Diagnosis-associated antimicrobial prescribing patterns in a resource-limited veterinary district setting: evidence from Mongar, Bhutan (2023–2026)

Routine veterinary records in resource-limited settings rarely link antimicrobial prescribing to clinical diagnosis at the individual-animal level, a gap identified in Bhutan’s National Action Plan on Antimicrobial Resistance. To our knowledge, few studies from Bhutan have linked diagnosis, antimicrobial selection, and documented client advice within routine veterinary records. We analysed retrospective clinical records from the district veterinary service in Mongar, eastern Bhutan (the District Veterinary Hospital and 16 Renewable Natural Resources Extension Centres): 8,881 clinical visits and 1,063 antimicrobial line items (January 2023–June 2026). We described prescribing by recorded diagnosis and the documentation of course-completion advice by WHO AWaRe tier, using proportions with Wilson 95% confidence intervals, Cramér's V, and a route-adjusted logistic regression (the single primary model); other analyses were secondary or exploratory. No clinical or microbiological outcome data were available, so results describe prescribing patterns, not appropriateness. Of 8,881 visits, 1,022 (11.51%) received an antimicrobial. Pharmacological class differed by diagnosis system (Cramér's V = 0.26, small-to-moderate): tetracyclines predominated in cattle, penicillins in dogs and cats. Prescribing proportions by diagnosis ranged from 0% to 80% against a baseline of 11.3%, with wide intervals for small groups (urinary tract infection 80.0%, 95% CI 60.9–91.1%, n = 25). Course-completion advice was documented for 55.7% of Access-tier and 23.7% of Watch-tier line items (difference 32.1 percentage points, 95% CI 25.6–37.9). Adjusted for route, the Watch-versus-Access odds ratio was 0.71 (95% CI 0.48–1.04), compatible with a substantial reduction or no difference. Exploratory route-specific odds ratios were 0.45 (95% CI 0.25–0.78) for injectable/other and 1.30 (95% CI 0.68–2.72) for oral items. The food-producing versus companion-animal difference in Watch-tier share reversed if oxytetracycline was classed as Access. In Mongar, the antimicrobial prescribed varied with recorded diagnosis. These are diagnosis-associated prescribing patterns; without clinical or microbiological data they cannot show whether prescribing was appropriate. Course-completion advice was documented less often for Watch-tier items, particularly injectables, but the indicator measures what was written in the record, not whether or how well clients were advised, so this is best read as a documentation gap that structured recording could address.