You may have heard about the widely reported study from the Harvard School of Public Health that shows that drinking coffee lowers one’s risk of developing diabetes. This would be nice, except it isn’t actually true. The full report published in the Annals of Internal Medicine concludes that “long-term coffee consumption is associated with a statistically significantly lower risk for type 2 diabetes.” This is not not the same thing as saying that caffeine consumption protects against diabetes.
This may seem like nitpicking, but it is a very important distinction. Quoting from the report again: “the design of this study does not permit the researchers to be certain that coffee decreases incidence of type 2 diabetes mellitus. It might be something else about coffee drinkers that protects them from diabetes“. In other words, don’t increase your coffee consumption purely to lower your risk of developing diabetes. There are much, much better ways to do this - like exercising. This is not to say that it isn’t a very interesting finding.
It would be nice if this was simply another example of the media making a hash of science reporting, but the misleading language actually originates with Harvard’s press release.
Years ago a friend forwarded me one of those internet “health warnings” about the links between antiperspirant and breast cancer. I did what I usually do with that kind of rubbish - spent about five minutes on the web and sent a reply to my friend highlighting the flaws in the “warning”, complete with links to authorative references and a few sites that debunked the hoax in detail.
My friend wasn’t the only person gulled by this hoax and some years back a study was released that showed no link between breast cancer and antiperspirant use. Not satisfied with that result, and possibly spurred on by the discovery of parabens (a class of chemicals widely used in deodorants) in breast tumours, a Dr Phillipa Darbre of the University of Reading has published a study that claims to show a link between breast cancer risk and a combination of shaving one’s armpits and deodorant use.
First off, parabens are not only found in deodorants. Pretty much any kind of goop that you put on your skin or in your hair is going to contain parabens. Secondly, the fact that parabens are found in breast tumours does not mean they caused those tumours. Breast tumours (and most other kinds of tumours) are awash with Dihydrogen Monoxide (DHMO), yet nobody is trying to suggest a causal relation between DHMO and cancer, let alone fund expensive studies.
But the real stupidity of all of this is the structure of the study. There is no way it could possibly have proved anything. The study examined 400+ breast cancer patients and grouped them according to axillary regime, finding those in the frequent shaver plus deodorant group contracted breast cancer at a much younger age.
This may seem pretty conclusive, but you have to consider that a far more likely alternative hypothesis also fits the facts. It seems reasonable to conjecture that younger women pursue a more intensive underarm regime than older women. If so, this would explain Dr Darbre’s findings perfectly. A control group of women without breast cancer would have helped settle this question, but the study did not include one. Also, age of onset is not necessarily a good epidemiological measure in any case - and particularly not in a study lacking suitable controls.
In this case, one suspects that the study’s protocol was deliberately flawed and Dr Darbre’s language certainly suggests someone has a barrow to push.