Development and validation of a MedDRA-based signal detection construct for drug-induced neurobehavioral impairment.

Background: The FDA's Guidance for an industry evaluating drug effects on the ability to operate a motor vehicle describes the need to evaluate drug-induced neurobehavioral impairment. However, the guidance does not provide instruction or the means to implement the analysis. Objective: The proposed Construct provides a means to operationalize the evaluation of adverse event datasets for this disorder. Design: A noninterventional pharmacoepidemiologic methods development study using routinely collected health data. Methods: Preferred terms for the Construct were derived by mapping the domains of concern cited by the Guidance to the corresponding MedDRA terms. Additional terms were derived from the MedDRA preferred terms cited in the warning section of labels from drugs with a warning about driving or operating machinery. The performance of the Construct was evaluated by comparing adverse events for the FDA Adverse Event Reporting System (FAERS) database for drugs considered to be of high risk versus low risk for this condition using descriptive statistics and a nonparametric test. Results: A Construct of PTs (neurobehavioral impairment term list; NITL) was obtained after identification of 55 drugs with the Warning in their FDA-approved label for driving or operating machinery based on drug-induced neurobehavioral impairment. Additional terms not covered in the labels were derived from the five domains of interest cited in the Guidance for a total of 52 PTs. Comparison of the difference in the proportion of FAERS reports with the NITL terms in drugs with and without the relevant labeling was nominally significant ( Conclusion: A MedDRA-based Construct was generated for the selection of adverse events related to drug-induced neurobehavioral impairment. The Construct discriminated high- from low-risk drugs based on publicly available adverse event profiles. Further evaluation with clinical trial data will be needed to facilitate its use for signal detection in safety analysis.