ABSTRACT
Cluster randomized trials (CRTs), also known as group randomized trials, have become increasingly common in certain areas of public health research, such as evaluation of lifestyle interventions, vaccine field trials, and studies of the impact of hospital guidelines on patient health. Depending on the primary question of interest and the chosen unit of inference, CRTs can be analyzed using cluster-level summaries or at the individual level using mixed models or generalized estimating equations. Cluster randomized designs usually fall into one of the three main categories: unmatched, matched, or stratified. For each of these three design types, the chapter presents sample size formulas for continuous outcomes, proportions, and incidence rates. In unmatched designs with relatively small number of clusters, it is possible that a chance of imbalance can occur, leading to large values of between-cluster variation and subsequent loss of power.
