Reliability and limitations of permanent tooth staging techniques

Reliability, or repeatability, of permanent tooth staging techniques is usually expressed as Cohen's Kappa. This single value obscures information about the quantity and allocation of disagreements. In this study we assess and compare intra-observer reliability of permanent tooth staging techniques described by Nolla, Moorrees et al. and Demirjian et al. The sample was panoramic radiographs of healthy dental patients made up of 100 males and 100 females aged 6-15 years. All permanent teeth on the left side (excluding third molars) were scored twice. Weighted Kappa and percentage agreement were calculated. Results show Kappa values for all teeth combined as 0.918, 0.922 and 0.938 for Demirjian (number of teeth N = 2682), Nolla (N = 2698) and Moorrees (N = 2674) respectively. A comparison of Kappa values between upper and lower teeth showed marginally higher values for upper incisors and lower molar for all three scoring methods. Small differences in Kappa values were noted between tooth types with the upper first molar having smaller values than other teeth. Percentage agreement ranged from 81 % (Moorrees), 86 % (Nolla) to 87 % (Demirjian). Tooth stage differences between first and second assessments were not more than one stage. Our findings show that Demirjian scoring is marginally more reliable than Nolla or Moorrees. We suggest that (1) data for reliability are tabulated in full to show the quantity and allocation of disagreement between first and second readings, and (2) that the reliability sample is sufficiently large with a wide age range to include multiple different tooth stages.