Results
Women were enrolled at the colposcopy clinic between August 1, 2017 and January 30, 2018. Of the 20 patients who agreed to participate in the study, three did not fulfil the inclusion criteria. There were no differences in the background variables (Table 1) or baseline pain levels (6.7±2.2 units versus 6.4±3.0 units, respectively; two-tailed Wilcoxon rank test p = 0.686; data not tabulated) of the sham first group and trial first group. We used Cronbach’s alpha test to confirm the statistical validity of calculating the average pain intensity from the various foci under each condition (Table S1). Then, we tested whether the manipulation order affected the difference between the baseline pain level and trial manipulation pain level. For the seven women in the trial first group, the average change in pain intensity was 0.86. However, for the 10 women in the sham first group, the average change was 1.5 (P < .512). Therefore, the manipulation order did not have a significant impact on the change in pain levels. Following these results, data from all subjects were gathered in one group. Then, we compared the average pain levels (Table 2) and showed that posterior fornix pressure led to a statistically significant decrease in the average pain level elicited by the vestibular Q-tip test compared with the sham manipulation pain level and baseline pain level (decreases of 1.12 and 1.13 units in pain intensity, respectively; P = .003). There was no significant difference between the average baseline pain level at baseline and sham manipulation pain level (P = .332), as anticipated. However, the overall decrease in the average pain level during trial manipulation was 18.4% ± 2.2%, which was below the 30% threshold that we considered clinically significant (Table S2). Further examinations of the data revealed that the decrease in the average pain level ranged between 0% and 78%; in other words, there was at least one subject who experienced no influence of trial manipulation and some subjects who experienced a significant influence. Then, we calculated the overall decrease in the average pain level for the 11 patients who reported any degree of improvement during trial manipulation, resulting in an average change of 28.4% ± 2.2%. Five patients experienced more than 30% improvement in their pain level.