Media headlines are notoriously sensationalist when it comes to health and nutrition news - they grab your attention, but they often base their content on very flawed research or completely misinterpret the results. Just Googling a simple nutritional question such as whether we should be drinking coffee, you often find yourself with very conflicting messages, headline media reports, and online ‘experts’ with a vast array of different opinions.
Many of us will then leave it at that – believing that the jury must be out on that particular question. Or worse still, we can often take some of these headlines too literally and change our habits to damaging effect. But don’t you sometimes just want to sift through the rubbish and find out what evidence we should actually believe? Well, the main thing to realise is that there is a kind of hierarchy of scientific evidence, which should dictate how much notice we give any particular article / headline.
1. Youtube videos, blogs, personal anecdotes – obviously anyone could say anything without being held accountable so this is a very unreliable source of information
2. Expert opinion – someone with the title of Dr or Prof writing an editorial or letter without the specific research to back it. Obviously experts are in a better position than most to comment, but unless there is good research to support their statements, they could easily be mistaken or their opinion could be out-dated
3. In vitro studies – these are studies that use isolated molecules, cells, or tissues in a test tube or petri dish to show how a mechanism might work. This can allow the components being studied to be observed more closely and conveniently than if they were in vivo (in humans or animals). However, this does not replicate the same conditions as if it were in vivo, so there is no telling that the same results would occur in the body.
4. Cross-sectional studies – this is a type of observational research that looks at data collected from a population (or a sample population) at a specific point in time. They can show if there is an association between two factors, but it’s impossible to say what is causing that association. For example, cross-sectional data may show an association between caffeine consumption and insomnia. It would be tempting to conclude that caffeine therefore causes insomnia, but can it rule out that maybe people who already suffer from insomnia simply tend to drink more caffeine to get them through the day?
5. Case control studies – this is another type of observational study, but where researchers compare patients with a disease or outcome of interest (cases) with patients without the disease or outcome (controls). They look back retrospectively to compare how frequently the exposure to a risk factor is present in each group to determine the relationship between the risk factor and the outcome.
6. Cohort studies – Working the other way round now, researchers observe a group of people without the disease or outcome of interest, to see who develops the disease or outcome over time. Using the caffeine example, none of the participants would have insomnia to start, then the researchers would observe who developed insomnia over time and whether this was associated with consuming caffeine. This can therefore give us a better idea of causality.
7. Randomised controlled studies - Now the research is getting very high quality because bias is minimised. Participants are randomly assigned to two (or more) groups to test a specific treatment or intervention. The experimental group receives an intervention (e.g. caffeine pill) and the control group receives either no intervention or an alternative intervention. Ideally this alternative would be a placebo (such as a sugar pill) so that the participants do not know which group they are in to reduce bias – this is known as ‘blinding’ the trial. ‘Double blinding’ the trial means even the researchers don’t know which participants receive the treatment so they cannot even unintentionally influence their outcome.
8. Meta analyses and systematic reviews – these are the 'gold standard' of research and we should base our answers on the outcomes of these, where they exist. They take a defined research question, and use systematic methods to identify all of the relevant studies that meet the strict quality criteria. Meta analyses then combine the results from these numerous studies to estimate the overall effect of a treatment / intervention. This provides us with an average result of a number of studies amounting to a large number of participants - sometimes tens or even hundreds of thousands.
Now that you know which types of evidence to look out for and which to ignore, how can we go about looking for these kind of research papers? Well, they tend to be published as part of scientific journals, which are available online. If you type Google Scholar into your browser, you can use this search engine specifically designed to browse scientific journals for free. Here are some great tips on how to get the most out of your searching. Unfortunately, many journals require a paid subscription to view the full articles. However, every paper has an abstract (an outline of the research including their main findings) available to view which usually provides enough information to answer your question. If you wanted to then go on to read the paper in full then you could check whether your local library has access to the journal, or if you are studying then check if your school, college, or university offers access.
But, perhaps most importantly, you now know whether to take a media headline or online article seriously. If they refer to, for example, just one cross-sectional study, then you should take their interpretation with a pinch of salt. If however they base their news on a meta-analysis, then you can be pretty sure that this is some of the best evidence we have available on a given topic.