Data mining of clinical oral health documents for analysis of the longevity of different restorative materials in Finland.

Evidence-based dentistry has shown that different restorative materials have different survival times. Our primary hypothesis is that this should be revealed from normal dental records by the use of data mining technique and a practice-based dentistry approach analysed in a scientifically sound way. Dental records from 1626 patients and altogether 19,892 restorations in three Finnish age cohorts were analysed. Survival curves (Kaplan-Meier) were drawn for each of the restorative materials. Median survival times for amalgam and resin-based composites were more than 15 years in older cohorts. More than 60% of silicate cement restorations were replaced within 5 years, and more than 50% of glass ionomers within 7 years. There was a significant reduction in the longevity of amalgams in the 1980 cohort. Data mining of digital oral heath documents would be a useful tool to analyse survival curves of new restorative materials in a practice-based manner in real-life conditions.