A total of 500 subjects will provide 80% power to test the hypothesis that the absolute value of the correlation between two measures (e.g., BMQ Necessity and DAS28) is ≤ 0.3 (i.e., H0:│r│ ≤ 0.3 ) versus the alternative that it is \> 0.3 (i.e., H1:│r │ \> 0.3 ) assuming the absolute value of the true correlation is 0.40. A total of 500 will be enrolled to allow for approximately 5% being lost from analysis due to lack of post baseline data.
The Pearson method will be used for correlation between the BMQ total score for Necessity and the DAS28 and between the BMQ total score for Concern and DAS28. The Spearman rank method will be used for correlation between the BMQ total scores and Safety. For these latter analyses, subjects will be classified into one of 3 safety categories: (1) No AEs, (2) Non-serious AEs, (3) Serious AEs. 95% confidence intervals for the correlations will be provided.
