Unsure of what to believe when you hear of a new diet or supplement that will cause or cure disease? You're not alone!
With so many ideas and opinions on health and well-being available on social media, mainstream media and from well-meaning loved ones, it can be tricky to determine whether a claim is credible or not.
I absolutely appreciate many attempting to research and interpret evidence that is presented to them, but having the know-how, time and resources to do so properly is not always realistic. However, it is important to be cautious before spending hundreds of dollars on a new supplement, or drastically cutting out products from your diet based on the benefit or risk it may pose.
So, I wanted to give some basic guidelines to follow when you hear a new health claim, read a new headline or are provided with a study as “evidence”.
First off, let’s define a couple of terms I mention:
Risk of bias: Not only refers to personal bias as normally used, but also an inherent bias in the design or conduct of a study that may misrepresent the results. For example, the selection of participants in a trial was not random as they only sourced participants from a recent marathon- this group is more likely to be active and eat healthier than the general population and so, studying only this group can bias the results and be an inaccurate reflection of the general population’s disease risk.
Correlation and causation: Correlation is when things simply happen at the same time, causation is when one thing causes another to occur. Say researchers found that coffee drinkers were 10x more likely to get liver disease. However, when more studies were conducted, it was found that coffee drinkers were also more likely to be alcoholics. Alcoholism is more likely to lead to liver disease, meaning that it's more likely coffee drinking did not cause liver disease, they merely occurred together in the study.
Peer Review: This is where experts review a study before it is published to assess the validity of a study. It isn’t a perfect process, but further reduces errors being missed before publication. You want the studies to be peer reviewed for this reason.
This image shows a condensed version of what's known as the 'hierarchy of evidence'. It's used to represent the strength and reliability of evidence, considering things like risk of bias and causation. It can be useful to refer to generally and may also explain why health professionals seem to quickly dismiss claims of a “cure” or risk of disease.
Here are some things to consider when you’re next presented with a new claim:
Is it actually scientific evidence? Note what isn't on the hierarchy.
Stories from family friends, YouTube experts, random website articles etc. I think it’s useful to learn from each other and share experiences, it may even lead to future research questions, however, it is difficult to determine true causation on our own.
For example, when your Aunt has quit sugar and lost 5kgs, has more energy and stronger hair and nails, its very easy to blame sugar alone. Though, when you have a closer look, it really comes down to cutting down discretionary foods and replacing these with nutritious foods (including those naturally higher in sugars) and a diet higher in protein and iron- there wasn’t a need to follow a very restrictive diet after all.
Is the claim illogical or defies human physiology?
If so, cheer- you don’t need to worry about the remaining questions, save your time!
If you’re unsure whether it could defy human physiology or not, don’t fret. It’s good to note that this may be something you are not well versed in, and is a common way people are misled by claims. Just ask someone in the know!
Does the study actually study humans?
If not, a conclusion cannot be made regarding human health. We have different physiology to rats and tend to have a bit more going on than a group of isolated cells. These studies can useful in developing questions for future studies, but that's it.
Is the claim supported by a well-designed, peer reviewed Randomised Controlled Trial (second tier)?
This might look promising, but still requires more evidence. The results need to be replicated in different studies and supported by a body of evidence to rule out other contributing factors.
RCTs is where a group of people are randomly divided into two groups and given either an intervention e.g. a particular supplement daily or a placebo e.g. sugar tablet daily to determine if the supplement reduces risk of disease. Ideally these studies will be “double blind”, meaning both the researchers collecting the data and participants are “blind” to which group is which.
Typically, as single studies, these are more reliable in determining the cause of an effect than the studies in lower tiers but are not flawless. This is where you’ll often hear “looks promising…but more research is needed”.
This isn’t to say that the other tiers between animal/in vitro studies and RCTs are invalid. These studies look at the question in different ways e.g. may look backwards by comparing a group with an existing cancer and one without to assess variables like diet, race, exercise etc, or by looking forward by comparing and monitoring those on a vegan diet vs those consuming animal products for a particular outcome. It can be difficult to determine the causation and rule out bias and correlation though, as human are difficult to study and there are many variables.
Is the claim supported by Meta-Analyses and Systematic Reviews?
Yay! If the claim is supported by these, there is a good body of evidence behind it and it can be considered.
This type of evidence is where the single RCTs are assessed, and the results are synthesised to find the actual effect/outcome in question.
These studies aren’t perfect either and are still scrutinised, particularly when it comes to disease risk. One example of what is to be considered is the size of the effect- determining the absolute risk. If the study says you are 50% more likely to have a heart attack if you eat bread, this sounds scary. However, if the original risk of you having a heart attack is already very low, irrespective of your bread-eating preferences, then the overall and absolute risk may only be 0.1%- a little easier to stomach.
Is the claim expensive, dangerous or impractical?
If the claim is fully backed by a body of evidence and seems good to go, but is too expensive or has 1000 side effects, there is no point following it. No one food or supplement is necessary for health. In a similar way, if the claim is something harmless, cheap and accessible, like adding a pinch ginger to your dishes, sure, go for it. However, it can’t be attached to a claim of reducing or increasing your risk of disease until properly assessed.
The aim of this post isn’t to make you feel in capable of researching, but rather emphasise that nutrition research is tricky, humans are tricky and there is A LOT to consider. If you aren’t bothered to worry about the above, don’t panic- it’s what we do!
I hope this was helpful,