RCT's Flashcards
What do RCT’s investigate?
The effect of any sort of intervention.
What standard of investigation are RCT’s often regarded as?
The ‘Gold Standard’.
What three questions are often asked with RCT’s?
Superiority, non inferiority, safety effect.
What is the purpose of randomisation?
To balance known and unknown covariates between treatment and control groups.
How can you remove allocation bias?
Through randomisation.
What value of n is needed for simple randomisation to be a viable form of randomisation?
200+
Why is varying block sizes good in blocked randomisation?
Stops practitioner bias.
Block sizes are always random in blocked randomisation. True or false?
False. They can be preset.
What is the advantage of using larger block sizes in blocked randomisation?
Order of events is less predictable.
When would the latin square technique be used with block randomisation?
When there are repeated measures of a treatment for a particular individual.
Do you stratify before or after blocking?
Before.
Stratification variables should be of _____ significance.
Prognositic.
Name four variables you can stratify by.
Gender, age, disease status, assessment centre.
What is the goal of minimisation?
To maintain balance of stratification variables during allocation.
When is minimisation and effective randomisation technique?
Small studies with multiples stratification factors.
When interventions vary between regions (eg vaccine interventions, verbal GP interventions) what sort of randomisation should be used?
Cluster randomisation.
What is an advantage of cluster randomisation?
It reduces risk of the intervention contaminating the control group.
Why is blinding necessary?
As randomisation alone does not prevent differential treatment later on.
Who is blinded in a triple blinded trial?
Participants, investigators, assessors.
Why are sham procedures not used anymore?
They are unethical.
Do parallel randomised trials require the same amount of people in each arm?
No.
What sort of comparisons are made in a parallel studies?
Between person comparisons.
When are crossover studies used?
When you need more statistical power (when the studies are small).
What type of conditions are crossover studies good for?
Chronic.
What determines the order of interventions in a crossover design?
Randomisation.
Why is variability minimised in a crossover RCT?
As each subject is their own control.
Factorial designs look at multiple outcomes for one question. True or false?
False. They can look at multiple outcomes and multiple questions.
What do the treatments in the factorial studies have to be?
Independant/ complementary.
What sort of effect(s) can be studied with a factorial design?
Independant or interacting.
What needs to be defined prior to a non-inferiority trial?
Need to define what is statistically ‘not different’.
What is the general rule of thumb in a non inferiority trial?
A quantity must be significantly smaller (1/2 or a 1/3) than the minimal clinical difference that can be used to calculate the required sample size in a superiority trial.
What is randomised in cluster randomisation?
Groups rather than individuals.
What sort of level of interventions do cluster randomisation trials look at?
Population level interventions.
What happens in a trial that has ‘adaptive design’?
Multiple arms get dropped as the trial procedures enriching the study. This means the trial can be modified during study.
What is increased in a adaptively designed trial?
Time and cost efficacy.
What sort of statistics are used in adaptively designed clinical trials?
Baysian.
Eligibility criteria in RCT’s allow for bias to be minimised. true or false?
False. The limited nature of the eligibility criteria introduce bias.
Who has the sponsorship of a trial?
Those legally responsible for the trial, not those paying for it.
What sort of things can be modified as a adaptively designed trial procedures (3 things) ?
- Sample size.
- Dose.
- Focus onto a particular group.
Name four reasons why an individual may be lost to follow up.
- Withdrawal.
- Discontinuation of treatment.
- Unable to locate.
- Mortality.
What are the two types of analyses that can be used in a RCT.
- Intention to treat.
2. Per protocol.
What is the definition of ‘intention to treat’?
All randomised participants are included in the analyses of the group they were allocated to.
What is the definition of ‘per protocol’?
Only those who were randomised according to the protocol and those who followed the protocol correctly are included in the trial.
What must trial results be?
Transparent. Everyone must be accounted for.
Despite varying methods both ‘intention to treat’ and ‘per protocol’ analysis will give rise to the same results. True or False?
False.
In both the US and the EU publication of trial results is a legal requirement. True or false?
False. It is in the EU.
What stage of a clinical trail looks at drug metabolism and bioavailability?
1st.
What can the 1st stage of a RCT also be called?
First in man.
What is the main aim of stage 4 in RCT’s?
Post marketing surveillance to determine rare side effects and long term effects regarding morbidity and mortality..
Who are the HRA and what do they do?
Health Research Authority- UK ethics board.
What does CONSORT stand for?
Consolidated standards of reporting trials.
RCT’s allow for greater confidence in what compared to observational studies?
Causality.
What is reduced in a RCT compared to an observational study?
Confounding.
