Antimicrobial use (AMU) is widely recognized as the main driver of antimicrobial resistance (AMR), thus calling for efforts to reduce AMU to effectively combat AMR. Therefore, this project aims to investigate the drivers of AMU in the Danish pig production, to guide targeted AMU interventions.
Antimicrobial prescription data in the Danish pig production from 2018 to 2024 was obtained from the Danish Veterinary Medicine Statistic Program database (VetStat), while herd information was obtained from the Central Livestock Registry of Denmark. Data management efforts were aimed at mitigating spurious variation in the outcome variable caused by the inherent uncertainties associated with register-based data errors. Given the cross-classified multi-level data structure – where AMU (measured as Animal Daily Dose per pig-day) is clustered within pig herds, which are grouped in farms overseen by different veterinarians, who in turn are grouped in veterinarian practices – a linear mixed modeling approach was employed to determine the variance in AMU attributable to the different levels. Fixed effects like age (sows, weaners, finishers), herd size, drug dispensing (oral, parental) and production type (production, free-range, and organic herds, as well as breeding and rearing herds) were also included.
In a preliminary analysis of the data from 2022 focusing on weaners, we computed a nested two-level model using the mean AMU prescribed by the most frequent veterinarian in 2022 for each herd. These initial findings indicate that the variation in AMU between herds can primarily be attributed to between herd differences rather than between veterinarian differences, suggesting that farm practices drive AMU in the Danish pig production more than veterinarian practices.
Next, we aim to develop the complete cross-classified multi-level model, integrating all relevant random and fixed effects, allowing for a detailed analysis of the drivers of AMU in Danish pig production.
Josefine Ostenfeld Nielsen’s presentation
