Multivariate extreme value modelling of laboratory safety data from clinical studies

Authors


Harry Southworth, AstraZeneca, Clinical Information Science, FE2 A/2, Parklands, Alderley Park, Macclesfield, Cheshire, UK.

E-mail: harry.southworth@astrazeneca.com

Abstract

Generally, in the interpretation of clinical safety laboratory data, it is extreme values that indicate potential safety issues. We illustrate the application of multivariate extreme value modelling to such data. Applying the methods to a clinical trial dataset, we find unexpected extremal relationships that have potentially important implications for the interpretation of such data. Copyright © 2012 John Wiley & Sons, Ltd.

Ancillary