An independent longitudinal retrospective analysis of the timing, number, and cost of tooth restorations was conducted for all subjects (n= 313) in the age cohort of a municipal health center, in which two consecutive field trials of up to 3 years on the caries-preventive effect of the use of xylitol chewing gum were conducted (Ylivieska study, 1982-85). A computer program calculated the post-eruptive timing of the first restorations on each permanent tooth surface, and survival analysis methods were applied before, during, and after the trials. The cost of restorations was then calculated. The total number of new restored surfaces was 4.0 per child in the xylitol group and 9.3 in the controls during the decade after onset of the trial. The number of restorations on the occlusal surfaces of the first molars leveled off at 40-50% in the xylitol group and at about 60-70% in the control group. Plateaus for the second molars were at 30-40% and 60-70%, respectively. The time lag for the first restorations was longer in the xylitol group for the proximal and other smooth surfaces. A marked difference in favor of the xylitol group was observed in the proximal surfaces of the upper incisors (for example, the mesial surface of the upper mesial incisor for girls: log-rank chi-square = 11.1,p<0.001). The preventive effect was most pronounced in teeth that had erupted during the trial. This analysis indicated that participation of subjects in the xylitol chewing gum trials resulted in a significant reduction in the number of first restorations and hence in costs during the decade after the onset of gum use.