcastic, not-so-funny post that goes a little like: “I told you [someone tagged], that B people are smarter than A” and then they post some news article that references a study about how people who laze in bed are smarter than their morning equivalents. *
There are generally two categories of those articles. First is the “haha” one, an example of that one would be the “Ice-cream for breakfast makes you smarter” article. The second one is the “it is the opposite of what you thought” article like this Guardian article on how less cushioning in your running shoes is better for you. While these articles are not always based on highly scientific methods, there is often some research or a study behind them.
But there is a third category of news items that is also prone to reference ‘a study’. Unlike the “haha” articles which tend to cite some obscure quasi-scientific journals, the third type tends to cite studies from more respected journals and involve more relevant and important topics, from disease to income inequality and to love and happiness. These articles are often good and the newspaper format is indeed a great format for transferring information from academic journal articles to the public. Usually this is the case, but unfortunately sometimes the journalists seem to be more interested in the headline than the content.
Taking one or more pots of Snus per day increases your Diabetes risk by 70%
This was one of the main messages in a BBC article from yesterday. For those who do not know Snus, it is a non-smoking tobacco that people (mostly in Sweden) put under their lip for a nicotine hit. While there is a consensus amongst scientists and medical doctors that smoking is awful, there remains a debate about the evils of Snus. One group of health professionals argue that Snus is bad because it is addictive and there are some theoretical and empirical evidence of the harm it causes (from gum health to potential cancer causes). Then there is the group of health professionals who think it is a good alternative and a substitute for cigarettes (they argue it is the lesser of two evils).
The first group – al tobacco is bad – group should feel vindicated by the BBC article. A new piece of empirical evidence in their anti-tobacco arsenal. 70% more likely to get diabetes if you take Snus. There is however a small problem with this. This BBC article is just a “haha” article disguised as proper article, and here is why.
First, the article states a 70% increase in risk of getting diabetes. This most certainly sounds big to most people. But what is the significance of this big figure? I could not find figures for Sweden, but somewhere it said that the average American has a 11% chance of developing diabetes before she/he turns 70 years old. Therefore, if the average Swede has the same chances of developing Diabetes as the average American then, based on this headline, the average Swede who stuffs a can of Snus in his lip each day will have around 19% chance of developing Diabetes. Now for a ton (50g x 365days x 50years) of tobacco stuffed in a lip, 19% does not seem too high. A ton of sugar would probably not do you any better.
Second, who on earth stuffs a ton of tobacco in her lip before the age of 70? The answer, somewhere around 1.3% of Swedish humans. One or more can a day is most certainly excessive and based on the data in the study, more common consumption is between 1 and 4 cans a week. It looks like roughly 7% of Swedish people do precisely that. Since this is a much more substantial group that it would have made much more sense by the BBC to report on the increased risk of getting diabetes for that group, which according to the study, is only 8%. Ideally, BBC would have made a clear statement that would inform the reader that “the average Swedish person has a 11% chance of developing diabetes before 70, while most Snus using Swedes have a 12% chance”. But where is the thrill in that? And more importantly who is click and share that unexciting headline on Facebook?
Do Snus for a part of your life, quit and you will reduce the likelihood of developing diabetes
Thirdly, BBC should report more widely on the results from the study. Only reporting on a part of the results is misleading. If you were to download this article and read it (which is what I have done), what you would find is that people who used to do Snus and no longer do are 16% less likely to develop diabetes than people who never did Snus (for a short moment I and all the other former Snusers are now thinking “Scoooore”). But how can that be? It makes no sense if Snus leads to diabetes in the long-run that it is preventive if you do it only for a part of your life.
That must mean that there is something else than Snus that makes the quitters less likely to develop diabetes? It is pretty obvious that it is not the Snus intrinsic quality that makes it so that if you do it for like 10 years then stop that it becomes some sort of diabetes prevention magical drug. No, it is rather some characteristic (genetic or behaviour) of the people who have the will power to stop that makes them less likely to develop this disease. And if that is the case for those who stop, the same case can be made about those who do not stop. They could well have genetic or behavioural characteristics that cause diabetes.
Irrelevant, overstated and possibly not causal
If you take the three points made earlier together the point that this article is trying to make becomes a bit clearer. The headline stated a 70% increase in the chance of developing diabetes if people take Snus:
- The people that the article is talking about only represent around 1.3% of the Swedish population;
- Since the chance of getting diabetes before turning 70 is around 11%, there is only a 12% chance that the majority of Snusers (1-4 cans a week Snusers) get diabetes; and
- The study is most likely missing some crucial variable that has the potential of overstating the impact of Snus on diabetes.
There are other issues that is possible to pick from the study (inconstancy of results across the pooled studies for example, control variables, etc.). But the problem again is not the study itself. It is done using as good science as possible and indeed the authors are careful not to make any grand statements. The problem also is not necessarily BBC journalism. At least they are trying to tell real news from real scientific research. But they live in a click-and-count world where the market success metric is not quality of neutral news reporting but likes and views. And for some reason, big exaggerated headlines win those while good hones reporting of facts does not. Finally, I have great respect for the BBC and usually they are good, but sometimes I worry if they are going the same way as everyone else.
Thanks for reading,
*This author is probably not entirely innocent of posting this kind of unfunny BS.