7. Random allocation

Question 1

Randomisation can be done at the level of…? (4)

Answer 1

• Participant
• Group or section (cluster)
• Sequent of interventions)
• Start of the intervention

Question 2

What are the goals of randomisation..? (2)

Answer 2

• Every participant has equal chance to be allocated to either group
• The process of allocation should be unpredictable and preferably also concealed

Question 3

Randomisation allocates treatment without bias, when…

Answer 3

Properly designed, conducted and analysed

Question 4

Allocation bias

Answer 4

Systematic difference between participants in how

they are allocated to the treatment groups (e.g. if researchers would allocate or patients could chose)

Question 5

Selection bias

Answer 5

Systematic difference between participants

recruited to the trial and those who are not recruited

Question 6

What question can you ask to detect selection bias?

Answer 6

Was the sample recruited representative for the patient population?

Question 7

Which types of randomisation are restricted?

Answer 7

• Stratification

- Minimisation

Question 8

How does minimisation work?

Answer 8

• Choose the factors important for balancing carefully
• The first participant is allocated using simple randomization
• For each subsequent participant, either which treatment group lead to better balancing in
  chosen factors (computer algorithm)
• For the next to be enrolled participant, the numbers
  of participants with the same characteristics
  will be summed up in each group
• Imbalance is minimized by allocating this person
  to the group with the smallest total OR at random if the totals are the same

Question 9

What can be done to improve unpredictability in minimisation?

Answer 9

Introduction of a random element of 0.8