A meta-analysis groups results of several trials to increase statistical power and provide an overall pooled effect estimate.
When different categories were evaluated separately but presented in the same analysis, a test for heterogeneity is required to show that there is little difference between them.
A commonly used heterogeneity test is the Q statistic, which results in a corresponding P value. (Cohran's Q)
In general, the null hypothesis (H0) denotes that there is no difference between the groups studied. A small significant P value (<0.05) suggests that there is a difference and that H0 should be rejected.
However, large nonsignificant P value (>0.05) suggests that there is no difference between the groups studied so H0 cannot be rejected.
Another commonly used heterogeneity test is the I2 index (derived from the Q statistic; I² = 100% x (Q-df)/Q. I²).
An I2 index = 0 represents no heterogeneity.
I2 index = <25% = low heterogeneity
I2 index = 50% = moderate heterogeneity
I2 index = 75% = high heterogeneity
Simple or basic regression
Meta-synthesis
