Table 2.
Challenges and lessons learned from large-scale Canadian feedlot cattle AMU projects.
| Design |
| Surveillance allows for emerging trends to be identified |
| Targeted research is critical prior to, and after surveillance system implementation to optimize data collection, summarization, reporting and dissemination |
| A specifically designed and dedicated system is required to efficiently collect and store individual-animal administered and in-feed AMU data at the national level |
| Piloting is critical to identify meaningful data points, without overburdening data suppliers |
| Surveillance design must be flexible so changes in feedlot production can be successfully managed |
| Surveillance should include pathogens of both public health and animal health importance to benefit the most stakeholders |
| The ability to compensate producers and veterinarians for their time and access to inventory or sites encourages ongoing participation |
| Sustainable, long term funding to support human and infrastructure resources and system maintenance is needed |
| Sampling and data collection |
| Schedule sampling so diagnostic infrastructure is not overwhelmed |
| Individual-animal administered AMU data access and compilation is straightforward because computerized systems have been specifically designed for this |
| In-feed AMU data access and collection and compilation can be time consuming because feed data collection systems have not been designed specifically for this – multiple indirect data sources often required to estimate in-feed AMU |
| Sample collection at standard commercial handling timepoints promotes compliance and improves data accuracy |
| Composite samples are comparable to individual animal samples for E. coli AMR |
| Stakeholders asked to supply/upload AMU data need to understand the data formats ahead of time so that administrative changes can be made to facilitate collection and compilation |
| Prescriptions, dispensing records and AMU data do not measure the same things |
| In-feed cohort exposure assessment (particularly in the last of feeding) is difficult due to present production practices (animal mixing and sorting) |
| Compilation of feedlot AMU data aggregated to the lot-level has been the most useful to date |
| Metrics and indicators |
| Standardization per 100,000 animals was important for appropriate AMU data interpretation |
| The number of animal daily doses has been the most useful metric to date for Canadian feedlot systems |
| Average weight of heifers and steers (~360 kg) was the best weight estimate to use for approximating actual AMU in the Canadian feedlot system |
| Subset data for individual-animal administered AMDs may be representative of census data at feedlots, depending on the goal/objectives |
| Summarizing AMDs of low human health importance separately from those of moderate/high importance seems most useful and least biased compared to other countries |
| Dissemination |
| Site-specific as well as aggregated result reports provide value to producers for ongoing participation |