Fig. 1.

Forest plot of the overall treatment effect (log odds ratio) for the Guillain-Barre syndrome (GBS) example. The inverse-variance random-effects method is applied in combination with the seven heterogeneity estimators. The between-study standard deviations τ are presented alongside each estimator The confidence intervals are calculated as $\hat{\theta }\pm {\hat{\sigma }}_{\theta }\cdot Z_{1-\alpha /2}$. The Mantel-Haenszel analysis is plotted as well.