klusterrandomiserte

Klusterrandomiserte, also known as cluster-randomised trials, are a type of experimental study in which the unit of randomization is a group or setting rather than an individual. Clusters can be schools, hospitals, communities, or workplaces, and all individuals within a chosen cluster receive the same allocation. This design is used when the intervention is delivered at the group level, when individual randomization would risk contamination between participants, or when policy or environmental changes are implemented through institutions.

In planning a klusterrandomiserte trial, the number of clusters and the average cluster size influence statistical power. Because participants within a cluster tend to resemble each other, outcomes are correlated. Statistical analysis must account for this intra-cluster correlation, using methods such as mixed-effects models or generalized estimating equations, or cluster-level summaries. The sample size must be inflated by the design effect, DE = 1 + (m−1)ICC, where m is average cluster size and ICC is the intraclass correlation coefficient.

Common designs include parallel cluster trials, where clusters are randomized to intervention or control, and stepped-wedge trials, in which clusters transition from control to intervention at different time points. Blinding is often challenging.

Advantages include reduced contamination, practical implementation of group-level interventions, and clearer evaluation of real-world effects. Limitations involve the need for more clusters to achieve adequate power, potential baseline imbalance with few clusters, complex analysis, and ethical considerations around consent at cluster and individual levels.

Reporting follows guidelines such as CONSORT extensions for cluster-randomized trials. These emphasize explicit description of cluster units, randomization, and analysis that accounts for clustering.