Abstract

A cluster randomized controlled trial is a research design in which organizational, social, or geographically defined units, rather than individuals, are randomized between or among comparator interventions. Interventions are often targeted at caregivers rather than those receiving care. Applying the definition of a “research subject” found in the U.S. federal regulations governing human subjects research may yield counterintuitive results. We explicate the identification of research subjects by evaluating different types of cluster trials and provide examples to illustrate the relevant concepts. We also present the results of an empirical study of a representative cohort of reports of cluster trials performed in the United States, which confirm that investigators often do not correctly identify the subjects of their research