Marginal Methods for the Design and Analysis of Cluster Randomized Trials and Related Studies

March 24, 2022
9:00 am to 10:00 am
Virtual

Event sponsored by:

Biostatistics and Bioinformatics

Contact:

LaDonna Huseman

Share

Xueqi Wang, PhD Candidate

Speaker:

Xueqi Wang, PhD Candidate

Cluster randomized trials (CRTs) are used to study the effectiveness of complex or community-level interventions across a diverse range of contexts. To provide meaningful population-level intervention estimates, marginal models are typical used to analyze CRTs and other studies for which clustering of outcomes arise, such as individually randomized group treatment (IRGT) trials. Using four motivating studies, this dissertation addresses unanswered statistical questions in the design and analysis of CRTs and IRGT trials within the marginal modeling framework. Specifically, we develop sample size methods for CRTs with four-level structure and for longitudinal IRGT trials, develop methods for the analysis of small CRTs with time-to-event outcomes, and provide a cautionary tale about the analysis of CRTs with variable cluster sizes. The proposed methods are assessed through simulation studies and applied to real data examples.

Mentor: Liz Turner

Zoom Link: Please contact ladonna.huseman@duke.edu for details on how to join.