Selection of question items for screening patients with temporomandibular disorders and estimation of their validity

Abstract

There have been no questionnaires in epidemiologic studies that estimated cross-validation of question items for screening temporomandibular disorders (TMD). The purposes of this study were to select question items for screening of TMD, and to assess their validity.
Methods: We analyzed 2, 360 dental patients (TMD prevalence rate: 11.7%) who visited any of four dental treatment facilities during one year from October 2005. Cross-validation, criterion-based validation, and estimation of cut-off value for diagnostic accuracy for TMD were calculated.
Results: Four question items (five-point numeric rating scale) were selected from 20 predetermined items by factor analysis followed by Mokken analysis. The validity was confirmed by construct-validity, cross-validity, criterion-related validity, polychotomous item response theory and Cronbach'α. The cut-off value of the total value of the four items was 8.5, indicating that the sensitivity was 0.746 and the specificity was 0.811.