Study Hypothesis and Testing

For each of the 2 doses of fingolimod, the null hypothesis that no difference in ARR would be found between participants who received fingolimod therapy and those who received glatiramer acetate therapy could be rejected if the observed significance for the between-treatment comparison was P < .05 (after adjustment for multiplicity). The 2 doses of fingolimod were tested against glatiramer acetate, 20 mg, hierarchically through a step-down procedure (ie, the initially approved fingolimod, 0.5 mg, dose was tested against glatiramer acetate, 20 mg, at an unpaired 2-sided significance level of P = .05). If this test result was rejected, the lower dose of fingolimod (0.25 mg) was tested against glatiramer acetate, 20 mg (at an unpaired 2-sided significance level of P = .05). The study was not designed to detect a difference in treatment effect between the 2 doses of fingolimod. Nominal P values were provided for secondary efficacy analyses to assist the evaluation of treatment effect even if the primary efficacy end point did not reach statistically significant differences based on a predefined hierarchical testing procedure. However, the interpretation of results in assessing treatment effect in these cases should be made with caution.