RESUMEN
BACKGROUND: The difficulties of caries diagnosis and the level of examiner's experience may influence the treatment decision, resulting excessive and inefficient intervention. AIM: This randomized experimental study evaluated caries treatment decisions made by undergraduate (UG) and postgraduate (PG) students when examination was performed through conventional visual clinical examination (CVCE) and when supported by ICDAS (IC) or Nyvad's (NY) criteria. DESIGN: Four UG and four PG initially analyzed, by CVCE, 300 primary incisors and molars surfaces of 25 children aged 5-12 years, and choose between no treatment, non-operative, and operative treatment. Students were randomized between IC (2UG; 2PG) and NY (2UG; 2PG) and, after one week of receiving training, repeated evaluations and treatment suggestions. Reference standard was established by two researchers' consensus. Sensitivity, specificity, ROC analysis, and kappa were calculated. Treatment decisions were analyzed by means of contingency tables. RESULTS: Only sensitivity showed statistical difference (P < 0.05). Operative treatment had high percentage for initial lesions at first evaluation for both criteria; second evaluation showed higher percentage of non-operative treatments for the same lesions. CONCLUSIONS: IC and NY present satisfactory performance in primary teeth by UG and PG, and their use may assistance UG to adopt a less interventionist approach for initial enamel lesions.