Design: Cluster-randomised controlled trial, 44 communities
Population: Community providers and patients in Burkina Faso and DR Congo
Primary outcome: Change in Watch-group antibiotic use
Effect: Adjusted prevalence ratio 0.33 (95% CI 0.14-0.78)

Rising use of WHO Watch-group antibiotics in community settings drives resistance. CABU-EICO, a cluster-randomised trial in 44 villages in Burkina Faso and DR Congo, tested a co-created behavioural bundle aimed at every level of community provider, from health centres to pharmacies and informal vendors.

Over three rounds across nine months, the intervention combined community health-education campaigns with provider education and feedback built around the WHO AWaRe antibiotic guidance. Watch-group antibiotic use fell from 26.8% to 17.1% in intervention clusters while rising from 13.4% to 21.2% in controls, an adjusted prevalence ratio of 0.33. Patient-management scores from simulated visits were essentially unchanged, so the fall was not at the cost of under-treatment.

The relevance beyond these settings is the model: engaging the whole chain of dispensers, including informal vendors, rather than hospital prescribers alone, can shift community antibiotic use without harming care.

Cluster-randomised trial in 44 villages in Burkina Faso and DR Congo.
A behavioural bundle targeted all community dispensers, including informal vendors.
Watch-group antibiotic use fell from 26.8% to 17.1% with the intervention.
Use rose from 13.4% to 21.2% in controls (adjusted prevalence ratio 0.33).
Simulated-patient management scores were unchanged, so care was not compromised.

Why it matters

It shows broad-spectrum community antibiotic use is modifiable outside the hospital, where much of the resistance pressure arises.

The statistics, in plain English

An adjusted prevalence ratio of 0.33 means roughly two-thirds lower odds of Watch-group use after accounting for the opposite trend in controls. The wide confidence interval (0.14 to 0.78) reflects a cluster design with relatively few villages, so the size of the effect is uncertain even though its direction is clear.