Statistical Analysis of Cumulative Serious Adverse Event Data From Development Safety Update Reports

Abstract

Background: Monitoring safety in clinical trials by regulatory authorities and sponsors involves the clinical review, often subjectively, of large data sets of different types of safety information which may require considerable resources. Methods: This study investigated a means of statistically guided clinical review of safety data provided in the Cumulative Table of Serious Adverse Events of a Development Safety Update Report (DSUR). A simple statistical approach that treats every adverse event as independent of all others and uses a reference prior, which avoids infinite estimates of relative risk but does not unduly influence posterior inferences, was used with fixed rules of relative risk to identify a serious adverse event preferred term as a potential risk. Results: This simple model, using cumulative serious adverse event (SAE) data from 5 DSURs, identified a small group of potential risks that included some not reported by the sponsor as well as most of those reported by the sponsor in the DSUR Summary of Important Risks. Conclusions: The method provides a systematic and objective approach to analysis of cumulative SAE data that could help to identify potential risks that need further investigation by a regulatory authority or sponsor.