Inter-rater Reliability of Caries Diagnosis by Three Different Modalities

"Objectives: " This study determined concordance among skilled dental clinicians in the diagnosis of dental caries. Their evaluation of extracted teeth was based on three modalities: 1) Clinical (visual-tactile), 2) radiographic, and 3) DIAGNOdent (Laser fluorescence). "Methods: " This in vitro study utilized 30 extracted premolars and molar teeth. The teeth were brushed with water using a tooth brush to remove debris and mounted in wax for evaluation. Six clinicians examined each tooth, saw the same radiographic films of each tooth and separately utilized a laser fluorescence (DIAGNOdent) device to determine the presence of caries. Data were analyzed to yield Cohen’s Kappa and interaclass correlations coefficients (ICC)."Results: " Over raters, 76% of teeth were considered decayed based on visual and tactile exploration. For ratings based on such exploration, kappa ranged from .20 to .87 (Med= .49) and the ICC was .54. Over raters, 40.6% of teeth were considered decayed using x-ray data. Based on the x-ray data, kappa ranged from 0 to .81 (Median= .24) and the ICC was .24. Over raters, 91.1% of teeth were considered decayed using DIAGNOdent readings. Based on the DIAGNOdent data, kappa ranged from .14 to .78 (Median= .52) and the ICC was .49. "Conclusion: " The DIAGNOdent detected more carious lesions than the visual and radiographic methods. While clinical and DIAGNOdent assessments yielded moderate ICCs of about .5, the x-ray data were much less reliable. However, none of these evaluation modalities yielded values above .6, usually considered a minimum for reproducibility. The study was designed to determine inter-rater accuracy rather than caries accuracy