It does not seem to take long before someone in a conversation mentions that they are “Banting”, or in other words, partaking a diet consisting mostly of meat and fat on the advice of one William Banting (circa 1863) and bolstered by overzealous diet gurus who employ after the fact fatsplaining based on what our ancestors supposedly ate. Nevermind the fact that it would be difficult to know exactly what the people in the Palaeolithic era ate, proponents of this diet (like the semi-plagiarised Atkins diet before it) seem to be spreading misinformation so fast that their personas are having to play catch up. I might be accused of being a carbohydrate lover to those who seem to have taken a detour off the information superhighway, but without delving into ad hominem attacks, the topic of whether the diet is better than what is recommended by the medical establishment is quite well-covered here.
Of course, there is the problem of study fatigue, whereby different studies are thrown at opponents without consideration of their context or how well they were conducted, which leads to ample fodder with little substance. This approach tends to sow confusion which makes the scientific process look like a hyperactive rabbit attempting to organise papers whilst simultaneously verifying if these papers are true (which it kind of is anyway). The point is, it is difficult to know what we should be eating which makes it easier for soothsayers to peddle their quackery as being based on legitimate evidence. Let’s face it: conducting a good study and getting good results is hard, even for less tricky subjects. But while there have been good studies on diet conducted, it is important to keep these points in mind:
1. It is hard to tell people what to eat for an extended period of time.
There are the ethical questions to consider for anyone willing to participate in a study which examines the effect of what you put in your mouth. You can’t really dictate what a person should eat which means that you either are left with an observational study of what people would have eaten without the prying eyes of the researcher, or you have a group of people who want to lose mass (or have some other health problem) and so you randomly assign them to a diet and see how they get on. Either choice is fraught with statistical peril; an observational study cannot infer causality, whilst the conclusions of the latter randomised controlled design study cannot be generalised to people not currently load shedding (or whatever group examined).
2. Researchers are not God.
Which means they are neither omniscient nor omnipotent. They cannot compel someone to complete the study if they do not want to, and cannot know if someone ate what they were supposed to. They may pray that enough people in their study followed the guidelines sufficiently to get some statistical significance going on (and even that has its troubles).
3. It is difficult to know how clinical measures at the end of the study will translate into clinical outcomes.
At the end the researcher may have some measurements indicating general health e.g low-density lipid (LDL) levels, high-density lipid (HDL) levels, kilograms lost, etc. But how does any differences in those numbers fair in ten, twenty, or thirty or more years’ time in terms of health outcomes? Sure, there are longitudinal studies but those can only infer a correlation between an activity and a health outcome.
4. Studies must usually be BIG.
The only way to account for many of the variables affecting health (amounts of exercise undertake, fitness level, genetics, usual diet, etc) is to get a lot of participants so that you can detect differences between groups given different diets. Not always practical, especially when organisations consider meta-analysis to be a good substitute.
And that’s not to mention problems of lack of a control group, confirmation bias, participants conforming to researcher’s expectations, the fact that not every study conducted is published especially if it has a negative result, overreaching conclusions not supported by the evidence, using multiple statistical tests in order to find one which gives a significant result, the list goes on.
So the next time someone berates you for loving your sandwiches more than their antibiotic infested muscle feast, just smile and remember: the truth is invariably more complicated and life is too short to constantly debate about eating the odd bit of cake.