Lecture 21: Randomised Control Trials Flashcards
What type of studies are randomised control trials?
analytic and intervention
What are our two groups that we compare in a randomised control trial?
treatment group and comparison group
What does the “randomised” part of RCT involve?
participants are randomly allocated into groups
What does the “contolled” part of RCT involve?
always have a comparison (control) group
What does the “trial” part of RCT involve?
testing effect of treatment/interventions
What does a trial involve?
we take a sample from the source population and assign participants to the exposed (treatment) or comparison (control) group
What does randomisation involve and how does this differ from random selection?
Randomisation is the same as random allocation which is when people are randomly allocated into either the treatment or control group. There must be an equal chance for each participant to be in either group.
Random selection is the random selection of people from a source population into the sample
What is the advantage of randomisation?
if enough people are randomly allocated, there should be the same proportion of a confounder in each group (for both known and unknown confounders).
What does successful randomisation mean?
confounding is an unlikely reason for differences in outcome between groups
What are the steps to carry out an RCT?
- Identify source population
- Randomly select sample who don’t have the outcome of interest
- Randomise the sample to either the intervention or control group
- Follow up over time to see who develops the outcome
What two things can be done to protect the randomisation?
- concealment of allocation
- intention to treat analysis
What is meant by “we can protect randomisation by concealment of allocation”?
It is important that the allocation sequence is concealed and unpredictable. This means that by looking at someone, we can’t know the allocation of that person and each allocation is independent of another
What is meant by intention-to-treat analysis and why is it important?
If people are randomised into the treatment group, you analysis their results as if they are in this group, even if they did not take the treatment. This is important because it can reflect the “real-world” effect of intervention ie. not everyone is going to take the treatment in real life
What are three potential sources of bias?
- blinding
- loss to follow up
- non-adherence
Why is blinding important?
We need to not let the researcher or the participant know which group they are in because other wise people may react differently depending on which group they are in
What is non-adherance?
When participants do not do what they are supposed to which can include doing what the other group is doing
What are the advantages/strengths of an RCT?
- it is the best way to evaluate an intervention
- we can calculate incidence and therefore relative risk and risk difference
- we can demonstrate a causal association
How can loss to follow up lead to bias?
People who leave the study which can lead to bias and confounding as those who leave the study may be different from those who stayed
What are the two ways to randomise and what do these involve?
- Cluster randomisation: when subgroups are randomised rather than individuals
- stratified/block stratification: where participants are randomised within blocks
What are 6 limitations of an RCT?
- Many exposures we cannot randomise
- Need clinical equipoise
- Expensive
- Time consuming
- Participants may not represent the general population
- Not good for rare outcomes
What is clinical equipoise?
genuine uncertainty about the benefit/harm of the intervention
In terms of clinical equipoise, what three things its it unethical to do?
- give known harmful interventions to people
- give interventions known to be less effective than current treatments
- waste resource and risk people’s well-being if you already know the answer
In an RCT, the exposure needs to be
modifiable
eg. age and vaccination status are not exposures because they can not be changed
What is the problem with RCTs being highly selective?
Often a very select group of people are included which can affect generalisability
