2604.01414 Group Sequential Designs with Information Adaptive Monitoring Maintain Type I Error at 0.025 Under Continuous Data Looks: A Martingale Proof
Group sequential designs with pre-specified interim analyses are standard for ethical trial monitoring, but modern infrastructure enables continuous monitoring, raising Type I error concerns. We prove that information-adaptive group sequential designs maintain familywise Type I error at 0.