Comparison of surveys and use records for quantifying medically important antimicrobial use in 18 U.S. Beef Feedyards

The objective of this study was to evaluate agreement between medically important antimicrobial use metrics derived from in-person surveys of feedyard management as opposed to metrics derived from production unit-level antimicrobial use records. Survey respondents were asked to estimate values which would allow calculation of the metrics of regimens per animal year (Reg/AY) and milligrams of antimicrobial per kilogram of liveweight sold (mg/kg-LW). At the study level, values were calculated by antimicrobial class within the use categories of in-feed use, control of bovine respiratory disease (BRD) and individual animal treatment. At the feedyard level, values were calculated by total overall use and total use within use category. Feedyard level correlation coefficients between survey and record values for total use were 0.76 (p = .0004) and 0.73 (p = .0009) for Reg/AY and mg/kg-LW, respectively. Correlation coefficients for use category within metric ranged from 0.25 (p = .3224) to 0.65 (p = .0061). Comparing feedyard level survey and record values for total Reg/AY and mg/kg-LW using a Wilcoxon signed-rank test resulted in p-values (95% CI) of 0.3247 (-1.06, 0.25) and 0.7019 (-14.49, 10.387), respectively. Evaluation of comparative rankings as total use by metric indicated that for Reg/AY, only two of the five top-ranked feedyards were consistent between the two data sources. The relationship for mg/kg-LW demonstrated the lowest two and highest three values were consistent between sources; however, the ranking similarities appear to markedly decline in the middle ranks. This report demonstrates that survey-based antimicrobial use data may closely reflect summary values determined from records across multiple beef feedyards. However, individual feedyard relationships between their record and survey values vary widely and ranking by survey may lead to different conclusions as to highest and lowest use than ranking by data collected from use records.