Increasing the sensitivity of caries clinical trials by applying logistic modelling to specific teeth.

An alternative statistical approach is proposed for analysing data from longitudinal clinical trials. Caries increments on those teeth which are known to be particularly susceptible to caries for the population under study are represented by a linear logistic probability model. Expression of the caries increments in the form of a probability model should lead to a better understanding of the relative importance of the various factors which contribute to the caries increment in the population under study. When this model was applied to caries increments on the permanent second molars of children aged 11-13 yr at the baseline of two 3-yr trials, the sensitivity of statistical analysis was up to five times greater than that obtained with traditional statistical methods.