Page 43 - ESHRE2019

THE MOST IMPORTANT THING • Assessing risk of bias is hard.
• It cannot be reduced to an algorithm.
• You need either a) some understanding of study
design or b) a willingness to learn (quickly)!
What is Risk of Bias (RoB)?
Risk of bias (RoB) is a systematic bias within a trial (internal validity). Bias makes us misjudge the size (and possibly, direction) of the treatment effect in the study sample. Not the same as:
Imprecision
• Caused by random error (not systematic error)
• Highly precise results (narrow CI) can be biased
• Sample size affects precision, not bias!
• •
Quality
Bias can occur in well-conducted studies
Not all methodological flaws introduce bias
Poor reporting
• Good methods may have been used but not well reported
38
PRECONGRESS COURSE 17 I VIENNA, AUSTRIA – 23 JUNE 2019 41