Dentist‐specific variation in diagnosis of caries – a multilevel analysis

BACKGROUND: There are few studies on practice variation within dentistry. This contrasts with medicine where numerous studies exist. A major finding from the field of medicine is that the size of the physician-specific effects depends on the degree of uncertainty in diagnosis. The physician-specific effects are small for diseases where the diagnostic criteria are unambiguous. Conversely, the effects are large for diseases where the diagnostic criteria are less unambiguous. OBJECTIVES: To assess the size of provider-specific variation in the diagnosis of caries among children and adolescents in Norway and to determine whether this variation depends on uncertainty in diagnosis. METHODS: Data on caries diagnosis for 709 611 children and adolescents aged 6-18 years were analysed using multilevel regression. Level-1 was patients and level-2 was public dental officers and dental hygienists. Caries was measured according to the following localization of the lesion: in the outer half of the enamel, in the inner half of the enamel, in the outer third of the dentine, in the middle and inner third of the dentine, to the pulp. The degree of uncertainty in diagnosis is expected to be least the deeper into the dentine the lesion goes. Our sample included 87.5% of all individuals aged 6-18 years. RESULTS: The provider-specific variation, measured as the intraclass correlation coefficient, ranged from 15% for caries lesions localized in the outer half of the enamel to 2% for caries to the pulp. CONCLUSIONS: The size of provider-specific variation in the diagnosis of caries is fairly low. The size of the variability is dependent on the level of diagnostic uncertainty, which is coherent with the practice style hypothesis.