Insights into toxicology, safety pharmacology, and drug dependence testing: The performance and predictive values of nonclinical tests.

Abstract The pre-marketing testing of drugs and chemicals is now a paradigm of contribution to the safeguard of public health. As pointed out by the ICH concept paper E14/S7B, there is a need for science-based frameworks that allow a better design, conduction, and/or interpretation of nonclinical and toxicity test, particularly, in order for those assays to better influence nonclinical and clinical evaluations, or decisions. Critical issues related to the performance and predictive values of nonclinical tests were highlighted and discussed in this article, specifically, to help pharmacologists, toxicologists, and regulators in the evaluation of the reliability of such testing, and basing the prospective decisions on the true predictivity of selected screening tests. This review addressed two common mistakes in drug and chemical testing, namely, (a) the assumption of either sensitivity and specificity as automatically “predictive,” and (b) the reporting of the predictive values disregarding a truly representative prevalence. This review also discussed a statistical basis for applying for (industry standpoint) or granting (regulatory standpoint) authorization for the waiving of selected nonclinical tests. Furthermore, this review can be guiding for those who are entering into the field of drug and chemical testing. Though application, investigation, and enhancement of the conceptual framework discussed in this review, nonclinical testing is expected to provide a higher contribution to drug and chemical development, regulatory science, and public health.