By Dr. Sean Delanghe BSc. (Hons), DC

Hi, I write a lot of articles about health, injuries, exercise and pain. But there is an important question behind everything I write:

How do we actually know that this is good advice?

We live in an interesting time. The internet has given almost everyone access to an enormous amount of health information. Google can find you an article, TikTok can give you a 30-second explanation, and someone with 100,000 Instagram followers can confidently tell you that their favourite treatment is “backed by science.”

But there is a big difference between having an opinion about health and having good evidence to support that opinion.

When I write an article, I am generally not starting with, “Here is what I think.” I’m one person who is, despite my best efforts, full if biases!

Instead, I start by asking:

What does the research actually show?

And not all research is created equal.

There is a general hierarchy of evidence that researchers and clinicians use to help determine how much confidence we should place in different types of information.

If you don’t want to read more, long story short, the vast majority of my articles are based on systemtic reviews and meta-anlaysis- the absoutely highest form of medical evidence.

But what exactly does that mean? Here’s a summary in order ot highest worst to best.

1. Professional or Expert Opinion

“I have treated thousands of patients and this is what I see.”

This is often where health information starts.

A doctor, physiotherapist, chiropractor, researcher or other professional may have decades of experience and develop a strong clinical opinion about what works.

Experience is valuable. Clinicians learn a tremendous amount from seeing real patients.

The problem is that experience can be misleading.

If 100 patients receive a treatment and 80 improve, it might seem obvious that the treatment worked. But what if many would have improved anyway? What if they were also exercising, sleeping better or simply recovering naturally?

Without a comparison group, it can be difficult to know what actually caused the improvement.

Bottom line: Expert opinion can be useful, particularly when research is limited, but it is relatively vulnerable to personal bias and other explanations.

2. Case Reports and Case Series

“We treated this person, and this happened.”

A case report describes an individual patient. A case series describes a group of patients with something in common.

These can be useful for identifying unusual conditions, unexpected side effects or interesting new observations.

For example, if a doctor sees an unusual reaction to a medication that has never been reported before, documenting it can alert other researchers to a possible problem.

But there is an obvious limitation:

There is usually no comparison group.

We don’t know what would have happened if the patient had received a different treatment—or no treatment at all.

Bottom line: Useful for generating questions and identifying unusual findings, but generally weak evidence for proving that a treatment works.

3. Case-Control Studies

“Let’s compare people who have the problem with people who don’t.”

Researchers can take a group of people who have a particular condition and compare them with people who don’t.

They then look backwards to see whether the two groups were exposed to different risk factors.

For example, researchers might compare people with a particular type of back pain with people without it and ask whether their previous activity levels, occupations or other factors were different.

This can be particularly useful for studying relatively uncommon conditions.

The limitation is that researchers are looking backwards, which can introduce problems such as recall bias and other confounding factors.

Bottom line: Useful for identifying associations and potential risk factors, but generally cannot prove that one factor caused another.

4. Cohort Studies

“Let’s follow people and see what happens.”

A cohort study follows a group of people over time.

For example, researchers could follow thousands of runners for five years and record how much they run, their training habits and whether they develop injuries.

This can provide useful information about relationships between exposures and outcomes.

However, it is still an observational study.

People aren’t randomly assigned to different behaviours. A runner who trains 100 km per week may also sleep differently, eat differently, have different genetics and participate in different sports than someone running 20 km per week.

So if the 100 km runner develops more injuries, we can’t automatically say that mileage caused them.

Bottom line: Very useful for studying risk factors, prognosis and long-term outcomes, but associations don’t necessarily prove causation.

5. Randomized Controlled Trials

“Let’s actually test the treatment.”

This is where things get particularly interesting.

In a randomized controlled trial (RCT), participants are randomly assigned to different groups.

For example:

• Group A receives Treatment X
• Group B receives Treatment Y or a control
• Researchers compare the outcomes

Because participants are randomly allocated, the groups should be relatively similar at the beginning of the study.

That helps reduce the influence of other factors that could affect the outcome.

This is one of the main reasons randomized trials are generally considered strong evidence for determining whether a treatment actually causes an improvement.

But RCTs aren’t perfect. They can be expensive, take years to complete and sometimes involve highly selected patients who don’t perfectly represent people in everyday practice.

Bottom line: For questions such as “Does this treatment work?”, a well-designed randomized controlled trial provides some of the strongest individual pieces of evidence available.

6. Systematic Reviews

“Let’s find all the good studies and look at them together.”

Now we get to one of the most useful levels of evidence.

A systematic review doesn’t usually conduct a new experiment. Instead, researchers systematically search for all of the relevant studies addressing a specific question.

They then assess the quality of those studies and summarize what they found.

Imagine you have 15 randomized trials looking at whether a particular exercise program helps people with knee pain.

One study says it works. Another says it doesn’t. Several show a small benefit.

Instead of choosing whichever study supports your preferred opinion, a systematic review looks at the entire body of relevant evidence.

This matters because individual studies can produce misleading results through chance, differences in methodology or other limitations.

Bottom line: For many treatment questions, a well-conducted systematic review of randomized controlled trials is considered one of the strongest forms of evidence available.

7. Meta-Analysis

You will often hear the term meta-analysis alongside systematic review.

They aren’t exactly the same thing.

A systematic review is the process of finding, assessing and summarizing relevant research.

A meta-analysis is a statistical method that can sometimes be used within a systematic review to combine the numerical results of multiple studies.

In simple terms, instead of saying:

Study 1 found a small benefit.
Study 2 found a moderate benefit.
Study 3 found little benefit.

Researchers can statistically combine the results to estimate the overall effect.

This can give us a more precise estimate of what the research collectively suggests.

But- putting bad studies together doesn’t magically create good evidence.

A systematic review can still be limited if the studies it contains are poorly designed, inconsistent or answering slightly different questions.

Bottom line: Meta-analysis can strengthen our understanding of an overall effect, but the quality of the underlying studies still matters.

So What’s the Hierarchy?

For a straightforward treatment question, you can think about the evidence roughly like this:

Lower confidence

Expert opinion
↓
Case reports / case series
↓
Case-control studies
↓
Cohort studies
↓
Randomized controlled trials
↓
Systematic reviews / meta-analyses of randomized trials

Practicel Applications

Knowing this hierchay is a massive step in the right direction, but not the full picture. This is not a simple “higher is always better” pyramid.

Different types of studies are better suited to answering different questions. A randomized trial might tell us whether a treatment works. A cohort study might tell us what happens to people over 10 years. A case report might be the first clue that a very unusual side effect exists.

When I write about a health topic, I try to work my way up this evidence ladder. I try to find the best available research, with particular attention to systematic reviews and randomized controlled trials when they are appropriate for the question.

Then I look at the quality of that research, whether the results are consistent, how large the effect actually is, and whether the findings apply to the average person sitting in front of me.

As always, if you have any questions, free free to reach out or book online HERE.