Evidence-appraisal glossary
Cluster randomized trial
A cluster randomized trial randomly assigns whole groups (clusters), such as clinics, schools, or villages, rather than individual people, to the intervention or control arm. Everyone within a cluster gets the same assignment. It suits interventions delivered at a group level, or where randomizing individuals risks contamination between arms.
Also called: Group randomized trial, Cluster RCT, Cluster-randomised controlled trial.
A cluster randomized trial (CRT) randomizes intact groups instead of individuals. Researchers choose this design when an intervention naturally acts on a whole group (a hand-hygiene policy, a clinician-training program), when treating individuals separately would let the intervention leak across arms (contamination), or for administrative convenience.
When reading a CRT, check three things. First, how many clusters were randomized: with few clusters, chance imbalance between arms is common, so look for stratified or matched allocation and baseline comparisons. Second, whether the analysis accounts for clustering. People in the same cluster resemble one another (measured by the intracluster correlation coefficient), which shrinks the effective sample size (the design effect) and widens confidence intervals; correct analyses use multilevel or mixed models or GEE, not simple individual-level tests. Third, watch for recruitment bias, since participants are often enrolled after their cluster's assignment is known. The CONSORT cluster extension lists these safeguards.
This is a plain-language methodology definition for reading research. It is general education, not medical advice.