Effect of different diagnostic thresholds on dental caries calibration

OBJECTIVES: (a) To analyze the intra- and inter-examiner reproducibility (reliability) of a calibration trial, at different diagnostic thresholds of dental caries; (b) to verify the accuracy (benchmark validity) though sensitivity (S), specificity (SP), positive (PPV) and negative predictive (NPV) values. PARTICIPANTS: A group of dental examiners (n=11), who had previous experience in epidemiological surveys and six to seven-year-old children. Children were selected according to the dmft and dental caries activity. METHODS: Theoretical and clinical training and calibration exercises were arranged for a total of 28 hours. WHO criteria including the active initial lesions (IL) were used. MAIN OUTCOME MEASURES: WHO and WHO+IL diagnostic thresholds according to tooth and dental surface. RESULTS: Excellent mean results of intra and inter-examiner Kappa values were found for both diagnostic thresholds, according to tooth and surface, during the calibration phase. The most relevant errors were related to IL diagnosis and to the first permanent molars. When assessed against a benchmark examiner, moderate to high validity values were observed (0.71-1.00), with some loss mainly for sensitivity and positive predictive value, when including IL. CONCLUSION: It was possible and feasible to use the proposed methodology of this study in epidemiological surveys, even with the inclusion of IL. However, further examiner calibration studies are still needed in order to improve and establish a methodology of calibration with this new diagnostic threshold.