Review of Recent Methodological Developments in Group-Randomized Trials: Part 1-Design

Area of Science:

• Biostatistics
• Epidemiology
• Clinical Trial Design

Background:

• A 2004 review by Murray et al. summarized methodological developments in group-randomized trials (GRTs).
• Significant advancements in GRT design and analysis have occurred in the subsequent 13 years.
• This work serves as a companion to a review focusing on analytical developments.

Purpose of the Study:

• To review and highlight methodological developments in the design of group-randomized trials (GRTs) over the past 13 years.
• To update the comprehensive review of GRT methodologies from 2004.
• To discuss both established and emerging GRT designs.

Main Methods:

• Review of methodological literature published in the 13 years following the 2004 Murray et al. review.
• Discussion of developments in traditional GRT topics such as clustering and matching.
• Exploration of novel designs including constrained randomization, stepped-wedge, pseudocluster, and network-randomized GRTs.

Main Results:

• Identified advancements in established GRT design principles.
• Detailed new randomized designs that offer alternatives to standard parallel-arm GRTs.
• Highlighted designs like stepped-wedge, pseudocluster, and network-randomized GRTs that necessitate accounting for clustering in design and analysis.

Conclusions:

• The field of group-randomized trial design has evolved significantly since 2004.
• New designs provide flexible alternatives to traditional parallel-arm GRTs.
• Proper accounting for clustering remains critical in both the design and analysis of various GRT formats.