Evaluating Meta-Analyses Flashcards
Advantages of meta-analyses
Increased power
Clarify direction of literature
How to avoid publication bias?
One must systematically search for negative trials for inclusion
How to get around the problem of having different types of studies in a meta-analyses?
Clear inclusion criteria
Forest plots
Heterogeneity statistics
What happens if a study has missing data?
Cannot be included in a meta-analyses
When does one explore subgroup findings?
As a way to explain failure to find overall effect - these can be studied in further RCTs
Disadvantage of doing further analysis on subgroups
Potential for confounding and misleading findings
Why does subgroup analysis lead to potential for misleading findings?
Randomisation between treated and control groups does not extend to subgroups
Reason for sensitivity analysis
Process of selection, inclusion and aggregation of data may affect main findings of meta-analysis
What does sensitivity analysis do?
Explores the ways in which the main findings are changed by varying the approach to aggregation
What should good sensitivity analysis do?
Mode the effect of excluding various categories of studies
Examine how consistent results are across sub-groups
What happens if a meta-analysis does not have sensitivity analysis?
Reader has to make guesses about likely impact of selection, inclusion and aggregation of data on the conclusion of the study
Where will line of no difference be on a forest plot if final estimate is standardised mean of Cohens d?
Placed on 0
Where will the line of no difference be on a forest plot if point estimate is relative risk or ratio measure
1
How do we know if a study gives precise results form a forest plot?
Horizontal line across the square will be shorter
What does it mean if a studys line crosses the line of no effect on a forest plot?
Results are not statistically significant
