Randomised controlled trials

Question 1

What is a clinical trial?

Answer 1

An experiment in which a treatment is administered to humans in order to evaluate it’s efficacy and safety

Question 2

What are the different types of clinical trials that can be administered?

Answer 2

Uncontrolled - everyone is given the treatment
Controlled - compare placebo group with the treatment group
Randomised controlled trial - allocations to groups is determined by chance

Question 3

What are the benefits of a randomised controlled trial?

Answer 3

Ensures that group receiving treatment A is similar to group receiving treatment B
Avoids selection/allocation bias

Question 4

What are the different types of blinding that can be used in trials?

Answer 4

Single blind - patients do not know what treatment they are on

Double blind - both the patients and the observers do not know what treatment the patients are on (not always possible e.g. surgery)

Question 5

What are the parallel vs. crossover methods?

Answer 5

Parallel group - randomise individuals into one of two groups (A or B) and the administer treatment A and treatment B and record the outcomes

Crossover - randomise the treatment sequence i.e. provide treatment A to group A and treatment B to group B and then after specified amount of time, provide treatment B to group A and treatment A to group B - the patient is provided with both types of treatments

Question 6

When would you use parallel vs cross-over grouping?

Answer 6

Parallel group - when the effect of the treatment is irreversible
Cross-over - when the effect of the treatment is reversible (because otherwise they would have already been potentially treated once they start taking the second medication)

Question 7

What are the advantages of cross-over trials?

Answer 7

Each patient is their own control - less room for error
Do not require as large a sample size
Better for subjective measurements

Question 8

What are the disadvantages of cross-over trials?

Answer 8

More time consuming - will take twice as long

Risk of carry-over effects (from treatment A to treatment B time period)

Question 9

How can carry-over effects be avoided?

Answer 9

Can administer a washout period between the treatment A period and the treatment B period

Question 10

What are cluster randomised trials?

Answer 10

This is where you randomise pre-existing groups to one of two treatments e.g. villages, schools, general practices

This avoids contamination and enhances compliance (community spirit)

Question 11

What is meant by relative risk and how is this calculated?

What value is used to depict this?

Answer 11

The relative risk of e.g. death is the risk of death in the treatment group/the risk of death in the control group

If the treatment has no effect then the RR = 1

Question 12

What different methods can be used to maximise compliance?

Answer 12

Ensure correct selection of patients - do not select patients that are too ill
Double blind trial design
Administer a run in period where all get the treatment - used to identify those who cannot tolerate it

NB. - these are not always possible/practical

Question 13

How can you use risk to determine how many patients you should treat to prevent one patient from getting the disease?

Answer 13

Calculate absolute risk difference i.e. difference between placebo group risk (3.5) and treatment risk (1) so 3.5-1=2.5
Divide this into 100 so 100/2.5 = 40
SO 40 patients should be treated to prevent one patient from having the disease

Question 14

What is the problem with too small a sample size?

Answer 14

If the sample size is too small then the study may not detect a real effect - the study does not have enough statistical power

Question 15

What is meta-analysis and what is it used for?

Answer 15

Meta-analysis is bringing together all of the evidence to more powerfully estimate the effect
Summarise the results of individual studies

Question 16

What are the issues associated with meta-analysis?

Answer 16

Heterogenecity e.g. a variation in the study results due to difference in study design, characteristics, dosage etc.
Publication bias e.g. journal editors, peer reviewers etc

Question 17

What is ‘registering’ a trial?

Answer 17

All trials are registered prospectively - must be registered to be published

Question 18

What are the advantages of registering a trial?

Answer 18

Allows in planning of new trials
Don’t have duplication of research
Can have collaboration between research groups
Optimal use of research funds by funding agencies
Avoid subjecting patients to trials seeking evidence that is already available
Improve opportunities for methodological research
Help to detect publication bias in meta-analyses

Question 19

What is ‘intention to treat’ analysis?

Answer 19

Analysing all subjects that were recruited into the trial and assigned a treatment group, regardless of whether they complied

Question 20

What is ‘on treatment’ analysis?

Answer 20

Analysing only subjects who actually took part in the treatment

Question 21

Should you use ‘intention to treat’ or ‘on treatment’ analysis when interpreting data?

Answer 21

Should use intention to treat

Question 22

How do you work out how many people with a disease must be given treatment to prevent one person with the disease?

Answer 22

Work out absolute risk (risk in placebo group-risk in treatment group)

100/absolute risk

Question 23

How do you determine a great enough sample size?

Answer 23

Should have a statistical power of at least 80% - generally should be 85%