Behavioural antimicrobial stewardship interventions in outpatient settings: a systematic review and meta-analysis of physician-targeted approaches.

BACKGROUND: Antimicrobial stewardship (AMS) interventions are increasingly implemented in outpatient settings to improve the rational use of antibiotics and mitigate antimicrobial resistance. OBJECTIVES: To assess the effect of physician-targeted behavioural AMS interventions on outpatient antibiotic use. METHODS: Data Sources: PubMed/MEDLINE, Cochrane Library, APA PsyNet, Epistemonikos (PROSPERO CRD420251110732). STUDY ELIGIBILITY CRITERIA: Randomized controlled trials (RCTs) and non-RCTs evaluating AMS interventions. PARTICIPANTS: Primary care physicians. INTERVENTIONS: AMS interventions with at least one behavioural component. Assessments of risk of bias: The revised Cochrane risk-of-bias tool and the Joanna Briggs Institute tool were used. Methods of data synthesis: We categorized AMS interventions along two dimensions, into seven broad types and by number of components (single, combined, multicomponent). Adjusted and crude estimates were calculated in random-effects meta-analyses to assess overall, recommended/first-line, and non-recommended/inappropriate antibiotic use, stratified by study design. Univariable random-effects meta-regression was conducted to explore whether intervention effects varied by AMS components. Publication bias was assessed using Egger's test, funnel plots, and sensitivity analysis. RESULTS: Eighty-five studies were included, 43(50.6%) RCTs and 42 (49.4%) non-RCTs. Among RCTs, behavioural AMS interventions reduced overall antibiotic use (0.89; 95% CI: 0.84-0.94) and non-recommended/inappropriate use (0.61; 95% CI: 0.49-0.75) and increased recommended/first-line use (1.35; 95% CI: 1.16-1.58). Corresponding estimates for non-RCTs were (0.88; 95% CI: 0.77-1.01), (0.54; 95% CI: 0.47-0.62), and 1.52 (95% CI: 1.34-1.71). In meta-regression, each additional intervention component was consistently associated with a greater reduction in non-recommended/inappropriate use (RCTs 13%; 95% CI: 0.76-0.99) and overall use (non-RCTs 7%; 95 CI: 0.86-0.99). Incentives and accountability, communication and patient engagement also showed statistically significant reductions in overall and inappropriate use among RCTs, but were based on fewer studies. CONCLUSIONS: Outpatient physician-targeted behavioural AMS interventions reduce inappropriate antibiotic prescribing and increase recommended/first-line use, with more modest effects on overall volume; intervention effectiveness increased with the number of components.