Figure 4: The effectiveness of Kampo formulae on symptoms other than the target; the secondary endpoints. The effectiveness of the Kampo formulae was evaluated by changes in intensity of the symptoms from baseline to the end of administration, which measured by VAS. Differences in effectiveness between the formulae for symptoms other than the target symptom are shown. Data are means (dots) and 95% confidence intervals (bars). The effectiveness of the formulae was not statistically significant, except for Keigai-rengyo-to, which worsened general fatigue by a marginal significance (P =  .07 by ANOVA with Bonferroni correction).