Event Date:
08/28/2020
Description:
Speaker 1: Alex Karanevich, PhD (EMB Statistical Solutions)
Title: Overview and Optimization of an Adaptive Bayesian Two-Stage “Drop-the-Losers” Trial Design
Abstract:
When it is desired to test several treatment arms against a common control arm, a two-stage adaptive design can be more efficient than a single-stage design. Adaptive designs require more statistical support and are less straightforward to plan and analyze than a standard single-stage design. We discuss one such adaptive design, a two-stage “Drop-the-Losers” design, which attempts to pick the best treatment arm and compare it with the control arm. We discuss trial assumptions, power and sample size calculations, and how to optimize allocating subjects among the two stages. We also show how a freely-available R Shiny application can be used to determine these operating characteristics. The software is freely available at https://github.com/stefangraw/Allocation-Power-Optimizer.
Speaker 2: Hengrui Sun, DrPH, MD (FDA)
Title: Pediatric Drug Development and Bayesian Method
Abstract:
Since the establishment of the Best Pharmaceuticals for Children Act (BPCA) and Pediatric Research Equity Act (PREA), significant progress in pediatric drug development has been seen, yet delays in pediatric drug approvals still exist, especially for the younger pediatric patients. Obtaining substantial efficacy and safety evidence of the drug for the children specific diseases or shortening the gap between the adult approval and the pediatric labeling remain challenging. The challenges in the pediatric drug development are mainly due to the special characteristics of the pediatric patient population and the considerations in protecting this population. These challenges manifest in the clinical trials as limited sample sizes and a lack of suitable control groups, which in turn affect the subsequent data analyses and efficacy interpretation. On the other hand, the adult or older age group data are available and hard to ignore sometimes. In this presentation, Bayesian strategies will be discussed, and case examples will be presented.
Speaker 3: Joe Marion, PhD (Berry Consultants)
Title: Design of a Single Arm Trial in a Rare, Progressive Disease Using Natural Histories and a Disease Progression Model
Abstract:
Rare diseases present a variety of challenges for traditional study designs. These difficulties include limited patient populations, heterogeneous disease development, and an urgency to treat that makes randomization to placebo difficult. In this talk, we provide an example design that addresses these issues for studies in Mucopolysaccharidosis (MPS) IIIA, a rare genetic disease that leads to neurodegeneration in children. First, we develop a Bayesian disease progression model (DPM) using natural history data that characterizes the cognitive decay of children with MPS IIIA. We then leverage the DPM to design a single-arm study, estimating the rate at which an investigational therapy slows the progression of the disease relative to the natural history control. The efficiency of this design, relative to a more traditional design, is demonstrated through simulation.
Slides will be stored in http://www.bayesianscientific.org/kol-lecture-series/
Join Zoom Meeting
https://diaglobal.zoom.us/j/2203599026
Meeting ID: 220 359 9026
One tap mobile
+16468769923,,2203599026# US (New York)
+16699006833,,2203599026# US (San Jose)
Dial by your location
+1 646 876 9923 US (New York)
+1 669 900 6833 US (San Jose)
+1 253 215 8782 US (Tacoma)
+1 301 715 8592 US (Germantown)
+1 312 626 6799 US (Chicago)
+1 346 248 7799 US (Houston)
855 880 1246 US Toll-free
877 853 5257 US Toll-free