Cluster trials in implementation research: estimation of intracluster correlation coefficients and sample size

The design considerations specific to implementation research studies are described, focusing particularly on the estimation of sample size requirements and on the need for reliable information on intracluster correlation coefficients for both effectiveness and efficiency outcomes.

Abstract

The cluster randomized trial with a concurrent economic evaluation is considered the gold standard evaluative design for the conduct of implementation research evaluating different strategies to promote the transfer of research findings into clinical practice. This has implications for the planning of such studies, as information is needed on the effects of clustering on both effectiveness and efficiency outcomes. This paper describes the design considerations specific to implementation research studies, focusing particularly on the estimation of sample size requirements and on the need for reliable information on intracluster correlation coefficients for both effectiveness and efficiency outcomes.