“Draw a cluster randomized CONSORT flow diagram with 24 sites assessed, 4 excluded, 20 randomized to 10 intervention and 10 usual-care sites, enrolled participant counts, retained sites and participants, and final analysis counts.”
Read the decisions behind it.
Start with sites because they are randomized.
Show enrolled people inside each site arm.
Report sites retained separately from people followed.
State both sites and participants in the final boxes.
Use this form for cluster randomized trials in which practices, schools, wards, or other groups are allocated.
Other examples in this family.
Common questions
What is different about a cluster CONSORT diagram?
It reports both the clusters allocated and the participants recruited or analysed within those clusters.
Should site losses be shown separately?
Yes. A lost site and a lost participant have different implications and should not be combined.
What can count as a cluster?
A clinic, school, ward, practice, community, or any group randomized as one unit can be a cluster.