Figure 1

Brief pain inventory response. Proportions of patients with (a) 30% and (b) 50% reduction in the average brief pain inventory score. ***P < 0.001 vs. placebo. With both of the outcome measures, between-groups (duloxetine-treated vs. placebo-treated) analysis showed that the group of patients receiving duloxetine exhibited significantly greater numbers of responding patients than among those receiving placebo, but this was only true among the patients in the more severe tiredness group. There was no significant treatment-by-tiredness subgroup interaction between the three subgroups (P > 0.1).