Dose-Finding Designs for Early-Phase Clinical Trials and Outcome Dependent Sampling for Longitudinal Studies of Gene-Environment Interaction

study design;early-phase clinical trials;outcome dependent sampling;gene-environment interaction;Statistics and Numeric Data;Science;Biostatistics

Sun, ZhichaoBoonstra, Phil ;

University of Michigan

关键词: study design;    early-phase clinical trials;    outcome dependent sampling;    gene-environment interaction;    Statistics and Numeric Data;    Science;    Biostatistics;   

Others  :  https://deepblue.lib.umich.edu/bitstream/handle/2027.42/133222/zcs_1.pdf?sequence=1&isAllowed=y

瑞士|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

In the first project, we extend the nonparametric biased coin design (BCD) for studying a single agent to a two-stage adaptive procedure that can be easily implemented for dual-agent Phase I trials. The basic idea of our design is to divide the entire trial into two stages and apply the BCD, with modification, in each stage. Through simulations we show that our design is competitive with four contemporary parametric approaches and promotes patients safety by limiting patient exposure to toxic combinations.In the second project, we propose two designs for Phase I/II trials when the dose-efficacy curve plateaus within the dose range of interest. We incorporate multiple sets of pre-specified efficacy probabilities and use Bayesian model averaging to address misspecification in the dose-efficacy pattern. Dose assignment is determined adaptively by maximization of the posterior selection probability among the set of admissible doses. The simulation results demonstrate that our designs identify the OBD effectively and allocate patients around the OBD frequently when compared to a competing approach designed for non-monotonic dose-efficacy curves. To investigate GxE interaction in longitudinal studies, in the third project, we propose exposure enriched outcome trajectory dependent designs that inform sample selection by leveraging individual exposure and outcome trajectory