Statistical considerations related to the use of caries model systems for the determination of clinical effectiveness of therapeutic agents.

Caries model systems bear great promise for the eventual determination of clinical benefits of therapeutic agents without the time and expense involved in the conduct of a full-fledged clinical trial. However, before such models can routinely be applied for such purposes, there are several issues which must first be resolved, including the consideration of a number of statistical matters. These include the assessment of the validity and reliability of the models to be so used. A procedure presented by Proskin, Chilton, and Kingman provides a quantitative means by which a conclusion concerning clinical effectiveness can be derived on the basis of data obtained through the use of a model system given that certain fairly general underlying conditions prevail. The present paper provides some discussion and recommendations of the author related to statistical matters concerning the use of models, and discusses the derivation and logic of the Proskin, Chilton, and Kingman procedure.