Caries risk analysis: considering caries as an individual time-dependent process

To estimate the probability that a tooth does not develop a cavitated lesion (= end-stage caries) within time periods of interest. Additionally it was of interest to assess the heterogeneity between patients, in relation to the time periods required for their teeth to develop a cavitated lesion.

A prospective study using patient data of a clinical pedodontic course.

Forty-one patients who had been followed for at least 3 years were selected. Before initiation of therapy, affiliation to a specific risk group was checked. Nine individuals who had belonged to a high-risk group on test were excluded. Of 1968 teeth, 1378 were eligible for data evaluation. The statistical techniques applied originate from time-to-event analysis methods, including a parametric frailty approach providing estimates of the relative caries risk for each individual to develop a cavitated lesion.

In the selected group of patients the probability that a tooth remained free of a cavitated lesion within five years ranged from 0.89 to 0.95 applying a 95% confidence interval. In the present sample the true caries risk of the patients to develop a cavitated lesion varied considerably although such variation was not expected from the individual examinations at baseline.

The frailty model accounts sufficiently for the heterogeneity of each patient. The applied time-to-event methods facilitated reliable estimation of both the probability that a tooth remains free of a cavitated lesion within a certain time and the relative caries risk of an individual to develop a cavitated lesion.