Page 61 - ESHRE2019

5. Incomplete outcome data
Common errors
“There were many missing data therefore high risk”
“There were many more missing data in the control group compared to the intervention group therefore high risk”
Neither are necessarily a bias, need to consider the reasons for missing data/drop-outs.
Note: Protocol violations are not necessarily patient drop-outs.
5. Incomplete outcome data
LOW RISK
• No missing data
• Small numbers of drop-outs
• Large number of drop-outs but reasons balanced across groups/not related to
intervention or participants outcome
UNCLEAR RISK
• High drop-out rates and reasons not reported
• Unclear how many women actually randomised/analysed
HIGH RISK
• High drop-out rates and reasons not balanced
56
PRECONGRESS COURSE 17 I VIENNA, AUSTRIA – 23 JUNE 2019 59