Induction of cytochrome P450 enzymes is associated with an increased risk of clinical drug drug interactions (DDIs), particularl y those interactions involving CYP3A. The risk of a clinical DDI presents both a safety and efficacy concern, particularly when considering increasing polyp har macy in an aging population. As clinical DDI studies are costly and time consuming, it is important to ensure predictions of in vivo induction potential from in vitro data are as accurate as possible. In addition to fold change and other static models for risk assessment, correlation methods such as calculation of the relative induction score (RIS) can be used to predict the magnitude of CYP3A4 clinical induction risk from in vitro data. This involves calibrating the RIS values for a set of known inducers against their clinically measured induction (AUC ratio) within each hepatocyte donor. Cyprotex has generated RIS data sets for multiple hep ato cyte donors. To understand the accuracy of RIS predictions compared to other basic methods including R3 , six additional test compounds with clinically available CYP3A4 induction data were assessed, and models evaluated on number of false negatives and false positives and correlation of predic ted AUCR with observed AUCR for quantitative prediction of clinical induction risk.