Fig. 7.

Funnel plot conducted using RevMan showing that pub-lication bias is unlikely in the BDI baseline and up to 6 weeks post intervention scores due to the symmetry of the plotted points shown. The dotted lines represent the 95% confidence intervals while the middle vertical line is the overall effect. Each study is represented by a dot, with the standardised MD result plotted on the x-axis, and their precision/standard error (SE) on the y-axis.

MD, mean difference; BDI, Beck Depression Inventory.