Adaptive designs offer the possibility of delivering more efficient and effective drug development decision making. However, these designs can sometimes be technically very complex and confusing to the non-statistician. In this case study several Phase II/III adaptive design options are evaluated and a strategic recommendation made that was accepted by the regulatory authority.
Effective Phase II design is essential in that it often supports critical Phase III decision making. In this case study, Phase II design options were examined to provide the best balance between practical feasibility and effective decision making.
Sometimes regulatory authorities require large randomised, controlled safety studies to rule out potential serious risks associated with drug products. The design of such studies is not always straightforward, however, and if the risk of concern is very rare then extremely large studies can result. This case study shows how statistical argument was used to arrive at a feasibile study design adequate to rule out a clinically worrying increase in risk.
Rare events pose challenges for effective meta-analysis. If unexpected rare events are observed in a development programme, understanding the degree of excess risk associated with a drug is often critical. Traditional fixed effects approaches are illustrated by example, along with random effects and Bayesian approaches.