Development of a new chewing function questionnaire for assessment of a self‐perceived chewing function

Abstract

To develop a new unidimensional questionnaire for patients' self-assessment of their chewing function (chewing function questionnaire: CFQ) and to test its psychometric properties. A focus group of 7 dentists and 15 prosthodontic patients generated a pool of 30 relevant items. A 5-point Likert scale (0-4) was used. A focus group reduced some redundant items, and 21 items remained for pilot testing after which additional four items were eliminated. Then, the 17-item questionnaire was tested on a larger sample (200 participants). Factor analysis was obtained to investigate the dimensionality of the questionnaire. Finally the 10 items remained, and the CFQ was unidimensional. To test the psychometric properties, the CFQ was administrated to 224 individuals: 100 patients were already removable denture wearers (RDWs), 24 patients needed a removable prosthodontic treatment, and 100 individuals had natural teeth (NT group). Discriminative validity was tested between the RDWs and the NT groups. Convergent validity and internal consistency were tested on 200 subjects. The internal consistency was assessed by calculating the Cronbach's alpha coefficient and the average interitem correlation. The test-retest reliability was tested on 60 subjects. Responsiveness was tested on 24 patients who received new removable prosthodontic treatment. The initial factor analysis revealed that the items had been grouped in three different dimensions. By eliminating items with complexities greater than one, additional six questions were eliminated. One more item with the lowest correlation was further eliminated. The final questionnaire included 10 items. Discriminative validity showed significant differences between the NT group and the RDWs, as predicted (P < 0.001). Convergent validity was confirmed by the Spearman's rank correlation. A significant positive association (P < 0.001) was found between the CFQ and one general question about chewing difficulties, as well as between the CFQ and five items from the OHIP49 related to chewing. The test-retest reliability showed high intraclass correlation and no significant differences between the two administrations of the CFQ (P > 0.05). The internal consistency showed satisfactory Cronbach's alpha values (0.916 for all subjects, 0.742 for the NT group, and 0.852 for the RDWs). Responsiveness was confirmed by a significant difference between the baseline and the follow-up score (P < 0.001) and a high effect size (0.95). The developed CFQ provides an assessment instrument of the hypothetical construct invoked to explain the chewing function. Psychometric properties are satisfactory demonstrating that the instrument is suitable for the assessment of a self-perceived chewing function.