RESUMO
Aim: To assess inter-examiner reproducibility in the detection of 20 occlusal caries in permanentteeth using three diagnostic codes and criteria: WHO (1997), Nyvad and ICDAS-II. Methods:Three graduate students (G) and two undergraduate dental students (UG) without previousexperience in the use of Nyvad and ICDAS-II were trained by a reference examiner. Examinervalidity was assessed by consensus agreement between the investigators. Two cutoff pointswere used for ICDAS-II and Nyvad to represent reproducibility values: the A- lesion, B- cavitylesion. Results: According to the examiners consensus, the kappa values ranged from 0.71 to0.85 for G group and from 0.85 to 0.95 for UG group (ICDAS-II). For the Nyvad index valuesvaried from of 0.77 to 1.00 (G) and from 0.65 to 0.74 (UG), for the WHO index, values obtainedranged from 0.66 to 1.00 (UG) and 1.00 (G). Using a cutoff A, interexaminers reproducibility(ICDAS-II) ranged from 0.73 to 0.87 (G) and 1.00 (UG). According to Nyvad criteria, the kappavalue ranged from 0.78 to 1.00 (G) and from 0.70 and 0.90 (UG) when compared to theconsensus. The kappa values using the cutoff point B ranged from 0.66 to 1.00 (G), 0.76-0.89(UG) in ICDAS-II and the in Nyvad criteria varied from 0.87 to 1.00 (G) and from 0.65 to 0.88(UG). Conclusions: Reproducibility values ranged from good to perfect. The reproducibilityrevealed precise answers in the occlusal caries lesions diagnosis according to the criteria used.The best use of reliability tools for examiner training was important regardless of the examinersbeing undergraduate or graduate students.