Abstract
To estimate the 6-month proportion of antibiotic self-medication among adults with recent antibiotic use in Narayanganj Sadar, Bangladesh, and examine associated characteristics. This community-based cross-sectional study enrolled 400 adults from 50 clusters across four sampling strata between January and April 2026. The most recent antibiotic course within the preceding 6 months was the index course. Self-medication was defined by the absence of prescribing involvement from a Bangladesh Medical and Dental Council-registered physician. Design-aware modified Poisson regression estimated adjusted prevalence ratios, accounting for clustering and stratification. Self-medicated antibiotics were classified using the World Health Organization Access, Watch, and Reserve framework. Self-medication occurred in 244/400 participants (61.0%; 95% confidence interval 55.6%-66.4%). Higher prevalence was associated with pharmacy access (adjusted prevalence ratio 1.50, 95% confidence interval 1.29-1.74), poor antibiotic knowledge (1.46, 1.24-1.73), lower educational attainment (1.30, 1.09-1.55), and lack of antimicrobial-resistance awareness (1.22, 1.06-1.41). Watch or Reserve agents accounted for 145/201 classifiable episodes (72.1%) and 117/152 objectively verified episodes (77.0%). Antibiotic self-medication was common among recent antibiotic users. Pharmacy access and limited antimicrobial literacy were associated characteristics rather than demonstrated causal determinants. Higher-priority AWaRe exposure remained substantial.