YSPH Biostatistics Seminar - "Clustered Sequential, Multiple, Randomized Trials for Developing Cluster-level Adaptive Interventions: Estimation and Sample Size Considerations"

NOTE: BIS 526 students are required to attend in person (47 College St., Room 106A). All others are requested to attend via Zoom.

Speaker Daniel Almirall, Ph.D.

Title- "Clustered sequential, multiple, randomized trials for developing cluster-level adaptive interventions: Estimation and sample size considerations"

Abstract

Individual-level adaptive interventions, also known as dynamic treatment regimens, are used to sequence treatment decisions based on a person’s static and changing health information. We introduce cluster-level adaptive interventions (CAIs): pre-specified decision rules guiding the provision of intervention at the cluster level, but with the goal of improving outcomes for individuals within a cluster. We present clustered sequential, multiple-assignment, randomized trial designs (SMARTs), which are used to develop optimized CAIs. In a clustered SMART, clusters are randomized multiple times, at critical intervention decision points; and primary outcomes are at the individual-level. We develop a weighted least squares regression approach to compare the CAIs embedded in a cluster-randomized SMART. We also develop a sample size calculator for designing cluster-randomized SMARTs when the goal is to compare two CAIs that begin with different first-stage interventions. The methods are illustrated using the Adaptive School-based Implementation of CBT Trial (ASIC), which is currently underway. The goal of ASIC is to develop a CAI to improve the adoption/implementation of cognitive behavioral therapy (CBT) by school professionals across Michigan.