From the Editor
Have you ever been depressed?
If so (or even if not), you probably know of the somewhat simplistic “Low serotonin theory of depression” and how it’s increasingly fallen out of favor. And you may also know that trials of SSRIs have been criticized for publication bias. So in a nutshell, while some people are definitely helped by SSRIs (and even pulled away from the brink of very dark situations), both the science and evidence is weaker than was thought a couple decades ago.
Nutrition theories have similar issues. The brain is so complex that not even the most well-trained cognition researcher knows what the hell is really happening. But the rest of the body is also complex. The difference is that few people pretend to know exactly how the brain works, while everyone and their mom (and brother, and grandma) has an opinion on nutrition issues.
Not that these opinions hold no water. People do know how they acutely feel when eating different foods. Tummy upset, energy levels, and whatnot. People don’t, however, know how different organs and tissues are impacted by the foods they eat, and it’s quite rare to do strict personal experiments on foods and nutrients and then get some kind of testing done.
So in the absence of obvious health effects from specific foods and nutrients, coupled with too many studies for any one researcher to read, there have been many competing nutrition hypotheses throughout the years. And many of these hypotheses have fallen. Dietary cholesterol and heart disease, the evils of even moderately high salt intake, and so on.
But that’s how science goes. It’s not the changing of hypotheses that’s holding up progress (as many lay people would think), rather it’s the behind the scenes action. If a big study showed something, were there other ones that showed the opposite but didn’t get much attention? Who funded which study, and might that have impacted the study design?
And now back to depression.
The truth is that (don’t you hate it when people say “the truth is”? as if they know some truth that you had never considered) these theories are made by humans to explain complex phenomena that are deeply, deeply interrelated. Babies aren’t usually born depressed, or at least I suspect not. They shoot out, get cooed over, and eventually get fed breast milk or formula. The path from there becomes way less dichotomous. Stress, relationships, medications, germs, and so on -- they all go in discrete buckets to us humans, but that doesn’t mean germs in your gut can’t be part of the reason you’re stressed or depressed.
That’s why interdisciplinary research is so crazy important. If you’re a nutrition geek and only read about bodybuilding nutrition, you’re not just missing out on some really fascinating stuff, you might be hurting your health. The brain may be complex, the gut might only be in nascent stages of research, and the gut-brain axis even more so. But that’s part of the fun. If you like research, the era of internet-enabled health exploration is quite an exciting time to live in. Drink deep, and drink well.
Editor-in-chief, Nutrition Examination Research Digest
There are three types of nutrition researchers in the world: those who enjoy biostatistics, those who tolerate biostatistics, and those who live in ignorant bliss of biostatistics.
It’s not the sexiest topic around, but when researchers at least tolerate biostatistics, studies they conduct vastly improve.
When one embarks on a new study, there’s a decent chance that a biostatistician will be pulled along for the ride. That wording was chosen carefully, because the relationship between the statistician and the nutrition researcher isn’t always a pleasant one. The statistician’s job is to make sure the biostatistical methods are sound, while the nutrition researcher is honing in on a specific health topic, and often hoping to build a career on that or similar topics.
So, when do these aims start to cause turmoil? Let’s take the example of a meta-analysis, meaning a quantitative pooling of data from multiple studies. Since we can access all those thousands (well, millions, actually) of studies on pubmed, we might as well combine data instead of siloing ourselves in individual studies, right? We analyze meta-analyses every once in a while for NERD because they’re especially helpful for piecing together topics that have large amounts of data.
Well, it turns out that a decent chunk of biostatisticians don’t even like the idea of meta-analyses.
Their reasons vary, but there are some major and quite valid arguments. Meta-analyses are not primary research, but they can contain some primary research studies (RCTs) that are very well performed. Therein lies the rub: a meta-analysis can group one or two great studies in with some mediocre studies, and even some pretty terrible ones. When studies are combined, their individual granularity is lost.
To top it off, meta-analysis is sometimes considered the platinum standard, even trumping a single well-done RCT. That’s despite inherent weaknesses in meta-analysing heterogeneous studies, plus the always-lurking issue of potentially missed studies due to publication bias.
That being said, there isn’t some expected level of animosity between nutrition researchers and biostatisticians. It’s usually a-okay. But nutrition is a fickle beast, and studying the impact of single nutrients or complex diets on varying diseases can be quite a statistical chore. It’s a pleasure to behold the rare nutrition researcher who likes biostats a lot, for they are as harsh when criticizing their own work as they are when reviewing other people’s.
Nutrition science, nutritional epidemiology, and biostatistics: they work better in concert than when siloed.
Editor-in-chief, Nutrition Examination Research Digest