Randomisation Flashcards
What is fixed randomisation?q
Randomisation methods are defined and allocation sequences are set up before the start of the trial.
What are types of fixed randomisation?
Simple
block
Stratified
What is simple randomisation
Assigning trial subjects to T/C using chance to reduce allocation bias and confounding factors
e.g. computer generated random numbers
when next subject is assigned a group, the previous allocation is not considered
What is a problem with simple randomisation
Confounding factors have equal chance of entering either group
HOWEVER in SMALL trials especially this may result in unequal group sizes and unequal distribution of confounding factors
What is block randomisation
Subjects are placed into blocks
Once a block is filled with subjects, they are allocated in equal numbers into different groups of the study
What are advantages and disadvantages of block randomisation
Advantage: treatment groups of equal sizes
Disadvantage: more complicated than simple randomisation, small block length makes it possible to predict the next allocation (so block of 2 never used)
What is stratified randomisation
Stratifies patients according to factors that could influence prognosis
Then randomises each stratum into the two groups
What are advantages and disadvantages of stratified randomisation
Good balance of participants in each group
May prevent type 1 error
What is cluster randomisation?
subjects are randomised as a pre-existing group rather than individually, into the same arm together
E.g. a clinical practice is a cluster
WHen is it appropriate to use CRT
When individual randomisation to treatment arms is not possible or the intervention is applied to the whole cluster
What is a disadvantage of CRT
Loss of statistical power and additional complexity in design, conduct, analysis
What is adaptive randomisation
randomised groups are adjusted as the study progresses to account for imbalances in numbers in the groups
When is adaptive randomisation useful
in small trials - to ensure good balance between groups for prognostic factors
WHat is minimisation in adaptive randomisation?
Only 1st person in truly randomised - itger participants are assigned according to important prognostic factors to minimise differences between groups
What is the right way to select primary outcomes?
Primary outcomes should be of impact to the patient or healthcare provider
Should be identfied via Delphi survey
