All ETDs from UAB

Advisory Committee Chair

Inmaculada Aban

Advisory Committee Members

Denise Esserman

Hassan Fathallah-Sheikh

Dustin Long

Fazlur Rahman

Document Type

Dissertation

Date of Award

2021

Degree Name by School

Doctor of Philosophy (PhD) School of Public Health

Abstract

Randomized clinical trials (RCT) are the gold standards for clinical trials as they reduce bias and minimize variability between different arms of a study. One of the drawbacks of these designs is their lack of flexibility to incorporate patient’s preference, which may reduce recruitment rates and/or reduce patient’s tolerance if they receive a non-preferred treatment. Designs incorporating choice allow a subset of participants to choose a preferred treatment. Some of these designs are Zelen’s (1979), Brewin’s (1989), and Rucker’s (Two-Stage RCT, 1989). Current methods to analyze two-stage RCT are based on an analysis of variance (ANOVA) approach that do not allow for inclusion of covariates in the model and they focus on specific outcomes (Normal, Binomial, and Poisson). In this dissertation, we extend the ANOVA approach for the design and analysis of two-stage RCT to a broader class of distributions including the regular exponential family. Using simulations, we evaluate the performance of the proposed methods for time-to-event outcomes following Weibull distributions with complete-case and right-truncated data, and make recommendations on situations when these methods may or may not be the best method to use. One of the major disadvantages of the ANOVA approach is its lack of flexibility to include covariates in the model as well as censoring for survival outcomes. An approach based on likelihood methods can remediate for these shortcomings. In our work, we propose an approach based on likelihood that allows for the inclusion of covariates, censoring, and multiple study arms in the model moving it beyond two-stage RCT to any clinical trial that incorporates patient’s treatment choice. We use extensive simulations targeting multiple outcomes (normal, binary, count, and survival) to compare between the ANOVA and likelihood approaches. We provide practical suggestions on situations when each method would be a better fit for the trial being designed. To illustrate the practical application of the likelihood approach, we apply it to a health education intervention trial, Women Take PRIDE study. Finally, we summarize the contributions of the dissertation, and discuss future research directions.

Included in

Public Health Commons

Share

COinS