Sample size calculation and statistical methods
Because of multiple comparisons, we applied the Bonferroni correction.
Statistical significance was set at 1.67% (α = 0.0167). We expected
statistically significant differences resulting from the baseline-trial
and trial-control comparisons and no difference resulting from the
baseline-control comparison. A difference between the two conditions was
viewed as clinically significant if it reflected a change of at least
30% in the pain level. Fifteen patients were required for this
crossover study. All calculations were performed using IBM SPSS
Statistics, version 25.