Yale School of Public Health
Designing Cluster ePCTs in Dementia Care Dyads to Assess Treatment Heterogeneity
Dr. Plourde is a biostatistician and an assistant professor of biostatistics at the Yale School of Public Health, a
faculty member of the Yale Center for Analytical Sciences, and a collaborator with the Yale Program on Aging.
Before joining the Yale faculty, she was a postdoctoral associate with the IMPACT Collaboratory in the Design
and Statistics Core. Her methodological research focuses on the design and analysis of pragmatic clinical trials
in aging and dementia, with particular emphasis on stepped wedge cluster randomized trial (SW-CRT) designs
that are increasingly used in pragmatic studies. Through this Career Development Award, she seeks to address
design-stage challenges arising from complex correlation structures, including power and sample size
determination for cluster randomized trials with multiple levels of clustering and multivariate outcomes.
Pragmatic trials of dementia care dyads often use cluster randomization, creating multiple sources of
correlation that complicate study design and analysis. Interest in designing cluster trials to detect
heterogeneity of treatment effects is increasing, yet existing methods do not accommodate dyadic
outcomes. This project will develop methods and practical tools to design cluster randomized trials
involving dementia care dyads that can detect meaningful subgroup-specific benefits even when an
overall average treatment effect is not observed. This Career Development Award will provide Dr. Plourde
with training and mentorship in dyadic trial theory and analysis, as well as in methods for designing trials
to detect heterogeneity of treatment effects, supporting her development into an independent
methodological investigator advancing the design and analysis of dementia care dyad trials. This training
will support the following Specific Aims: (1) To develop new methods for designing cluster randomized
ePCTs with dementia care dyads to detect treatment effect heterogeneity, and (2) To develop new
methods for designing stepped wedge ePCTs with dementia care dyads to detect treatment effect
heterogeneity. The methods developed in this award will strengthen the ability of dementia care ePCTs to
detect meaningful treatment effects within specific dyad subgroups, even when an overall average
treatment effect is not observed. This work will advance statistical approaches for designing pragmatic
trials of dementia care interventions involving dyads.