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.