Page 68 - ESHRE2019

RoB assessments
• Pre-define criteria for each judgment in the Review protocol (appendix)
• Avoid ‘unclear’ where possible
– If you have the information then make a decision
• Two authors independently assess each domain. Best to discuss the protocol first
• Pilot on a few studies first
• Don’t punish the study twice for the same thing
What’s the point?
Prioritise domains for your review
You must incorporate the RoB assessments in the text of the Review
Three possible analysis strategies
1.
2. 3.
Primary analysis restricted to studies at low (or low and unclear*)
– Predefine your criteria in the Protocol, e.g. excluding studies with at least one high risk domain
– Restrict to low risk of bias for certain domains (e.g. allocation). Present multiple (stratified) analyses
risk of bias
Present all studies and provide a narrative discussion of risk of bias *studies at ‘unclear’ risk should not be grouped with ‘low risk’ without a rationale
66
63