Background: Evidence on caries risk assessment (CRA) and recall intervals are limited in terms of caries prevention. Aim: To assess the effectiveness of a program on the incidence and regression of initial caries lesions. Design: A total of 296 children aged 1-12years old were assessed by calibrated examiners for Gingival Bleeding Index, Dental Plaque Index, dmf-t/DMF-T Index, initial caries lesions, and caries lesion activity. Children were classified as low, moderate, and high caries risk with different recall interval visits. Statistical analysis included Cox regression and Kaplan-Meier curves. Results: The mean (SD) time of total follow-up for the sample was 11.5(5.5). Higher dmft index at baseline showed a higher risk of new initial lesions (HR=1.93; P<0.0001). Higher number of active initial lesions, at baseline and during follow-up visits, is a higher risk predictor for new initial lesions (HR=9.49; P<0.0001), as well as for no arrestment of active lesions during follow-up (HR=1.32; P<0.0001). Each follow-up visit attended presented a 77% lower risk of initial lesions. The majority (94.8%) of patients did not show new initial lesions. Conclusions: The Program is effective on reducing the incidence and promoting regression of initial caries lesions in children.