Imprecision in orthodontic diagnosis: reliability of clinical measures of malocclusion.

The study examined the reliability among seven orthodontists in judging dental and facial aspects of malocclusion in a screening of elementary schoolchildren. Data included measures typically recorded during a clinical orthodontic examination: facial assessment of skeletal/incisor relationships and individual measures of morphologic malocclusion. Interexaminer reliability data were collected on 52 children. Pairwise comparisons between orthodontists were made using exact percent agreement and agreement within one category. Kappa statistics and one-sided Z-tests were used to evaluate observed agreement compared with agreement that would be expected by chance. Median Kappa statistics indicated that the reliability of maxillary and mandibular anteroposterior positions, incisor exposure, interlabial gap, and maxillary crowding was poor (K < 0.40). Acceptable reliability existed for mandibular anterior crowding, facial convexity, overbite, overjet, and molar classification (median Kappas ranged from 0.48 to 0.72). Excellent reliability existed only for evaluating the presence of a posterior crossbite (K = 0.79). The results caution that the language of clinical orthodontic diagnosis is imprecise.