Targeting community-level drivers of antimicrobial resistance in sub-Saharan Africa: the effect of a community-based intervention bundle on household transmission of Extended Spectrum Beta-lactamase-producing E. coli in rural Burkina Faso - a cluster randomised trial

BackgroundIn sub-Saharan Africa (sSA), invasive antimicrobial-resistant infections often originate from community-level acquisition. We assessed whether a behavioural intervention bundle targeting sub-optimal antibiotic use and hygiene practices reduced household-level acquisition of extended-spectrum beta-lactamase-producing E. coli (ESBL-E). MethodsWe conducted a cluster-randomised controlled trial in 22 village clusters in Nanoro district, Burkina Faso. We enrolled 12 randomly selected households per cluster to assess intervention impact on ESBL-E household-transmission. The intervention comprised three rounds at three-month intervals and combined WHO AWaRe-based educational feedback for formal and informal medicine providers with a community-wide WASH and antibiotic-use behaviour change campaign. Consenting household members provided stool samples before, during, and after intervention rollout, alongside a pre-post household WASH survey. We estimated intervention effects on ESBL-E acquisition using Bayesian Markov models. Cox frailty models assessed associations between WASH exposures and acquisition. ClinicalTrials.gov, NCT05378880. FindingsBetween Oct 11, 2022, and Feb 19, 2024, 1203 individuals were enrolled. At baseline, 57{middle dot}3% (346/604) of control and 48{middle dot}6% (291/599) of intervention household members were colonised. Pre-intervention acquisition incidence was 3{middle dot}8 per 100 person-days (95% credible interval [CrI] 2{middle dot}0-9{middle dot}9) in the intervention group and 3{middle dot}5 (95% CrI 1{middle dot}8-9{middle dot}6) in the control group. The intervention did not change the risk of ESBL-E acquisition in months 1-6 (hazard ratio [HR] 1{middle dot}02, 95% CrI 0{middle dot}78-1{middle dot}31), while we estimated a reduction in ESBL-E acquisition from months 6-9 (HR 0{middle dot}82, 95% CrI 0{middle dot}56-1{middle dot}14). Acquisition risk was higher in the rainy season (peak HR 1{middle dot}73, 95% CI 1{middle dot}49-2{middle dot}00), while improved sanitation was associated with lower risk (HR 0{middle dot}77, 95% CI 0{middle dot}59-1{middle dot}00). InterpretationFindings, though inconclusive, were consistent with a modest intervention-related reduction in ESBL-E incidence. Higher acquisition rates associated with the rainy season and poor sanitation highlight the need to tackle environmental drivers of AMR transmission in addition to antibiotic use in rural sSA. FundingJPI-AMR CABU EICO grant number: JPIAMR2021-053 Research in contextO_ST_ABSEvidence before this studyC_ST_ABSA systematic review of antimicrobial stewardship interventions in both community and hospital-settings published in 2019 found that only 23% (190/825) reported microbiological outcomes, underscoring a lack of evidence on how stewardship interventions affect antimicrobial resistance (AMR). Since then, this gap remains, and particularly for sub-Saharan Africa (sSA). Here, invasive AMR infections are frequently community-associated, with household transmission considered a dominant pathway for community-level AMR acquisition. Two systematic reviews on community level transmission of AMR bacteria, including one published in 2025, reported ESBL-E acquisition rates of 0{middle dot}17-0{middle dot}29 per 100 person-days, with half of individuals clearing their colonisation within 3-4 months. However, all of the included 11 studies were from high-income contexts. The Drivers of Resistance in Uganda and Malawi (DRUM) study provided important One Health insights into human, animal, and environmental reservoirs of ESBL-E and Klebsiella pneumoniae, and found high ESBL-E prevalences up to 60%. This study however, was observational and did not evaluate interventions, and to date, has not quantified community-level transmission. The published community-level interventions in sSA have largely focused on formal healthcare providers or prescribers to reduce sub optimal antibiotic use. A recent scoping review identified only seven intervention studies targeting general communities in sSA, highlighting a lack of rigorous evaluations of community-centred stewardship and AMR mitigation efforts more broadly and outside formal healthcare settings. None of these studies measured microbiological outcomes or effects on community transmission of resistant bacteria. Added value of this studyThe CABU-EICO trial is, to our knowledge, the first cluster-randomised AMR intervention in a rural, low-income setting to quantify the effects of an intervention targeting both providers and communities on community-level ESBL-E transmission-dynamics. Using repeated stool sampling from household members and a continuous-time multi-state modelling framework, we estimated household-level ESBL-E acquisition and duration of colonisation, and found evidence for a reduction in ESBL-E transmission following the intervention. Additionally, we quantified seasonal patterns in the risk of ESBL-E acquisition in our West African setting, showing a peak during the rainy season, despite reportedly lower antibiotic use during this time of year. By moving beyond prevalence-based outcomes and antibiotic-use metrics alone, our intervention evaluation provides a statistically more efficient and mechanistically informative framework for evaluating AMR interventions. Implications of all the available evidenceThis study shows that household-level transmission of ESBL-E is substantial in rural sSA and markedly higher than estimates from high-income settings, with clear seasonal peaks during the rainy season and increased risk associated with poor sanitation. Together with recent One Health genomic evidence demonstrating frequent transmission between humans, animals and the environment in Eastern Africa, these findings suggest that community-level AMR dynamics are driven by both antibiotic selection pressure and environmental exposure pathways. Effective AMR control in similar settings will therefore require, similar to our approach, integrated One Health strategies that combine antibiotic stewardship, with structural and environmental interventions, and that incorporate transmission as well as acquisition outcomes to fully capture intervention impact.