The aims of this study were to test the hypothesis that existing patient records serve as a source of data for survival analysis of dental health and to develop a system for routinely conducted survival analysis of dental health from tooth eruption to first caries attack. Patient records for three cohorts were collected from four towns in Finland: Oulu, Turku, Kuopio, and Kemi. The study population comprised 2103 children born and living in these towns. The Kaplan-Meier curves for dental caries were drawn separately for each tooth and for the maxilla and mandible (for each age cohort). Placements of the first restoration of all teeth in each age cohort were investigated. The 1960 cohort had an extremely high post-eruptive morbidity in all teeth. The 1970 cohort had relatively high immediate post-eruptive morbidity in both molars, but much lower than that of the 1960 cohort. The filling increment rates for second molars were lower than those of the first molars in the 1970 cohort. Characteristic of the 1980 cohort was a rapid increase in the caries increment of the premolars and second molars 4 years after eruption. The results indicate a big difference in tooth-by-tooth dental health in Finland. A huge decline in caries attack was seen from the 1960 cohort to the 1980 cohort, but a deterioration of dental health in premolars and second molars is clearly seen in the 1980 cohort in the 1990s.