Nutrition studies tend to go like this. Tomato causes cancer! Tomato cures cancer! Don’t eat more than one egg a day. Three eggs a day lower cholesterol. Red wine will extend your longevity – or not. Is coffee good for you? Maybe.
I think you get the point i made. We live in a world where we are flooded with information and advice based on nutrition studies. So how do you know which ones to trust, when you can get conflicting information on a daily basis?
NUTRITION STUDIES – GOLDEN STANDARD
There are certain criteria that must be meet for a study to be reliable. Putting on a white lab coat and designing charts isn’t enough.
For a study to be reliable it must be controlled, double-blind and randomized. If one of these criteria isn’t met, and there isn’t a good reason for it, then the results are to be questioned.
So let’s check the terminology and what does it mean when it come to nutrition studies and other.
This means that both the researcher, and the participating subject are unaware in which group the latter belongs. Be it in the control group or the one given.
There are two reasons for this. The first reason is the potential placebo effect if the subject knew he was receiving the medicine instead of the placebo. On the other hand the researcher could affect the results of the study with conscious or subconscious bias.
This means there are two groups. One which receives the medicine and one which receives the placebo (usually sugar pills).
Controlled studies are very important as they can show us if the medicine or the supplements tested is more effective, compared to the subject receiving a placebo.
This means that the subjects were picked at random to go in one of the groups. The experiment or the control one. Now this is very important. If the people can choose at hand, which group they prefer, they may show certain bias that could taint the results of the study.
A hypothetical example.
Maybe people who are more optimistic at nature, would rather try a new antianxiety medicine, compared to the old one given to the controlled group.
Or let’s say we have an athlete who would rather be in a group which will be given a new supplement that increases stamina and strength.
The researchers might wrongly conclude that the supplements was the reason for better results.
Similar problems might present themselves if the researchers are the ones who can choose who goes in which group.
A nice example are Wilsons HMB studies, such a fiasco.
CONFLICT OF INTEREST
Well we can’t go past the good old conflict of interest. If the aim of the study is to prove that sodas are healthy and the financier of the study is Coca-Cola, you should be suspicious.
That doesn’t mean that you should write the study off immediately, but it’s perfectly clear what the financier motives are and who in the right mind would want to finance a study that shows their products in a bad light.
And for the record, this isn’t a dig at Coke, I love Coke Zero. It’s a drink sent from heaven by the dieting gods!
But in reality there is a tendency to manipulate results to show what we want to show. Just look at the latest Wilson HMB study as the perfect example. That results are beyond believable and he has gotten major flack for it.
In medicine for example at least 30% of studies are financed by pharmaceutical companies. And this is a number where they were financed directly.
Who knows how many studies there are, where there are other type of conflict of interests. Does the researcher own stock in the company? Does a family member work there? Chances are the percentages would be even higher.
LENGTH OF THE STUDY
The length of the study also comes into question. Depending on the goal and the research of the study, some can easily be short-term. Then there are studies, which have to be long-term that could last years or decades before we can get a realistic picture.
For example, the Nurses Health Study, that began in 1976. It has been going on for 40 years and is now in its 3rd generation. It is focused on finding risk factors for cancers and other chronic diseases that plague woman.
CORRELATION DOESN’T MEAN CAUSATION
This is a popular phrase in statistics that emphasizes that despite correlation, one variable doesn’t influence the other. Correlation can be positive or negative, but it still doesn’t imply causation.
One of my favorite examples for this is shark attacks and ice cream sales. Both goes up during the summer months, which means there will be a strong correlation between them. But causation? I’d really like to listen to an argument, where ice creams sales influence shark attacks and the opposite.
Another important thing, is the reproducibility of a study. This means, this means that if I, you, Bob, or Sally do the study, we have to come to the same results as the original researcher. The study has to provide the same results over and over, as long as we follow the same methods and protocol.
NUTRITIONAL STUDIES – THE CONCLUSION
But the most important thing we have to remember, that nutrition studies don’t prove anything. Yes they DON’T PROVE anything. They merely suggest certain theories and mechanism. And even that only when you put them through a filter of conditions it has to meet first for its reliability.
Moderation. Yeah, who would have guess I recommend common sense. Instead of jumping from one extreme where you eat one egg a week or another where you eat 10 eggs a day, because one week eggs are healthy and the other week they are the devil.
You should eat everything, especially unprocessed food, but nothing too much, regardless of just how healthy it is in the latest study.