Keywords: Bayesian, historical controls, extrapolation, paediatric trials
Rare diseases and paediatric populations present several challenges and opportunities for clinical trial design and analysis, due to practical and ethical constraints on sample size. Last year, the EMA and FDA both held public workshops to discuss ways of leveraging data from adults or other paediatric populations to inform regulatory decision-making for paediatric medicines, and also established a joint working group on rare diseases with a remit that includes sharing best practice in the design and analysis of clinical trials in small populations. Significant challenges also exist in other settings such as areas of high unmet need, targeted therapies and personalised medicine, where recruiting patients for clinical trials can present severe difficulties due to ethical, logistical and patient burden issues. In this presentation we consider “small” trials where the feasible sample size that can be recruited has low power to demonstrate efficacy using a conventional designs. We will discuss the use of Bayesian designs using informative priors based on relevant external data to increase the power and precision of such trials, and propose a range of operating characteristics based on clinical trial simulation that can be useful to evaluate and compare designs. For example, it has been argued that the weight given by regulators to avoiding a type I error versus avoiding a type II error should not be equal across disease and clinical indications. We focus on the (weighted) average probability of an error (type I or type II) as a metric to calibrate trial designs in this setting. We also present various prior and posterior summaries of the available historical, current and total evidence which can be used to help sponsors and regulators assess the strength of the prior assumptions and the extent to which the current trial data can influence the final posterior inference. Methods will be illustrated by some recent case studies covering both paediatric settings borrowing efficacy data from adults, and confirmatory trials in adults borrowing historical controls.