Activity Number:
|
238
- SPEED: Biopharmaceutical Applications: Trials, Biomarkers, and Enpoint Validation
|
Type:
|
Contributed
|
Date/Time:
|
Monday, July 30, 2018 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Biopharmaceutical Section
|
Abstract #329291
|
Presentation
|
Title:
|
A Location-Adjusted Approach to the Covariate-Adjusted Response-Adaptive Allocation Design in Multi-Center Trials
|
Author(s):
|
Brian S Di Pace* and Roy T Sabo and David C. Wheeler
|
Companies:
|
Virginia Commonwealth University and Virginia Commonwealth University and Virginia Commonwealth University
|
Keywords:
|
multi-center clinical trials;
adaptive randomization;
spatial statistics;
Bayesian;
personalized/precision medicine
|
Abstract:
|
Covariate-Adjusted Response-Adaptive designs are used to allocate patients in clinical trials and adjust the treatment assignment for a set of covariates. This allocation strategy is useful when treatment responses and/or effects differ with respect to these covariates. Treatment differences could also be attributed to differences in social determinants of health that often, especially if unmeasured, manifest as spatial heterogeneity amongst the patient population. We propose the Location-Adjusted Response-Adaptive (LARA) approach to account for spatial variability in both treatment response and effectiveness. A Bayesian geographically adaptive regression model will incorporate spatially-varying regression coefficients and unstructured and/or spatially-structured random effects to calculate conditional probabilities utilizing treatment and residential information. These success probabilities are used to update weights for assigning patients between treatment groups as subjects are accrued to allocate patients most appropriately. We compare the LARA approach with several existing allocation strategies for a variety of scenarios where treatment success probabilities vary spatially.
|
Authors who are presenting talks have a * after their name.