The present study investigated whether the incidence or prevalence of filled teeth/approximal surfaces at one age could be predictive for the incidence in another period or for the prevalence at the age of 18. The study was conducted in 12-18-yr-olds in Norway. Regression analysis showed that the best prognosis for subsequent incidence of filled teeth/approximal surfaces could be made at the age of 15. By using regression analysis or discriminant analysis it was possible at the age of 15 to predict with high accuracy those who would acquire more fillings than the median at the age of 18. Discriminant analysis with one predictor variable is suggested for clinical use. The variable that discriminated best between above and below median number of new fillings in the period 15-18 yr was untreated lesions in the inner half of the enamel in the approximal surfaces of premolars and molars at the age of 15. From the use of simple prediction tools, it was concluded that individuals at the age of 15 with a low prevalence of filled teeth/filled approximal surfaces and without untreated approximal lesions would be subjected to a low incidence of new fillings until the age of 18.