The effect of cluster size variability on statistical power in cluster-randomized trials.

Stephen A Lauer,Ken P Kleinman,Nicholas G Reich,Andrew R Dalby

doi:10.1371/journal.pone.0119074

Stephen A Lauer, Ken P Kleinman + Show 2 more

Open Access

https://doi.org/10.1371/journal.pone.0119074

Copy DOI

Abstract

The frequency of cluster-randomized trials (CRTs) in peer-reviewed literature has increased exponentially over the past two decades. CRTs are a valuable tool for studying interventions that cannot be effectively implemented or randomized at the individual level. However, some aspects of the design and analysis of data from CRTs are more complex than those for individually randomized controlled trials. One of the key components to designing a successful CRT is calculating the proper sample size (i.e. number of clusters) needed to attain an acceptable level of statistical power. In order to do this, a researcher must make assumptions about the value of several variables, including a fixed mean cluster size. In practice, cluster size can often vary dramatically. Few studies account for the effect of cluster size variation when assessing the statistical power for a given trial. We conducted a simulation study to investigate how the statistical power of CRTs changes with variable cluster sizes. In general, we observed that increases in cluster size variability lead to a decrease in power.

Highlights

The cluster-randomized trial (CRT) is a common study design in public health research, in which interventions are administered to groups rather than to individuals
In situations where dividing a group of individuals into treatment and controls is unethical or impossible, a CRT design retains many of the strengths of an individually randomized study design [1]
Using the results from these simulations, we examined the effect of variability in cluster sizes on statistical power of CRTs and developed simple and concrete quantitative guidelines for researchers who design CRTs with high variability in cluster sizes

Summary

Introduction

The cluster-randomized trial (CRT) is a common study design in public health research, in which interventions are administered to groups rather than to individuals. In situations where dividing a group of individuals into treatment and controls is unethical or impossible, a CRT design retains many of the strengths of an individually randomized study design [1]. By comparing the outcomes of small populations (clusters), CRTs can observe the impacts of interventions on a community as a whole. The number of published articles utilizing CRTs has increased every year since 1997 (See Fig. 1). The Consolidated Standards of Reporting Trials (CONSORT) Group issued guidelines for conducting CRTs in 2004 [2], with an update published in 2012 [3]. One important component of CRT design is the sample size calculation; in which researchers must find the correct number of clusters to achieve sufficient statistical power.

Methods

Results

Conclusion