meta-analysis Flashcards
Describe systematic reviews
a comprehensive assessment of all the relevant research in order to answer a research hypothesis
Which are considered to be the highest standard of systematic review?
Cochraine reviews
Describe a meta-analysis
combines quantitative information from independent studies addressing the same question
When should a meta-analysis be carried out?
when the studies are too small
original studies are not generalisable
to make the effect estimate more precise
When should meta-analysis not be carried out?
when the studies are too small
When the design / demographics / research questions are too diverse
When there is no consistent outcome measure
When the studies are at high risk of bias
Describe fixed-effects meta-analysis
simplest
assumes no heterogeneity between the studies
Wi=i/Vi where vi is variance
Describe heterogeneity
due to different protocols
I squared test is the test for heterogeneity
Describe random effects meta-analysis
weightings here more heavily emphasise the smaller studies compared to the fixed effects model
Describe sub-group analysis
method of investigating heterogeneity
splits participant data into different groups
multiple testing risk
pre-specified > post-hoc
Describe sensitivity analysis
includes comparison of the differences between fixed and random effects models
always should be viewed as exploratory
Describe publication bias
positive findings are much more likely to be published
this leads to an overestimation of the effects
What other types of bias are there?
time lag bias location bias citation bias language bias outcome reporting bias
Describe individual participant meta-analysis
combining raw data from existing studies
good for observational studies
more politically difficult
