Inter-examiner variability in orthodontic treatment decisions for Danish children with 'borderline' treatment need

This study explored the variation between examiners in the orthodontic treatment need assessments of fifth-grade children with a borderline orthodontic treatment need. Each of three groups of children with borderline treatment need (n = 18, 19, and 19, respectively) were examined by one of three groups of orthodontists (33 in each group), whereby each of 56 children had 33 orthodontic treatment need assessments based on a clinical examination. This treatment need determination exercise was subsequently repeated with treatment need determined based on study casts and extraoral photographs. The proportion of positive treatment decisions based on the clinical examination was 49.3, 49.6, and 52.5 per cent, respectively, and 45.7, 46.3, and 50.5 per cent, based on the model assessments. There was a considerable disagreement between examiners in the treatment need assessments, whether assessments were based on a clinical examination or on a model-based case presentation. The average percentage agreement between two orthodontists for the treatment need based on clinical examination was 69, 66, and 61, respectively, corresponding to mean kappa values of 0.38, 0.32, and 0.22. When the model-based assessments were considered, the average percentage agreement between two orthodontists was 62, 58, and 69, respectively, corresponding to mean kappa values of 0.25, 0.16, and 0.37. Linear regression analysis of the orthodontists' treatment propensity as a function of their gender, place of education, years of orthodontic treatment experience, type of workplace, and place of work showed that only the orthodontic experience was influential for the model-based treatment propensity [β = 0.34 per cent/year (95 per cent confidence interval = 0.01-0.66)].