Non-prescription antibiotic dispensing in community drug retail outlets in Central Gondar, Ethiopia: a mixed-methods study

Antibiotics are prescription-only medicines in Ethiopia, yet community drug retail outlets may provide an important route of access without a valid prescription. Evidence on observed non-prescription antibiotic dispensing in Central Gondar is limited. To quantify non-prescription antibiotic dispensing in community drug retail outlets in Central Gondar, Ethiopia, and explore contextual factors that may help interpret observed dispensing practices. An explanatory sequential mixed-methods study was conducted from March to May 2024. The quantitative component used standardized simulated-client encounters in 120 of 134 eligible community drug retail outlets. Each outlet received one encounter involving one of three standardized clinical scenarios: acute uncomplicated cough/upper respiratory symptoms, non-bloody diarrhea, or mild urinary symptoms. A structured questionnaire assessed pharmacy personnel’s knowledge, attitudes, and perceptions. The qualitative component comprised interviews with 25 participants, including pharmacy professionals, clinicians, and regulatory officials. Bivariable and multivariable logistic regression were used to examine factors associated with observed non-prescription antibiotic dispensing. Because only 29 outlets were outcome-negative, Firth penalized logistic regression was additionally used as a sensitivity analysis. Antibiotics were supplied without a valid prescription in 91 of 120 assessed outlet encounters (75.8%). A total of 116 antibiotic items were supplied during these 91 encounters; amoxicillin accounted for 49 (42.2%), ciprofloxacin for 26 (22.4%), and metronidazole for 14 (12.1%). Patient assessment occurred in 24/120 (20.0%) encounters, while counseling was provided in 11/91 (12.1%) encounters in which antibiotics were supplied. In conventional multivariable logistic regression, non-prescription dispensing was associated with unlicensed outlet status (AOR 3.45, 95% CI 1.64–7.29), peri-urban location (AOR 2.21, 95% CI 1.04–4.71), and absence of a licensed pharmacist (AOR 2.76, 95% CI 1.32–5.74). Firth penalized logistic regression was used as a sensitivity analysis because of the sparse outcome-negative category. The Firth estimates were similar in magnitude and direction to those from the conventional model: unlicensed outlet status (AOR 3.32, 95% CI 1.55–7.25; p = 0.002), peri-urban location (AOR 2.16, 95% CI 1.00–4.76; p = 0.050), and absence of a licensed pharmacist (AOR 2.65, 95% CI 1.24–5.67; p = 0.012). The fourth covariate showed no evidence of an association (AOR 1.18, 95% CI 0.56–2.49; p = 0.660). Overall, the Firth analysis yielded findings consistent with the conventional regression model, supporting the robustness of the observed associations despite the sparse outcome-negative category. Qualitative findings highlighted perceived regulatory challenges, economic and competitive pressures, and customer expectations for antibiotics. Non-prescription antibiotic dispensing was frequently observed during standardized simulated-client encounters in community drug retail outlets in Central Gondar. The findings identify regulatory oversight, professional staffing, and community expectations as potentially relevant areas for future antimicrobial stewardship research. The study did not evaluate interventions, and the observed associations should not be interpreted as causal.