Results
Thirty participants were included in the pilot study. Sixty-three percent were female, and their ages ranged from 18 to 62 years (median, 48 years). Half of the patients (50%) had completed university studies, and high school and primary school were the highest levels of education completed by 40% and 10%, respectively. The number of missing items from the CSOC was 0.
All the patients found the instrument understandable (100% comprehensibility). However, 5 patients expressed concern about items 14 and 15, as they believed they did not account for their individual circumstances. These items inquired about work and academic activities, and the patients who raised issues with these items were either unemployed or retired. As a result, the research team included recreational activities and basic activities of daily living in the language of these items. The internal reliability for the 20 items was high (Cronbach α coefficient =0.85). The Cronbach α coefficient showed no improvement when individual items were deleted, meaning that no items negatively affected the internal consistency of the CSOC.
Table 2 contains the results of the expert ratings of relevance for each item. Items 5 and 6 were given an I-CVI of 0.63, meaning that they should be considered for elimination. Items 4, 9 and 13 received an I-CVI of 0.75, which indicates that further consideration should be given to these. After it was confirmed that eliminating these items did not affect internal consistency, the research group decided to exclude items 5 and 6. The members of the research group reviewed items 4, 9, and 13 (I-CV=0.75) and finally decided to maintain these in the second phase of the validation process. The OS-CVI was 0.85 and if we eliminate items 5 and 6, the OS-CVI increases to 0.88. One expert suggested eliminating item 5, which also had a low I-CVI. The expert panel suggested adding 3 items: (1) Presence of purulent or thick saliva; (2) Difference between parotid and submandibular gland; and (3) Use of drugs to relieve symptoms. The research group decided not to incorporate these additional items. The presence of purulent or thick saliva is similar to item 5, which was not considered to be representative. Differences noted in the affected gland was not considered appropriate within a QOL questionnaire. Finally, use of drugs to relieve the symptoms was included as part of item 18.
Once the CSOC was modified in light of analytical findings and expert feedback, the final version comprised 18 question (Supplementary Table 3) (total score 0-72).
One hundred and twenty subjects completed the presialendoscopy CSOC. Seventy-three percent were female, and their median age was 48.3 years. One hundred fifteen subjects (96%) completed the postsialendoscopy CSOC during the follow-up examination. The number of missing items from the pre and postsialendoscopy CSOC was 0. The time required tofill out the CSOC was 5.7 minutes (2-15 minutes) and 4.5 minutes (1-15 minutes) for pre and postsialendoscopy CSOC respectively.
The reliability determined by test-retest washigh (interclass correlation coefficient= 0.86, 95% CI, 0.77-0.91). The internal consistency for the 18 items was very high for both pre and postsialendoscopy CSOC (Cronbach α coefficient =0.90 and 0.94, respectively). Cronbach α coefficient showed no improvement when individual items were deleted.
Significant differences were observed when the total CSOC score was compared between 2 different groups of patients according with the frequency of sialadenitis episodes during a week (fewer than 2 episodes or more than 4 episodes).A greater difference was observed in postsialendoscopy CSOC (table 3).
Figure1 shows the correlation coefficient between total CSOC score and the different dimensions of SF-36 and VAS. The correlation with the SF-36 dimensions was negative meaning that lower values of CSOC are equal to higher values of SF-36. The correlation with VAS was positive.
The mean score for presialendoscopy CSOC was 28.6 (range, 3–61) and for postsialendoscopy CSOC it was 7.98 (range,0-41).