The 'Cariogram' is an interactive PC program for caries risk evaluation. It takes into account the interactions between caries-related factors and expresses a graphic assessment of the risk. The aim of this study was to assess the caries risk in schoolchildren using the Cariogram and to evaluate the program by comparing the caries risk assessments with the actual change in DMF. A 2-year prospective study on 446 schoolchildren, 10-11 years old, was conducted. At baseline, data on general health, diet, oral hygiene and use of fluoride were obtained. Saliva analyses included mutans streptococci and lactobacilli counts, buffer capacity and secretion rate. DMFT and DMFS were calculated from records and bitewing radiographs. Scores were entered and caries risk was assessed. Re-examination for caries was done after 2 years. The children were divided into 5 groups according to the assessed caries risk at baseline. Where the Cariogram predicted a 0-20% (high risk), 21-40%, 41-60%, 61-80% and 81-100% (low risk) chance of avoiding new lesions, 8, 35, 42, 73 and 83%, respectively, had no new lesions 2 years later. Logistic regression analyses were carried out. When the Cariogram was included, only two factors, the Cariogram (p