Background: To evaluate the clinical course and interventions required during two years of follow-up of dental surfaces of deciduous molars diagnosed, and consequently treated, by two different strategies: diagnosis made by clinical examination alone or associated with radiographs.Methods: This is a secondary analysis of a two-arm randomized clinical trial with two parallel groups related to the diagnostic strategy for caries detection in preschool children. 216 children (3 to 6 years old) were followed-up for two years. All dental surfaces were diagnosed by visual inspection and later, through radiographic assessment. Baseline treatment was made in accordance with the results obtained by visual inspection performed alone or combined with radiographic method, considering the allocated group. Dental surfaces with no restoration needs, or those restored at the beginning of the study were followed-up for two years. The treatment decision was made according to the allocated group. The outcome was the occurrence of failure (a new caries lesion or a restoration replacement) during the follow-up. Results: 4,383 proximal and occlusal surfaces of deciduous molars in 216 preschool children were diagnosed and treated according to the abovementioned diagnostic strategies and followed-up for 24 months. The assessment of radiographs made change the initial decision reached by visual inspection in about 30% of the surfaces when all types of interventions were considered. However, most disagreements occurred for initial lesions, where radiographs tended to underestimate them. Discordances between methods occurred in less than 5% of all surfaces when considered lesions requiring operative treatment. For discrepancy cases, the placed interventions guided by following the radiographic results did not present less failures against those made following only visual inspection. As a matter of fact, the use of radiographs in the diagnostic strategy for caries detection in children brought more harms than benefits due to the occurrence of false-positives, overdiagnosis and lead-time bias.Conclusions: Simultaneous association of visual inspection and radiographic assessment for caries detection in preschool children causes more harms than benefits, and therefore, visual inspection should be conducted alone in the regular clinical practice.Trial registration: clinicaltrials.gov platform: NCT02078453, registered on 5th March 2014