13.6.2.1 Controlling for confounding
13.6.2.2 Combining studies
13.6.2.3 Analysis of heterogeneity
13.6.2.4 When pooling is judged not to be appropriate