What is the importance of meta-analyses? (3)
A primary study:
- May not have a sufficient sample size of power
- Cannot be generalised and is not accurate enough
- Differences in definitions of disease, variations of drugs or drug doses, study design
What are the aims and procedures of a meta-analysis? (8)
1) Combine the results of independent trials
2) Calculate the average ‘pooled’ effect
3) Evaluate overall strength and direction
4) Provide unbiased conclusion
The aims:
More statistical power due to larger combined sample size, so more reliable and less biased
Takes into account standard error and weights with effectiveness.
What are the types of heterogeneity in a meta-analysis? (4)
Methodical heterogeneity
- differences in the way studies are conducted
Clinical heterogeneity
- differences in participant characteristics and patient population
What are some biases in meta analyses? (6)
- Publication bias
- Selective outcome reporting bias
- Language bias
- Low income and high income countries
- Single/ multi centre trials
- Industry sponsorship
What is publication bias and how can it be avoided? (5)
Not all studies conducted are published. A study is more likely to be published if the result is positive or if it is a large study
To avoid it:
Researchers should be aware of it, and encourage publication of all trials
Search unpublished trials in a trial registry
