Outcome stability of orthopaedic treatment of skeletal Class III malocclusion: A study of prognostic agreement between three models

Sa？etak Introduction: Long-term stability of the treatment outcome of skeletal Class III malocclusion is always not achievable, thereforeseveral prediction models of stability of orthopaedic treatment of the Class III malocclusion have been proposed. Aim: This cross-sectional study was aimed at the evaluation of the prognostic agreement and association between three differentprediction models based on skeletal parameters recorded on lateral cephalograms. Materials and methods: A total of 75 subjects (34 females and 41 males, age range 7-11, mean age 9.1±1.2years) having skeletal ClassIII malocclusion were included in the study. Prediction models reported by Baccetti et al. 2004 (prediction model 1), Moon et al.2005 (prediction model 2) and Yoshida et al. 2006 (prediction model 3), were considered. Percentage agreement and unweighted kappa coefficient evaluated the agreement between the prediction models in terms of stable or unstable cases. Multiple regressions were run to evaluate the association between the absolute scores obtained from each of the prediction model (prediction scores). Results: Percentage agreement ranged from 77.3% (models 2 and 3) to 89.3%, (models 1 and 2) and corresponding unweighted kappa coefficients ranged from 0.099 (models 1 and 3) to 0.205 (models 2 and 3). The prediction score from model 3 was significantly associated with those from both the other models; on the contrary, prediction scores from models 1 and 2 were not significantly associated. Conclusions: Agreement in terms of stability or instability outcome derived from these models is only partial and not satisfactory,even though significant association between the prediction scores denotes a common biological meaning of the models