Sample size determination
The sample size, using a formula of repeated measures
analysis of variance, was estimated on the basis of pain during
ICS, using a numerical rating scale at 3-month follow-up
and assuming 90% power, 5% of significance, and 15% of
attrition rate. To detect a clinically meaningful difference
between groups of two scores on a numerical rating scale,14
a total sample size of 42 was required.