Theory-Based Quantification of Behavioural Determinants of Antibiotic Prescribing in Primary Care in Vietnam Using Linked Prescribing and Survey Data

Background: Antibiotic overprescribing in primary care is driven by complex behavioural determinants, yet these remain under quantified. Understanding such drivers is critical for designing effective antimicrobial stewardship (AMS) interventions.<br><br>Methods: We linked routine prescribing records with a cross-sectional prescriber survey from 94 commune health centres in northern Vietnam. Behavioural factors were mapped to the Theoretical Domains Framework (TDF), and logistic regression examined associations between each factor with overall antibiotic prescribing and use of ‘Watch’ antibiotics (WHO AWaRe classification), adjusting for prescriber characteristics.<br><br>Results: We analysed 123,726 acute respiratory infection consultations by 127 prescribers. Antibiotics were prescribed in 97.8% of visits, and 76.2% of respondents underestimated their own proportions. &nbsp;Ten TDF domains were represented. High confidence in refusing antibiotic requests when not necessary (adjusted odds ratio [aOR] 0.45, 95% CI 0.22 - 0.95) and prior antimicrobial resistance (AMR) training (aOR 0.34, 0.16 - 0.70) were associated with reduced overall prescribing. For Watch antibiotics, confidence in refusing antibiotic requests (aOR 0.46, 0.25 - 0.83) and answering patient questions (aOR 0.47, 0.25 - 0.89) were associated with lower prescribing, whereas AMR training was associated with higher use (aOR 2.50, 1.16 – 5.37).<br><br>Conclusion: Prescriber self-efficacy in patient communication and prior AMR training influence both the quantity and quality of antibiotic use in Vietnamese primary care. AMS strategies should extend beyond knowledge transfer to strengthen patient-centred consultation skills, promote adherence to evidence-based guidelines, and implement audit-and-feedback mechanisms to reduce unnecessary prescribing.