A comparison of three continuous scales used to determine the prevalence of clinically significant dental fear

Objective: To assess the level of agreement and concordance of three different dental anxiety and fear scales and determine whether using commensurate cut-points reduces discrepancies in the estimated prevalence of high dental fear. Methods: A representative sample of the Australian adult population completed a computer-assisted telephone interview (CATI) with additional data obtained on a random sample of 25% of those completing the CATI. Participants completed three measures of dental anxiety and fear: the Dental Anxiety Scale, the Index of Dental Anxiety and Fear, and a single-item question. Cut-points were determined to equivalize the range of possible score responses defining a case of high dental fear for each scale. Results: A total of 1084 people completed the CATI and returned a questionnaire (response rate = 71.4%). Prevalence estimates ranged from 7.3% to 28.4% depending upon both the scale and cut-point used. The overall percent agreement between the scales was high, but the overlap between the scales in those people determined to have high dental fear was considerably lower. Using different cut-points to define a case significantly altered the sensitivity and specificity of the scale in predicting dental avoidance, problem-oriented visiting and fear of lack of control. Conclusions: Equivalizing scale cut-points does reduce variation in prevalence estimates of high dental fear across scales, but variation remains because of differences in the nature of, and content covered by, each scale. Dichotomizing scores on continuous scales to obtain prevalence estimates of high dental fear should be appropriately justified, correctly interpreted and supplemented by results describing the distribution of scores.