A practical guide to simulation for an adaptive trial design with a single interim analysis

Background The demand for adaptive trial designs is growing because of their flexibility and the potential for efficiency gains over traditional fixed designs. Adaptive trials allow planned modifications to the design based on accumulating data. Simulation is imperative in designing adaptive trials because analytical power formulae cannot account for data driven adaptations. Despite their popularity, the uptake of adaptive trials has been slowed by the lack of expertise and availability of training resources. Methods In this tutorial, we demonstrate how to simulate data from a simple adaptive trial with a single interim analysis, summarise the simulations, and use these results to balance the type I error and power to inform the study design and to determine the expected sample size. The simulation code, based on a real trial in hyponatraemia in children, is provided in both R and Stata programming languages. The code is written in modules to improve comprehensibility and enable simple changes to generate a range of adaptive designs. Discussion When using simulation to design an adaptive trial, the simulations must be tailored to the unique design requirements of the trial at hand. We hope that this tutorial will provide a starting point that will make the simulation process more accessible to both statisticians and clinicians.