Clinical Trial Design to reduce bias

Question 1

List the 7 forms of bias

Answer 1

1. selection bias
2. performance bias
3. detection bias
4. attrition bias
5. observer bias
6. reporting bias
7. publication bias

Question 2

What is meant by ‘Selection bias’

Answer 2

the systematic difference between baseline characteristics of the groups that are compared

Question 3

What is meant by ‘performance bias’

Answer 3

Systematic difference between the care that is provided or in exposure to factors other than the interventions of interest

Question 4

What is meant by detection bias

Answer 4

systematic differences between groups in how outcomes are determined

Question 5

What is meant by ‘attrition bias’

Answer 5

systematic differences between groups in withdrawal from a study

Question 6

what is meant by ‘observer bias’

Answer 6

failure to measure or classify the exposure or outcomes appropriately

This can be achieved through interview bias where the investigator is aware of the treatment/placebo and the measurements

recall bias - selectively reporting events by the participant

Question 7

What are 3 methods to avoid bias

Answer 7

1. randomisation
2. blinding
3. controlled groups

Question 8

What is meant by the ‘controlled group’

Answer 8

the group that does not receive the intervention (either placebo or the standard of care

Question 9

what 2 bias’ can occur in randomisation process

Answer 9

1. selection bias - when allocation is predictable - recruiting participants at specific time points to avoid a specific group
2. accidental bias - imbalance of risk factors or prognostic markers

Question 10

Name 3 types of randomisation

Answer 10

1. simple randomisation
2. blocked randomisation
3. stratified randomisation

Question 11

What does ‘simple randomisation’ consist of?

Answer 11

computer assigns different treatment group in a random order

e.g. ABABABABA

Question 12

What does ‘blocked randomisation’ consist of

Answer 12

it reduces the risk of accidental bias

creates equal sized sample groups

block of 4 randomisation for treatment A and B - one of the 6 arrangements are then selected randomly and 4 participants are assigned in block accordingly

Block factor must be concealed to maintain randomisation integrity

Question 13

what does ‘stratified randomisation’ consist of

Answer 13

randomisation is stratified by risk factors/prognostic

reduces the chance of accidental bias especially when a particular risk factor imbalance within the groups will affect the analysis