Methods for dealing with unequal cluster sizes in cluster randomized trials: A scoping review.

Area of Science:

• Biostatistics
• Clinical Trials Methodology
• Epidemiology

Background:

• Cluster-randomized trials (CRTs) often have participants varying across clusters.
• Existing methodological literature predominantly assumes equal cluster sizes.
• This assumption can impact statistical performance goals in CRTs.

Purpose of the Study:

• To conduct a scoping review of methodologies for unequal cluster size CRTs.
• To identify research gaps in the current literature.
• To inform future methodological development in CRT design and analysis.

Main Methods:

• Systematic search of EMBASE, Medline, Google Scholar, MathSciNet, and Web of Science databases.
• Inclusion of English-language articles on unequal cluster size CRT methodology published until March 2021.
• Data extraction on paper focus, CRT type, parameter values investigated, and conclusions.

Main Results:

• 79 out of 5032 identified papers met inclusion criteria.
• Most papers focused on parallel-arm CRTs (76%) and stepped-wedge CRTs (18%).
• Methodology for unequal cluster sizes is well-developed for parallel-arm CRTs with Gaussian/binary outcomes, but limited for other designs and outcome types.

Conclusions:

• Significant methodological work exists for unequal cluster size parallel-arm CRTs, but synthesis is challenging due to parameter variation.
• There is a notable paucity of research on unequal cluster sizes for other CRT designs and outcome types, especially binary outcomes.
• Further methodological development is needed to address these identified gaps in CRT research.