Figure 2.

Forest plot of secondary outcomes at 20 weeks.

Adjusted difference between arms (and confidence limits) divided by pooled standard deviation at baseline and multiplied by −1 if lower scores are associated with an improvement.