Evidence-Based Medicine and The Research Pyramid
Evidence- Based Medicine
Table of Contents
Pyramid of Evidence
Go ahead and sear this pyramid into your memory because you’ll be using it as a guideline forever.
Not all research is the same and you can setup studies in many legitimately different ways. This pyramid features the very best research studies right at the top. We’ll go over each level of the pyramid, talk briefly about what they mean, discuss some important principles in research, talk about a few example of evidence-based medicine in practice, and then close with what you can do to help combat the spread of false information and bad studies.
General Principles for Research and Evidence-Based Medicine
I’m going to introduce you to four general principles that apply to all research – peer review, size, bias, and reproducibility. A good research study is PEER REVIEWED, has a LARGE SIZE and CONTROLS FOR BIAS (meaning that there’s little or no bias). Finally all good research needs to be REPRODUCIBLE!
Peer Review: Never trust a study that hasn’t been peer reviewed! What is peer review? It’s when a panel of experts in the same field review your work. Researchers and scientists really kind of love arguing with each other. Even if they respect the other person’s work they’ll still happily challenge it if they think there are flaws. This is one of the benefits and why we have peer review! Work needs to be challenged by other experts and stand up to their challenges.
Size: How big a research study needs to be for it’s results to matter depends on some pretty complicated math. However, as a general rule be wary about the size of any research study that your research. This is one of the most common mistakes in research and in real life. How many times have you heard someone say a thing must be true because it happened to their friend? We would call that a sample size of n = 1. That’s not a statistically meaningful sample size. As a rule, the larger the sample size, the more meaningful your results will be.
There are many famous examples of bad research that used very small sample sizes. Bad research like this has caused a lot of damage in medicine and continues to do so!
Bias: You’ve heard of people having biases. Well they exist in research too! There’s lots of different biases. When we talk about bias in research we don’t necessarily mean discrimination – there’s other types too. In research, any type of action or question that influences the results of a study is considered a bias.
Reproducibility: Finally, no matter how good a research study is we have to ask ourselves is it reproducible? Science depends on us being able to retest, reverify, and repeatedly challenge even the best research studied – even the things we want to be true.
Sitting at the very top of the pyramid is the meta-analysis. This is a type of study that looks at other studies. Researchers conducting a meta-analysis take the very best evidence they can find. They use very specific criteria in their search, and share that criteria in their results.
The Meta-analysis sits at the top because it doesn’t just rely on one study and it uses a quantitative statistically analysis. It statistically looks at a multitude of studies and then determines if the findings are statistically significant. It is an even higher quality version of a systemic review. You can read more about meta-analysis here.
Some of the best evidence-based medicine comes from meta-analyses.
Did You Know: Physicians continuely learn and train, even after graduation.
*Statins: Medications that lower lipids (fats) in the blood and reduce the risk of cardiovascular disease.
Just one step down from the meta-analysis is a Systemic Review is also a very high-quality piece of evidence. Unlike the meta analysis a quantitative statistical analysis is not done with a systemic review. That’s the big difference between the two.
Question: Would you trust this systemic review about the determinants of weight loss?
Remember: Not everything needs to be a meta-analysis to be a good piece of research!
Randomized Controlled Trials
A Randomized Control Trial or RCT is what a lot of people think of when they think of research. In these trials you give one group of patients a placebo (fake treatment) and another group the real treatment. It’s random who gets what. You can even make it double-blinded (meaning you and the patient each don’t know who is getting the real drug vs the placebo). Making it double blinded helps to reduce biases.
Question: Why do you think we try to make research double blinded? Would you expect a person getting the real treatment to feel better? Would you feel comfortable giving a placebo cancer treatment to someone who has terminal cancer?
This is a type of study where we take a bunch of people who share a specific condition or trait – like smokers and follow them. In this example maybe we’re trying to figure out how many smokers will eventually develop lung cancer if they continue to smoke. This type of study is purely observational and not as strictly controlled as RCTs.
Case Control Study
Similar to a Cohort Study in that we have two groups but we look back over time and see what the differences are. In this case maybe we look at people who have lung cancer vs those those don’t have lung cancer. Then we search through their social history to determine how many people with lung cancer were smokers.
These are like a series of case reports and also at the bottom of the pyramid and are not bad pieces of work. They just lack statistical validity. So you shouldn’t be generalizing what you present in a Case Series to other people. A lot of crumby evidence online is someone generalizing a case report and saying that it’s true for everyone!
Advice for Using The Evidence-Based Medicine Approach Online
Interested in learning more about how evidence-based medicine guides treatment? Want to learn how medical students see this firsthand while in school? This is a great topic for you to discuss with our Medical School Mentors or your Personal BreakThru Mentor.
Unlock more content, more resources, and more mentoring by becoming a Premium BreakThru Member Today!
Want individual guidance? Apply today for six months of personal mentoring from a BreakThru Medical Student Mentor!
The Path to Medical School Can Feel Long And Unclear
But never forget that we’re here to help. Do you have a question about something you just read in our BreakThru Learner Series?
Use the BreakThru Feed to Ask and Inspire others with your questions. Our Medical Student Mentors are always here to help you, for free, anytime.
Where Mentoring Happens