Meta-analysis of cluster-randomized trials

In cluster-randomized trials, interventions are randomly allocated to clusters of eligible participants or multiple sites within the same participant. Hence, the unit of allocation is the cluster or within-participant site rather than the participant.1 For example, 2 education programs are randomly allocated to classrooms in which the unit of allocation is the entire classroom and not the individual students within classrooms. In orthodontics, multiple brackets are bonded on clusters of teeth to the same patient.