So, you want to call yourself eViDeNcE-bAsEd. This extends beyond posting complicated scientific articles on your Instagram story. It also helps immensely to be able to understand them, and where to find them. You cannot call yourself evidence-based if you don’t know what the evidence is, or how to get it.
Do you need to read scientific articles to be evidence-based? Absolutely not. However, finding and critiquing scientific articles and evidence is rarely taught in any personal training course in Ireland (that I am aware of!).
If you want to read articles, and find out a bit more about how to get access, keep reading!
Slow Down, Egghead: What is an Article?
Scientific articles are papers published based on scientific evidence, published in scientific journals. There are millions of articles and thousands of journals worldwide, covering a huge array of topics. Scientific articles allow us to keep up to date with the latest developments in an area (heyi evidence-based practice).
For an article to be published in a journal, it must undergo peer-review. This means that experts in the field evaluate the quality of the research, and identify weaknesses or errors. This in turn improves the quality of research overall. Peer-review in science is like peer-review in everyday life. You want to send a risky DM to your secret Instagram bae. So you run your message by your mates, who inform you it’s best to avoid mentioning that you also went to Tenerife in 2010. You remove any stalker-undertones from your DM, and you are more likely to elicit a response. Science is no different (but a lot less stalkery).
How to Access Articles
Scientific articles aren’t hiding in plain Google searches. She’s a little too refined for that. You can access the millions of scientific articles through academic search engines. Google Scholar is fantastic, as is ScienceDirect or PubMed. It is important to note that these sites are search engines. They don’t actually publish the articles, and are not journals.
Behold, you now have the first key to the kingdom of science.
Scientific Gated Communities: What Next?
So, you can search for articles. Now you can read them, right? Wrong.
The majority of articles aren’t fully available on most websites. Most websites will provide the abstract of an article. And this is precisely where you need to be careful. Reading an abstract does not mean you have read the article. As we will discuss in the next post, an abstract is useful to sum up the key points of an article, but that is where it ends. More of that in the post on critiquing and reviewing articles.
Back to accessing articles. Most websites such as PubMed will have a “click to view the full article” prompt, where you can pay to view the full article. Alternatively, most universities and colleges have affiliations with the big databases like ScienceDirect and PubMed. If you (or a friend) have a college log-in, you can use that to access the full text of an article.
Once you have found a title of an article, it may also be worth searching on ResearchGate. Most researchers will put the full-text of their papers for free up here, or if you send them a nice email they are usually more than happy to oblige.
Alternatively, there are a few pirate sites that hack paywalls. Taking me back to my LimeWire days. Obviously, I am not going to provide the name of the sites, but if you know you know 😉
How to Utilise Search Engines
To get to what you need to find, you need to be specific. Decide what exactly it is you want to learn about. For example, say you want to investigate what exercises are the best for building glutes. You wouldn’t just search for “best glute exercise”. Similarly, if you wanted to find out where the best pizza is in Dublin (it’s Vincenzos), you wouldn’t just Google “food”. Be specific, and include keywords in your search. Nail down your question, and this will help you get more relevant results. Because there’s a shit ton of research to sift through.
In addition, most search engines like Google Scholar have the option to sort by year. A general rule is to try and look for studies published in the last five years. Again, this is not definitive, and obviously depends on your question.
How Do I Know an Article Is Good?
You would be forgiven for thinking that if something is peer-reviewed and published it is accurate and valid. Hell no. A lot of idiots get published (indeed I stand a good chance!), and there’s a lot of misinformation out there. “Good” is subjective obviously, and study-dependent, but here are a few common considerations when deciding what articles to read.
Where Is It Published?
Making sure it’s from a peer-reviewed journal is the cornerstone. This rules out Wikipedia, random websites, all your favourite Instahuns and obviously this post itself *the fourth wall crumbles*. None of these are peer-reviewed. A peer-reviewed journal is step one.
Step two involves the quality of the journal itself. Not all journals are created equal. Generally, you want to look for articles published in a journal with a high impact factor.
What the hell is that? Good question.
A journal’s impact factor indicates the average number of times a study published in the journal will be cited. This is generally used to indicate the importance of a journal: the higher the impact factor, the more the journal has contributed to its field. For example: the British Medical Journal has an impact factor of 30, whereas the Irish Medical Journal has an impact factor of 1.2.
Using the impact factor alone to decide the “worthiness” of a study is not a perfect science. A lot of smaller journals are very specific to their field, and will have a lower impact factor. Again, it depends on your question. If you were looking to investigate the best glute exercises, whilst the BMJ would have a higher impact factor, the Highly Specific Journal of Glute Exercises for Hypertrophy* may yield better results. (*not a real journal sadly).
Impact factor can be used as a general guide, but it isn’t everything.
Who Else Cares?
The number of citations can also be helpful. The number of times a study is cited is given beside the article in Google Scholar (really handy!). In general, the more times a study has been cited, the more it has contributed to its field. This is similar to the impact factor we spoke about in the last section.
Number of citations can also be affected by a variety of factors. More recent articles will have lower citation scores, as they have obviously just been published. Depending on the nature of the article, it may be very specific to one area, and therefore won’t be as widely cited.
The author themselves can also be a good indicator. But again, this comes with experience. From reading around the field, you will get experience with who the notable figures are. The same names will crop up.
What Kind of Study is It?
Not all studies are the same, and each can be advantageous in its own regard. A quick Google of the “Hierarchy of Scientific Evidence” brings up the different types of study.
Case Reports, Opinion Papers & Letters
These tend to be the weakest on the chain of evidence. The reason for this tends to be their lack of reproducibility. One case in isolation cannot be generalised to a wider population. Similarly, opinions have no proven evidence. They can certainly be an excellent educated guess, and it can be useful to see the hypotheses of thought leaders in a field.
It doesn’t mean that we can dismiss the evidence of this type of research, at all. It is simply a cautionary tale against applying results to everyone. For example, if someone were to write a case report on my ACL surgery, we couldn’t generalise the results for everyone. My first ACL reconstruction used my patellar tendon and failed. My second reconstruction used my hamstring and was far more successful. Does that mean we can say patellar tendon grafts are more likely to fail? Hell no. I am just one case. One person does not a scientific law make.
Animal & In-Vitro Studies
These are second from the bottom in the pyramid of evidence. Again, this doesn’t mean we can write off all of their findings as invalid. For many suggested interventions on humans, we are required by ethics to prove their effects on animals. Sounds crass, and vegans look away, but if an animal is adversely affected by an intervention, chances are humans will be, and so we can point humanity in a safer direction.
Issues arise when it comes to generalising animal results to human populations. This is particularly evident in the field of chrononutrition (my secret nerd love). Many studies carry out interventions on rats. But rats are not like humans. They are nocturnal, and have four legs. Therefore, we cannot say humans will react exactly the same. But it brings us a bit closer!
In-vitro means in a test-tube or outside of the living organism. You may recall from biology in school examining cells under a microscope etc. That kind of buzz. Again, studying human cells in isolation can be useful, but does not give an indication of how the organism as a whole will react, and our cells rarely act in isolation.
Cross-sectional studies are observational studies. Variables are observed in a population at a particular point in time. The issue with these kinds of studies is that they only study a particular point time, and it is hard to discern what are the causes of outcomes. For example, if we were to study the prevalence of coronavirus amongst radiographers during the first wave of lockdown, we might note a number of variables. For example, those exposed to more positive patients, females, working more hours in the hospital or in public hospitals might be more likely to contract the virus. We cannot isolate a single variable as the cause for a particular outcome. For example if female radiographers have a higher incidence of contracting the virus, we cannot say that gender alone makes you more likely to contract the virus. It can be a number of other factors.
It can be useful to identify the influences on a certain outcome, but it can be difficult to draw exact cause and effect from cross-sectional studies.
Case-control studies are a step up from the cross-sectional study. In these studies, the population being studied is split into two groups. The incidence of outcome is observed and differences between the two groups are noted. They are typically used in studies where we haven’t much idea what causes a particular outcome. For example, it is used to assess risk factors for rare diseases.
They are observational, and this makes it difficult to observe the exact cause of a particular outcome.
Cohort studies are longitudinal studies. This means they examine their participants over a longer period of time. The population being studied are separated by a particular characteristic, and an outcome assessed over a period of time.
They provide stronger evidence than case-control or cross-sectional studies, in that multiple outcomes associated with multiple exposures can be studied. When you only observe something at one time point, you can miss out on influential variables.
However, as they are observational studies, they cannot offer definitive cause and effects. They can offer a good starting point though!
Randomised Control Trial
Randomised control trials (RCTs) are the absolute gear. During RCTs, a group of similar people are split into two. One group receives the intervention, and the other does not. The groups are both assessed and measured at specific times, and statistically analysed.
They are highly advantageous in that the statistical evidence reduces the chance of bias. Big pharma love a good RCT to assess the efficacy of certain drugs. You may have heard of cancer patients going for “clinical trials”. They reduce bias and confounding variables.
However, the results are generally population specific, and cannot be generalised to those outside of the strict criteria! They must be reproduced for different populations to be generalisable.
Meta-Analyses & Systematic Reviews
The big dogs. The daddy of scientific research (yes we are officially peak single in that we are calling science daddy). They combine the research of multiple other studies and from that draw conclusions. They tend to be based on RCTs. They synthesise the results of previous research to draw conclusions about the research.
Results tend to be generalisable and more precise at estimating the influencing variables on a particular outcome.
However, the almighty meta-analysis is not without fault. The inclusion criteria of a meta-analysis can influence the validity of its conclusions. For example, if a meta-analysis of a particular intervention only includes studies that are published in English in the last five years, we could miss out on some key findings.
This is particularly evident in nutritional science, where hugely impactful studies will be omitted by researchers for trivial reasons. This is fantastically explained by Alan Flanagan and Danny Lennon on the Sigma Nutrition podcast (click here).
Conclusions: Wait, What?
Hopefully from reading this, you will have gained an appreciation of the intricacies of scientific research, and understand that it is not as simple as posting a picture of an article. Bear in mind, this post was only about finding articles. We haven’t even gotten into reading them!
I hope you also understand that no study is without fault, and ultimately the articles you need depend on what exactly it is you want to know.
In general, when searching for articles look for:
- A peer-reviewed journal
- A high number of citations
- Publications in a journal with a high-impact factor
- Recently published articles
- Established authors who are experts in the field
- Meta-analyses where possible
I accept that this is definitely a bit heavy for most people to get to grips with, and if you are a PT you will most likely wonder why the hell you weren’t taught this in the first place. I promise it all becomes clearer the more of this you do. And hopefully you read on to the next post, where we actually read articles (and tear apart my undergraduate thesis in the process!).
Peace, love and science,