Episode 24, “STEP Statistical Modeling”
The latest "In the Interim…" episode features Dr. Elizabeth Lorenzi and Dr. Amy Crawford discussing the statistical design of the STEP adaptive platform trial for acute stroke.
STEP—an NIH-funded master protocol—employs Bayesian change point models and extensive simulation to define where endovascular therapy (EVT) benefits—or does not benefit—specific patient subgroups.
Key technical details:
● Modeling treatment effect boundaries along the NIH Stroke Scale using Bayesian change point modeling
● Importance of trial simulations to determine operating characteristics across patient populations before trial launch
● Evolving from 1D to 2D adaptive modeling as new evidence and clinical hypotheses arise
● Refining models through ongoing, simulation-driven work between statistical and clinical teams
Step in to the full episode today: https://www.berryconsultants.com/resource/24-step-statistical-modeling