![School of Public Health](../../assets/md5images/35d233b2c9a8d4e92cd89657d26cb679.gif)
Dissertations & Theses (Open Access)
Date of Award
Summer 5-2019
Degree Name
Master of Science (MS)
Advisor(s)
Ruosha Li, Phd
Second Advisor
Hongjian Zhu, Phd
Third Advisor
John Michael Swint, Phd
Abstract
One important goal of the pharmaceutical industry is to evaluate new therapies in a time-sensitive and cost-effective manner without undermining the integrity and validity of clinical trials. Adaptive seamless phase II/III designs (ASD) have gained popularity for accelerating the drug development process and reducing cost. Covariate adaptive randomization (CAR) is the most popular design in randomized controlled trials to ensure valid treatment comparisons by balancing the prognostic characteristics of patients among treatment groups. Although adaptive seamless clinical trials with CAR have been implemented in practice1, the theoretical understanding of such designs is limited. In addition, current approaches to control the Type 1 error rate in seamless trials are based on theories for complete randomization, which may be invalid under CAR and lead to a Type 1 error rate that deviates from the nominal level. Recently, Ma and Zhu (2019, unpublished) established the theoretical foundation for the adaptive seamless phase II/III trial with CAR and proposed a hypothesis testing approach to control the Type 1 error rate in such trials. In the current research, numerical studies were conducted to investigate the feasibility and advantages of the proposed approach in the seamless design with stratified permutated block (SPB) randomization. The simulation results revealed that the newly developed method well controlled the Type 1 error rate around the nominal level, improved the statistical power compared to the standard two sample t-test and increased the number of replications that the best treatment is selected for Stage II of the seamless trial under the SPB design compared to the complete randomization, which could promote its application in practice.
Recommended Citation
Wang, Mengxi, "Numerical Study For Seamless Clinical Trials With Covariate Adaptive Randomization" (2019). Dissertations & Theses (Open Access). 102.
https://digitalcommons.library.tmc.edu/uthsph_dissertsopen/102