Randomized control trial

Definitions & Key takeaways

A randomized controlled trial (RCT) is a type of experimental research design in which participants are randomly assigned to one of two or more groups, such as a treatment group or a control group. The goal is to determine the effectiveness of an intervention or treatment by comparing the outcomes between the groups. Patients who are in a control group receive a placebo instead of the active treatment. The groups are then followed over time to see if there is a difference in the outcome measures (such as symptoms, quality of life, or survival) between the groups.

Watch video only
Randomized controlled trials or RCTs are a type of study design that’s often used to figure out if there’s a causal relationship between an exposure and an outcome.
For example, let’s say we want to find out if a newly discovered drug, let’s call it Drug A, can prevent migraines for up to a year.
In this example, Drug A is the exposure and having a migraine is the outcome. In the most basic randomized controlled trial, the sample population might be randomly split into two treatment groups, an exposure group that receives Drug A, and a control group that receives a placebo.
The placebo looks and tastes like Drug A but is completely harmless and ineffective - like a tiny capsule filled with water.
After both groups get their treatments, researchers would compare the number of individuals in each group who got migraines over the next year.
Typically, the goal of a randomized controlled trial is to figure out if the intervention can help some target population - usually that’s just people in the general population.
That means that it’s important to perform the trial on individuals that accurately represent the general population. In other words, the sample population should be similar to the general population.
For example, if researchers want to find out if Drug A helps prevent migraines in women, then women are the target population.
And the randomized control trial should be done on women rather than being done on men. Furthermore, if the goal is to use Drug A for woman around the world, then the sample population shouldn’t just include women living in Vancouver, British Columbia.
Instead it should include women of all ages, races, and socioeconomic statuses from around the world. If the sample population and that target population are really similar, then the randomized controlled trial has high external validity, meaning that any conclusions made about the sample population can be applied to the target population, which is good news for the manufacturer of Drug A and hopefully means fewer headaches for women around the world!
Now, to make the sample population represents the target population, one tool that can be used is randomization, meaning that individuals get selected to enter the study through a process of chance.
To show how that works, let’s say they put the names of every women in the target population into a brown paper bag, a big brown paper bag.
So that’s 3.5 billion names in that bag - probably with a number of repeats. Then let’s say that you choose a thousand names out of the bag to include in the study - either by simply picking them or by using a computer program to make sure that it’s truly by chance.
That’s randomization. Randomization minimizes the chance of selection bias, which is when researchers decide for themselves who to include in a research study.
For example, the researchers might choose slightly older women with a median age of 60, simply because they have more time and are more willing to join the study.
Now, as it turns out the median age of a woman around the world is 30. So, in this example, the study results wouldn’t necessarily apply to the target population, and the study would have low external validity.
After figuring out who will participate in the study - the sample population - the next step is to divide the participants into treatment groups.
This is done by random allocation, where an individual is assigned to a treatment group once again by chance. The goal is to make sure that the treatment groups - whether there are two like in our study or many different treatment groups - all have similar characteristics.
That way the only difference is that one effect that we’re trying to study, which is that of Drug A compared to a placebo.
For example, let’s say that one treatment group had only women from Europe and another treatment group had only women from Asia.
In that situation, it’s hard to know if a difference in migraines is due to the medication or due to a biological or cultural differences between women from these different geographic regions.
Random allocation tries to ensure that the treatment groups have comparable characteristics, and if that’s achieved, then the study is said to have high internal validity.
The flip side of this would be non-random allocation. Let’s say that a researcher talks to a woman who is selected to be in the sample population, and he decides to give her Drug A for intentional or subconscious reasons.
Perhaps it’s because she suffers from terrible migraines and he would like to see her benefit from this experimental medication.
Often times this happens when a researcher believes that the drug is likely to help and when an individual is at high risk of the outcome.
Over time, the exposure group that gets Drug A and the control group that gets the placebo begin to look different. High-risk individuals that are likely to get more migraines end up in the exposure group, and low-risk individuals that are likely to get fewer migraines end up in the control group.
And in this situation, Drug A would look less effective because that group is getting more migraines. In an extreme situation, it could even look like Drug A causes migraines!
All of that is because the two treatment groups don’t have comparable characteristics - and therefore the study has low internal validity.
Blinding or masking, is another technique that increases internal validity in a study. In a blinded study, the participant does not know which treatment group they are assigned to - in other words, they don’t know if they are using Drug A or a placebo.
This is important because, if a participant finds out what treatment group they’re in, it might change their perception of the treatment.
For example, women in the Drug A group might say that they’re having fewer migraines, simply because they believe the drug is working.
On the other hand, women in the control group might say they’re having more migraines, since they know they’re taking a placebo.
In this example, the exposure group would have an underreporting of migraines, and the control group would have an over reporting of migraines, which will make it seem like Drug A is more effective at preventing migraines than it really is.
There’s also double blinding, which is when the researchers conducting the study also don’t know which treatment group a participant is in.
In a double blinded study, each participant is assigned a unique code, and the key for all the codes is kept locked up and hidden away until the end of the study.
Once all the data have been collected, the key is revealed and the researchers get to find out who was in each treatment group.
You can imagine how exciting it would be to look at the key for the first time, especially since most randomized controlled trials last at least several months, and in some cases they last several years!
Alright, as a quick recap, randomized controlled trials are used to determine if there’s a causal relationship between an exposure and an outcome.
Randomized controlled trials have high external validity if the sample population has similar characteristics to the target population, and high internal validity if the various treatment groups have similar characteristics.
Randomized selection of individuals increases external validity, while random allocation of treatment and blinding increase internal validity.