# Antimicrobial Stewardship in Food Animal Practice: Metrics and Benchmarks ## Key Takeaways - Antimicrobial stewardship in food animal practice hinges on quantitative measurement, necessitating the selection of appropriate metrics like Animal Daily Doses (ADD), Days of Therapy (DOT), and mg per Population Correction Unit (mg/PCU) to establish baselines and track interventions. - Dose-based metrics (e.g., ADD) standardize antimicrobial use by expressing it as defined daily doses per animal or population unit, facilitating comparison across different drugs, while duration-based metrics (e.g., DOT) count treatment days, directly reflecting clinical decision-making. - The correlation between ADD and DOT can be poor, particularly in hospital populations, with ADD often overestimating DOT due to variations in actual dosing and treatment duration, highlighting the importance of consistent metric selection and application. - Treatment incidence, expressed as ADD or DOT per 1,000 animal-days, is a critical population-level metric for standardizing use across different farm sizes and time periods, but benchmarks require identical metric definitions and denominators to be meaningful. - Metric selection is driven by production system, data availability, intended decision-making, and regulatory context, with veterinarians often preferring dose-based metrics for farm-level decisions, underscoring the need for tailored reporting formats. - Benchmarking requires careful consideration of denominators (e.g., animal-days at risk, biomass) and stratification by production class and user status to avoid obscuring significant variations in antimicrobial use between farms. --- Quantitative measurement is the foundation of antimicrobial stewardship in food animal practice. Without reliable metrics, a stewardship program cannot establish a baseline, detect change, or demonstrate the effect of an intervention. This article examines the metrics and benchmarks used to evaluate antimicrobial use in food animal production, the scientific logic that underpins each measurement approach, and the interpretive challenges that arise when comparing data across farms, species, and regions. It is written for veterinary researchers and practitioners who design, implement, or audit stewardship programs and who need a rigorous framework for selecting and interpreting quantitative indicators. The article addresses a specific question: which metrics should a food animal stewardship program adopt, and how should the resulting data be interpreted against meaningful benchmarks? It covers the conceptual basis of dose-based, duration-based, and population-level metrics, the relationship between antimicrobial use measurement and resistance surveillance, and the practical considerations that determine whether a metric is fit for purpose in a given production system. Clinical treatment guidelines are outside the scope of this article. ## At a Glance | Parameter | Consideration | |---|---| | Primary use metrics | Animal daily doses (ADD), days of therapy (DOT), mg per population correction unit (mg/PCU) | | ADD vs DOT correlation | Poor in hospital populations, ADDs often overestimate DOTs | | Data sources | Treatment records, farm self-report, practice management software, sales data | | Benchmark comparisons | Require identical metric definitions and denominators | | Reporting audience | Veterinarians prefer dose-based metrics for farm-level decisions | | One Health context | Human, animal, and environmental health sectors linked through WHO and CDC frameworks | | Metric selection drivers | Production system, data availability, intended decision, regulatory context | ## The Conceptual Basis of Antimicrobial Use Metrics Antimicrobial use quantification serves two distinct purposes. The first is epidemiological: measuring population-level exposure to support investigations of the relationship between use and resistance. The second is operational: providing individual farms or practices with data that support internal benchmarking and stewardship decisions. These purposes impose different requirements on the chosen metric. The [WHO One Health framework](https://www.who.int/health-topics/one-health) situates antimicrobial use measurement within a broader strategy that links human, animal, and environmental health. The [CDC One Health resources](https://www.cdc.gov/one-health/index.html) similarly emphasize cross-sector collaboration for zoonotic disease and resistance control. Within this framework, food animal antimicrobial use data are also a production concern, they are part of an integrated surveillance system that informs public health policy. The choice of metric therefore has consequences beyond the farm gate. A fundamental distinction separates dose-based from duration-based metrics. Dose-based metrics, such as the animal daily dose (ADD), standardize the quantity of an antimicrobial used by expressing it as the number of defined daily doses per animal or per unit of animal population. Duration-based metrics, such as days of therapy (DOT), count the number of days an animal receives an antimicrobial. The two approaches answer different questions. ADD answers "how much antimicrobial was used," while DOT answers "for how many days were animals treated." In theory, the number of ADDs should approximate the number of DOTs for a given treatment course. Empirical work shows that this assumption fails in practice. ## Dose-Based Metrics and Their Limitations The ADD metric was adapted from the human defined daily dose concept. It assigns a standard daily dose to each antimicrobial based on a reference body weight and a typical dosing regimen, then expresses total use as the number of these standard doses consumed. This approach allows different antimicrobials to be compared on a common scale and permits aggregation across drug classes. The [quantification of antimicrobial use on Pennsylvania dairy farms](https://pubmed.ncbi.nlm.nih.gov/30594359/) demonstrated the practical application of ADD and DOT metrics in a large production population. Across 235 farms, the study reported treatment incidences of 4.2 ADD per 1,000 animal-days and 3.3 DOT per 1,000 animal-days. The two metrics ranked farms similarly, but the absolute values differed, and the study noted that the choice of metric affected the apparent magnitude of use. This finding matters for benchmarking because a farm that appears to be a high user by one metric may appear average by another. The [comparison of animal daily doses and days of therapy](https://pubmed.ncbi.nlm.nih.gov/32120054/) across canine patients, large animal hospital patients, and dairy herds found that the correlation between ADD and DOT was poor for hospital populations, with Lin correlation coefficients of 0.16 and 0.18 for small and large animals respectively. ADDs most often overestimated the number of DOTs. The discrepancy arises because the ADD assumes a fixed daily dose and treatment duration for each drug, whereas actual prescribing varies with body weight, severity of illness, and clinician judgment. A metric that assumes a standard treatment course cannot capture variation in that course. ## Duration-Based Metrics and Treatment Incidence Days of therapy counts the number of calendar days on which an animal receives an antimicrobial, regardless of dose. This metric is intuitive and directly reflects clinical decision-making. It is also less sensitive to differences in body weight and dosing regimens across populations, which makes it attractive for comparing farms with different animal demographics. The treatment incidence approach expresses either ADD or DOT per 1,000 animal-days, which standardizes for population size and observation period. This denominator is essential for comparing farms of different sizes or for tracking use over time within a single operation. The [Fijian livestock farm study](https://pubmed.ncbi.nlm.nih.gov/34568535/) used a different population denominator, mg per population correction unit (mg/PCU), which estimates total biomass treated. The study found that annual antimicrobial use in Fijian livestock was 44 mg/PCU compared with a global average of 118 mg/PCU, but that use was concentrated in 56% of farms, with the remaining 44% reporting no antimicrobial use. This distributional finding illustrates a critical point: aggregate metrics can obscure substantial variation between farms, and benchmarking programs must account for the fact that some farms may have zero use. ## Metric Selection and Reporting to Veterinary Audiences The choice of metric is not purely technical. It affects how veterinarians interpret and act on the data. A [survey of Canadian swine veterinarians](https://pubmed.ncbi.nlm.nih.gov/33967283/) found that respondents preferred dose-based metrics over weight-based and frequency-based alternatives, and that they self-evaluated their understanding of dose-based metrics as higher. The veterinarians reported multiple objectives for antimicrobial use information, including improving use on their clients' farms and enabling comparisons with other farms. These objectives are not identical, and a metric that serves one may not serve the other. The same survey noted that veterinarians interpreted most results of antimicrobial use analyzes correctly, but the small sample size and the variation in objectives suggest that reporting formats should be tailored to the intended decision. A veterinarian advising a single client needs different information than a practitioner comparing regional trends or a researcher evaluating a stewardship intervention. ## The Evidence Base for Stewardship Interventions The [systematic review of behavior-change interventions for antimicrobial stewardship](https://pubmed.ncbi.nlm.nih.gov/37155592/) identified only 11 publications in the animal health sector compared with 290 in human health. The authors could not perform a meta-analysis because of variation across intervention type, study type, and outcome. This evidence gap has direct implications for metric selection. Without a robust literature linking specific metrics to improved stewardship outcomes, programs must rely on face validity and on the internal consistency of their measurement approach. The review assessed interventions using metrics across five thematic areas: antimicrobial use, adherence to clinical guidelines, antimicrobial stewardship, antimicrobial resistance, and clinical outcomes. This framework is useful for program design because it distinguishes between measuring what was done (use, adherence) and measuring what happened as a result (resistance, clinical outcomes). A stewardship program that tracks only antimicrobial use cannot demonstrate that its interventions improved animal health or reduced resistance. ## Benchmarking Antimicrobial Use Across Farms and Production Systems Quantitative benchmarking requires a denominator that reflects the population at risk and the period of observation. The most widely applied population-level metric in food animal work is the treatment incidence, expressed as animal-defined daily doses (ADD) per 1,000 animal-days or days of therapy (DOT) per 1,000 animal-days. In a survey of Pennsylvania dairy farms, treatment incidences of 4.2 ADD and 3.3 DOT per 1,000 animal-days were reported, with mastitis as the dominant indication and first-generation cephalosporins the most frequently used class ([quantification of antibiotic use on dairy farms in Pennsylvania](https://pubmed.ncbi.nlm.nih.gov/30594359/)). These figures provide a reference range for dairy operations in the northeastern United States, but they should not be treated as universal thresholds. Herd-level factors such as lactation stage distribution, season, and culling policy shift the denominator and the numerator simultaneously. The mg per population correction unit (mg/PCU) metric remains common in national and regional surveillance because it standardizes across species by estimated biomass. In a survey of Fijian livestock farms, estimated annual use was 44 mg/PCU, below a global average of 118 mg/PCU, but use was concentrated in 56% of farms while the remainder reported no antimicrobial use ([quantification of antimicrobial use in Fijian livestock farms](https://pubmed.ncbi.nlm.nih.gov/34568535/)). That distribution illustrates a core benchmarking problem: averages conceal the difference between farms with no use and farms with intensive use. A stewardship program that tracks only the herd mean will miss the high-use tail, which is often where the greatest reduction opportunity lies. ### Selecting the Denominator The choice of denominator changes the interpretation of the benchmark. Animal-days at risk is the preferred denominator for longitudinal herd monitoring because it accounts for changes in herd size over the observation period. Biomass-based denominators such as PCU are preferable for cross-species or national comparisons because they standardize for body weight. A third option, treatment incidence per 1,000 animals present, is simpler but distorts when animals are present for short periods, as in veal or broiler production. For swine practice, Canadian veterinarians surveyed about metric preferences rated dose-based metrics as easier to interpret than weight- or frequency-based alternatives, and they used antimicrobial use reports for two main purposes: improving use on their own clients' farms and enabling comparisons with other farms ([choosing which metrics to use when reporting antimicrobial use information to veterinarians in the Canadian swine industry](https://pubmed.ncbi.nlm.nih.gov/33967283/)). The same survey found that veterinarians interpreted most results correctly regardless of metric type, which suggests that the limiting factor in benchmarking is not numeracy but the alignment between the metric and the management decision the report is meant to inform. ## Key Performance Indicators for Stewardship Programs A stewardship program needs a small set of indicators that can be calculated from records already kept on farm. Table 1 presents core KPIs with their formulas and indicative target ranges. Targets are expressed as ranges instead of fixed values because production system, disease pressure, and regional regulatory context all shift the feasible benchmark. | KPI | Calculation | Unit | Indicative Target Range | Primary Failure Mode | |---|---|---|---|---| | Treatment incidence (ADD-based) | Total ADD administered / 1,000 animal-days at risk | ADD per 1,000 animal-days | Dairy: 3 to 6, Swine grow-finish: 5 to 15 | Overestimation when ADD exceeds actual daily dose | | Treatment incidence (DOT-based) | Total days of therapy / 1,000 animal-days at risk | DOT per 1,000 animal-days | Dairy: 2 to 5, Swine grow-finish: 4 to 12 | Underestimation when treatment records omit duration | | Critically important antimicrobial use proportion | Antimicrobial courses using WHO CIA classes / total courses | Percentage | Less than 10% for routine prophylaxis, higher for therapeutic use | Masking when CIA use is concentrated in few animals | | Dry-off therapy intensity | Total antimicrobial courses at dry-off / number of dry-off events | Courses per dry-off event | 0.8 to 1.0 in conventional herds, 0 in selective protocols | Confounding by herd expansion | | Treatment failure rate | Courses repeated within 14 days / total courses | Percentage | Less than 5% for mastitis, less than 10% for respiratory disease | Misclassification of new infections as failures | | Benchmark deviation | Herd treatment incidence / regional median treatment incidence | Ratio | 0.5 to 1.5 | Regional median not stratified by production system | The ADD and DOT metrics do not always agree. In a comparison across canine patients, large animal hospital patients, and dairy herds, the correlation between ADD and DOT was poor for hospital populations, with Lin correlation coefficients of 0.16 and 0.18 for small and large animals respectively, and ADDs most often overestimated the number of DOTs ([comparison of animal daily doses and days of therapy for antimicrobials in species of veterinary importance](https://pubmed.ncbi.nlm.nih.gov/32120054/)). The practical implication is that a program should select one primary metric and apply it consistently. Switching between ADD and DOT across reporting periods will create artefactual trends that are not attributable to changes in prescribing behavior. ## The Assessment Sequence for a Herd-Level Stewardship Audit A stewardship audit follows a defined sequence. The first step is to verify the completeness of treatment records. Farms that record treatments on paper calendars or in a single treatment log often undercount duration, particularly for chronic conditions such as mastitis where a course may be extended beyond the label duration. The second step is to calculate the primary treatment incidence metric for the audit period, usually the preceding 12 months to smooth seasonal variation. The third step is to stratify use by indication, drug class, and production stage. Stratification by production stage is essential in dairy because dry-off therapy and lactation therapy respond to different interventions. The fourth step is to compare the herd's metrics against the relevant benchmark. For dairy herds in the northeastern United States, the Pennsylvania survey provides a reference of 4.2 ADD and 3.3 DOT per 1,000 animal-days ([quantification of antibiotic use on dairy farms in Pennsylvania](https://pubmed.ncbi.nlm.nih.gov/30594359/)). For operations outside that region, the benchmark should be drawn from regional surveillance programs or from a purpose-built cohort of comparable farms. The fifth step is to identify the highest-yield reduction target. In most dairy herds this is mastitis therapy, and in most swine systems it is group-level water medication for respiratory disease. The final step is to set a single, measurable reduction target for the next 12 months and to specify the intervention that will achieve it. ## Monitoring Parameters and Their Interpretation Each monitoring parameter detects a different failure mode. Treatment incidence detects changes in the volume of use. The proportion of critically important antimicrobials detects shifts toward or away from highest-priority classes. Treatment failure rate detects inadequate dosing, incorrect drug selection, or emerging resistance. Benchmark deviation detects whether the herd is moving relative to its peers. The interpretation of these parameters depends on production system. In a dairy herd, a rising treatment failure rate for clinical mastitis may indicate that the predominant pathogen has shifted from gram-positive to gram-negative organizms, which would warrant culture-based therapy selection instead of a change in drug class. In a swine grow-finish site, a rising failure rate for respiratory disease may indicate that the water medication system is delivering subtherapeutic doses due to medicator malfunction or line obstruction. The same KPI therefore triggers different diagnostic workups in different systems. Documentation should record the metric, the denominator, the observation period, and any adjustments for herd size changes. The record should also note the data source, such as treatment logs, purchase invoices, or practice management software, because the source affects the reliability of the numerator. Farms that rely on purchase invoices instead of treatment records will systematically overestimate use because purchased product is not always administered. ## Species and Production System Modifications The correct metric and benchmark differ by species and production system. In dairy, the dry-off event is a distinct therapeutic episode that must be separated from lactation therapy. In swine, the batch or flow group is the natural unit of analysis, and the relevant denominator is pig-days at risk for the grow-finish period. In broiler production, the short production cycle means that treatment incidence per 1,000 animal-days will be low even when a high proportion of flocks receive medication, because the denominator accumulates rapidly. For broilers, the proportion of flocks treated and the proportion of production cycles with any antimicrobial use are more informative than treatment incidence alone. Patient status also changes the correct choice. A hospitalized large animal patient receives individualised therapy that is best tracked with DOT, because dose escalation and renal adjustment make ADD calculations unreliable. A group-housed production animal receives standardized protocols that are best tracked with ADD, because the dose is fixed by label and body weight. The comparison of ADD and DOT in hospital populations showed poor correlation, which supports the use of duration-based metrics where individualised dosing is common ([comparison of animal daily doses and days of therapy for antimicrobials in species of veterinary importance](https://pubmed.ncbi.nlm.nih.gov/32120054/)). Available equipment changes the feasible monitoring approach. Farms with electronic herd management software can generate treatment reports automatically. Farms with paper records require manual abstraction, which is labor-intensive and prone to omission. In the latter case, a targeted audit of a single production stage, such as the dry-off period, may be more feasible than a whole-herd audit and can still yield actionable data. The choice of audit scope should be driven by the quality of the available records, not by the ideal design. ## Recognized Failure Modes in Antimicrobial Use Measurement The most frequently encountered failure mode in food animal antimicrobial use measurement is the silent divergence between dose-based and duration-based metrics. [A comparison of animal daily doses and days of therapy across species](https://pubmed.ncbi.nlm.nih.gov/32120054/) demonstrated that ADD and DOT correlate poorly in hospital populations, with Lin correlation coefficients of 0.16 and 0.18 for small and large animal patients respectively. The ADD metric most often overestimated days of therapy, sometimes by substantial margins. This divergence matters because a stewardship program that tracks only dose-based metrics may conclude that use is declining when treatment duration has actually lengthened, or vice versa. A second failure mode is denominator distortion. When the population at risk is estimated from inventory instead of from actual animal-days at risk, seasonal variation in stocking density or mortality can produce artefactual changes in treatment incidence. A farm that reduces its standing population while maintaining the same number of treatments will appear to have increased its use per 1,000 animal-days. Detection requires examining the numerator and denominator separately before interpreting the ratio. A third failure mode is the aggregation of dissimilar production classes. [Quantification of antimicrobial use in Fijian livestock farms](https://pubmed.ncbi.nlm.nih.gov/34568535/) found that use was concentrated in 56% of participant farms, with the remaining 44% using no antimicrobials at all. Pooling such farms into a single benchmark obscures the distinction between farms with genuine stewardship problems and those with no measurable use. Reporting should stratify by production class and by user status before any comparative judgment is made. A fourth failure mode is the reliance on self-reported data without verification. [Quantification of antibiotic use on dairy farms in Pennsylvania](https://pubmed.ncbi.nlm.nih.gov/30594359/) relied on farmer recall or treatment records, a method that is practical but vulnerable to under-reporting and recall bias. Where electronic treatment records exist, they should be cross-checked against purchase invoices or dispensing records at least annually. | Observation | Likely cause | Discriminating check | | --- | --- | --- | | ADD and DOT trends diverge over time | Treatment duration changing independently of dose | Recalculate both metrics for the same period and compare per-case | | Treatment incidence rises despite stable drug purchases | Denominator shrinkage | Verify population counts against movement records and mortality | | Benchmark position worsens after a good year | Change in case mix, not prescribing behavior | Stratify by disease category and production stage | | Farm reports zero use but invoices show purchases | Recording gap or off-label sourcing | Reconcile treatment records with purchase and dispensing logs | ## Common Errors in Metric Interpretation Less experienced analysts frequently mistake a single metric for a complete picture. The ADD metric corrects for potency differences between drugs but says nothing about how long an animal was treated. The DOT metric captures duration but is insensitive to dose intensity. [Choosing which metrics to use when reporting antimicrobial use information to Canadian swine veterinarians](https://pubmed.ncbi.nlm.nih.gov/33967283/) found that practitioners preferred dose-based metrics and interpreted most results correctly, but the study also noted that different objectives require different metrics. A report intended to support within-farm trend analysis may need a different metric than one intended for cross-farm benchmarking. A second common error is comparing farms across production systems without adjusting for the denominator. A dairy herd with year-round lactating animals has a different denominator structure than a seasonal calving herd or a batch-raised broiler flock. Direct comparison of treatment incidence across such systems is misleading. The corrective action is to report the denominator explicitly and to restrict benchmarking to farms within the same production system and region. A third error is treating a single year's data as a trend. Antimicrobial use varies with disease outbreaks, weather, and market conditions. At least three consecutive years of data are needed before a directional claim is defensible. The corrective action is to plot rolling averages and to annotate known disease events on the same timeline. ## Limitations of the Current Evidence The evidence base for antimicrobial stewardship metrics in food animals is thin. [A systematic review of behavior-change interventions across human and animal health](https://pubmed.ncbi.nlm.nih.gov/37155592/) identified only 11 publications in the animal health sector compared with 290 in human health. The authors could not perform a meta-analysis because of variation in intervention type, study type, and outcome measures. This means that most recommendations about which metrics to use, and how to interpret them, rest on descriptive studies and expert opinion instead of on controlled trials. Expert opinion still differs on the preferred metric for cross-sector comparison. Some authorities advocate for dose-based metrics because they are easier to standardize across species. Others argue that duration-based metrics are more intuitive for clinicians and more closely reflect the actual course of therapy. The [WHO One Health framework](https://www.who.int/health-topics/one-health) and the [WOAH terrestrial animal health standards](https://www.woah.org/en/what-we-do/standards/codes-and-manuals/terrestrial-code-online-access/) both call for harmonised measurement, but neither prescribes a single metric for all purposes. Until a consensus emerges, practices should select metrics based on their specific objectives and document the rationale. ## Escalation and Referral Criteria Referral to a specialist or laboratory is warranted when metric divergence cannot be explained by data quality issues. If ADD and DOT trends diverge persistently despite verified records, the cause may lie in off-label dosing practices, compounding, or unreported extra-label use. In such cases, laboratory analysis of feed or water samples may be needed to confirm actual drug delivery. Regulatory reporting obligations vary by jurisdiction and production system. Veterinarians should consult their [national professional body's practice resources](https://www.avma.org/resources-tools) and their [local veterinary reference materials](https://www.msdvetmanual.com/) for current requirements. When a farm's antimicrobial use data suggest a pattern of sustained high use of highest-priority critically important antimicrobials, or when use cannot be reconciled with purchase records, the case should be escalated to the relevant regulatory authority. The threshold for escalation is not a specific numerical value but the presence of unexplained discrepancy or suspected diversion. ## Frequently Asked Questions ### How Should a Practice Prioritize Metric Implementation When Resources Are Limited? Begin with a single, reliable metric that existing records can support. If treatment records are complete, days of therapy (DOT) offers intuitive interpretation and aligns with human stewardship reporting. If records are incomplete, animal daily doses (ADD) calculated from purchase invoices may be more feasible, but recognize that ADD tends to overestimate actual treatment duration, as shown in comparisons across hospital and dairy populations ([Redding et al., 2020](https://pubmed.ncbi.nlm.nih.gov/32120054/)). Establish a denominator that reflects production volume, such as animal-days or biomass. Add a second metric only after the first is consistently generated. A simple, accurate system outperforms a complex one that collapses under staffing constraints. ### What Denominator Should a Mixed-Operation Practice Use for Cross-Species Comparisons? No single denominator serves all species equally. For swine, Canadian veterinarians prefer dose-based metrics and value farm-to-farm comparisons, but their objectives for the data vary, which complicates standardization ([Bosman et al., 2021](https://pubmed.ncbi.nlm.nih.gov/33967283/)). For mixed operations, report each species separately using species-appropriate denominators, then aggregate only at the level of total antibiotic mass per population correction unit if required by a program or regulator. Avoid combining treatment incidence rates across species, because differences in body weight, typical durations, and disease patterns render the pooled number uninterpretable. ### How Do Metrics Differ Between Intensively Housed and Pasture-Based Systems? Intensively housed systems, particularly swine and poultry, generate group-level treatment records and accurate population denominators, making treatment incidence straightforward. Pasture-based beef and small ruminant systems often treat individuals sporadically, and accurate animal-days require estimation of stocking rates and turnout periods. Dairy operations present a hybrid, with lactating cow treatments well documented but young stock and dry cow therapy frequently under-recorded. A Pennsylvania dairy survey found that farmer self-report captured substantial use but depended on recall or treatment records of variable completeness ([Redding et al., 2019](https://pubmed.ncbi.nlm.nih.gov/30594359/)). Pasture systems may therefore require periodic census instead of continuous denominator tracking. ### What Are the Minimum Records Needed to Begin Measuring Antimicrobial Use? A functional system requires four data elements per treatment event: animal or group identifier, drug identity, dose administered, and date. For group treatments, record group size and the number of animals actually treated. Without these elements, neither dose-based nor duration-based metrics can be calculated reliably. Practices that lack electronic records can start with a paper log or spreadsheet, provided a single staff member owns data entry and verification. Purchase records alone cannot distinguish treatments from other uses and will overestimate clinical use. The WHO One Health framework emphasizes cross-sector data sharing, which presupposes that basic farm-level records exist in a usable form ([WHO One Health](https://www.who.int/health-topics/one-health)). ### How Should a Veterinarian Explain Metric Results to a Producer Who Is Not Familiar With Them? Translate the metric into a production-relevant statement. Instead of reporting 4.2 ADD per 1,000 animal-days, state that the average cow received treatment for roughly four days per year, or that one in ten animals was treated in the last month. Show the producer their own trend over time before introducing comparisons with other farms. Comparisons should be framed as ranges, not rankings, because farm-level variation in disease pressure and record completeness affects the numbers. The AVMA practice resources offer communication guidance for veterinary-client interactions that can be adapted to stewardship discussions ([AVMA practice resources](https://www.avma.org/resources-tools)). Emphasize that the goal is identifying reduction opportunities, not assigning blame. ### When Should a Practice Seek External Support for Its Stewardship Program? Seek external support when internal metrics show no improvement over two consecutive reporting periods despite active intervention, when a farm's use exceeds regional benchmarks for a critically important antimicrobial class, or when the practice lacks the statistical skills to interpret its own data. External support may come from diagnostic laboratories offering benchmarking services, university extension programs, or veterinary organizations with stewardship initiatives. The systematic review of stewardship interventions found that the animal health evidence base is thin, so practices should expect to adapt approaches from human healthcare instead of adopt a validated livestock template ([Craig et al., 2023](https://pubmed.ncbi.nlm.nih.gov/37155592/)). Referral is also appropriate when a producer resists data sharing, because benchmarking requires comparable, transparent records. ## Related Clinical & Scientific Guides * [Wildlife Disease Surveillance: Designing and Implementing a One Health Program](/knowledge/veterinary-medicine/veterinary-public-health/wildlife-disease-surveillance-designing-implementing-one-health-program) * [Biosecurity Risk Assessment for Livestock Operations: A Practical Framework](/knowledge/veterinary-medicine/veterinary-public-health/biosecurity-risk-assessment-livestock-operations-practical-framework) * [Rabies Post-Exposure Prophylaxis in Veterinary Personnel](/knowledge/veterinary-medicine/veterinary-public-health/rabies-post-exposure-prophylaxis-in-veterinary-personnel) ## References and Further Reading - [Quantification of antibiotic use on dairy farms in Pennsylvania.](https://pubmed.ncbi.nlm.nih.gov/30594359/). 2019. - [Behavior-change interventions to improve antimicrobial stewardship in human health, animal health, and livestock agriculture: A systematic review.](https://pubmed.ncbi.nlm.nih.gov/37155592/). 2023. - [Quantification of antimicrobial use in Fijian livestock farms.](https://pubmed.ncbi.nlm.nih.gov/34568535/). 2021. - [Antibiotic usage in 14 equine practices over a 10-year period (2012-2021).](https://pubmed.ncbi.nlm.nih.gov/37587746/). 2024. - [Comparison of animal daily doses and days of therapy for antimicrobials in species of veterinary importance.](https://pubmed.ncbi.nlm.nih.gov/32120054/). 2020. - [Choosing which metrics to use when reporting antimicrobial use information to veterinarians in the Canadian swine industry.](https://pubmed.ncbi.nlm.nih.gov/33967283/). 2021. - [WHO One Health Initiative](https://www.who.int/health-topics/one-health). WHO. - [CDC One Health and Zoonotic Disease Resources](https://www.cdc.gov/one-health/index.html). CDC. - [MSD Veterinary Manual, Professional Edition](https://www.msdvetmanual.com/). MSD Veterinary Manual. ## Related Articles - [Antimicrobial Stewardship in Veterinary Practice: Guidelines and Implementation Barriers](/knowledge/veterinary-medicine/veterinary-public-health/antimicrobial-stewardship-veterinary-practice-guidelines-implementation-barriers) - [Food Safety Risk Assessment in Veterinary Practice](/knowledge/veterinary-medicine/veterinary-public-health/food-safety-risk-assessment-in-veterinary-practice) - [Antimicrobial Resistance Surveillance in Food Animals: Sampling Strategies and Data Interpretation](/knowledge/veterinary-medicine/veterinary-public-health/antimicrobial-resistance-surveillance-food-animals-sampling-strategies-data-interpretation) - [Food Safety Inspection in Veterinary Practice: Critical Control Points and Regulatory Compliance](/knowledge/veterinary-medicine/veterinary-public-health/food-safety-inspection-veterinary-practice-critical-control-points-regulatory-compliance) - [Zoonotic Disease Risk Assessment in Veterinary Practice](/knowledge/veterinary-medicine/veterinary-public-health/zoonotic-disease-risk-assessment-in-veterinary-practice) > This article is educational professional reference material for veterinary audiences. It is not a substitute for veterinary diagnosis, individual clinical judgment, current product labeling, or applicable regulatory requirements.