Using Clinical Trials to Learn about Drug Safety: Potentials and Pitfalls

This presentation explores how clinical trials are used to assess drug safety, highlighting both their strengths and limitations. It discusses the advantages of controlled conditions and detailed data collection, but also the challenges posed by restrictive study populations, missing data, patient withdrawals, and variability in exposure time. Key issues include how to handle confounders, interpret adverse event rates, and ensure fair comparisons across treatment groups. The concept of estimands is introduced as a way to better define causal treatment effects in the presence of intercurrent events, and alternative trial designs—such as single-arm, adaptive, and decentralized studies—are considered for their impact on safety assessment.