Page 51 - ESHRE2019

2. Subverting random allocation
“Randomised controlled trials appear to annoy human nature – if properly conducted, indeed they should” Schulz 1995
Anecdotal accounts of subversion (Schulz 1995)
• Holding envelopes to a bright light
• Feeling for different weight of envelope
• Opening the envelope prior to randomisation and re-sealing
• Obtaining a few allocations at once, even if from a third party source
• Breaking into filing cabinets to find original sequence! Most accounts of subversion occur with envelopes
Block randomisation
Keeps numbers allocated to each group similar to optimise study power
BUT
Introduces an element of predictability into the sequence
1
Control
2
Intervention
3
Intervention
4
Control
5
Control
6
Intervention
7
Intervention
8
Control
9
Control
10
Intervention
11
Control
12
Intervention
13
Intervention
14
Intervention
15
Control
16
Intervention
17
Control
18
Control
46
PRECONGRESS COURSE 17 I VIENNA, AUSTRIA – 23 JUNE 2019 49