Annually in the United States human population, approximately 2 million antimicrobial-resistant infections occur and result in roughly 23,000 deaths.1,2 Mitigating AMR is one of the most important challenges in public health and is a complex issue involving various dynamic social, political, and economic factors.3 A major driving force behind the development of AMR is the use of AMDs in human and veterinary medicine.2,4
In production animal medicine, AMDs are routinely used for disease treatment, control, and prevention.5 Recently, substantial efforts have been made in the United States to reduce use of medically important AMDs (ie, AMDs6 important for treating human diseases) in production animal medicine.7 Interventions include regulations that curtailed the use of medically important AMDs for growth promotion purposes in livestock and that increased veterinary oversight of the use of medically important AMDs for disease prevention in animals.8,9 Implemented in 2017, revisions to the veterinary feed directive regulations10 require veterinary oversight for the use of medically important AMDs in animal feed or water. Yet, there is a lack of data on US-specific veterinary prescription practices, with which potential impacts of such regulations and other efforts over time may be assessed, and data are limited regarding veterinarians' perceptions and attitudes about AMR and factors influencing the use of AMDs. Historically, sales data had been used to quantify AMD consumption and use; however, sales data do not accurately reflect AMD use,11,12 and accurate data reflecting AMD use (usage data) in production animal medicine are needed so that the impact of regulations and other interventions intended to reduce AMD use can be assessed. Such usage data, along with more general information on veterinarians' perceptions and attitudes about AMR, could guide development of effective tools and trainings intended to mitigate AMR and curtail nonessential use of AMDs. Therefore, the purpose of the study reported here was to assess production animal medicine veterinarians' prescription practices and identify factors influencing their use of AMDs and their perceptions of and attitudes toward AMR.
Materials and Methods
Data collection was conducted in 2 parts. First, a cross-sectional, online survey of a sample of production animal medicine veterinarians was used to gather data pertaining to prescription practices for AMDs, perceptions and attitudes regarding AMR, and influencing factors for the use of AMDs. Second, a subsample of survey respondents maintained prospective prescription diaries, which we later assessed and used to validate the AMD prescription data results obtained from the survey. In both processes, respondents were grouped according to their selected production setting disease scenarios and reported number of years as veterinarians. The survey and study protocol were reviewed by the Colorado Multiple Institutional Review Board, which determined them exempt from a full review under the Institutional Review Board protocol.
Cross-sectional survey
A questionnaire (Supplementary Appendix SI, available at avmajournals.avma.org/doi/suppl/10.2460/javma.257.1.87) was developed, then administered through an online survey platform.a An email invitation for practicing production animal veterinarians to complete the survey was disseminated primarily through 2 email lists maintained by: the American Association of Bovine Practitioners (approx 2,200 members) and Academy of Veterinary Consultants (approx 800 members), both of which included veterinary students and other animal health professionals as well as production animal medicine veterinarians. In addition, a poultry medicine expert recruited other veterinarians in poultry medicine through professional network email channels. Survey responses were voluntary and anonymous, and the survey was open from March 1 to May 31, 2017. One reminder email was sent 6 weeks after the initial invitation. Responses from individuals determined not to have been practicing in production animal medicine were excluded from analyses. Questions in the survey were divided into 3 sections: demographic information (eg, years as a veterinarian, geographic region within the United States, and professional sector), disease scenarios, and perceptions and attitudes regarding AMD use and AMR.
Production setting disease scenarios—Five disease scenarios, 1 for each of 5 specific animal production settings, including feedlot cattle, backgrounding cattle, dairy cattle, swine, and poultry, were created and used to assess respondents' treatment recommendations for a respiratory disease (ie, pneumonia), which is reported13–15 as the most prevalent disease condition. Species experts were consulted during the development of each hypothetical scenario, which included physical examination findings for an individual ill animal in a group of animals (exposed group), characteristics of the exposed group, and characteristics of a disease-free group at high risk of contracting the scenario disease.
The feedlot cattle scenario involved a 318-kg (700-lb) calf that had been moved to the feedlot's hospital pen and had a history of isolation from the group, inappetence, lethargy, nasal discharge, and fever (40.6°C [105°F]). This calf and 200 other calves had been purchased at auction 30 days earlier, administered vaccines and a broad-spectrum anthelmintic on arrival at the feedlot, and kept as a group in the same pen. Approximately 40 of the 200 (20%) calves in the group showed signs of respiratory disease. Four calves from the group had died, and results of necropsy indicated that 2 had fibrinous pneumonia and 2 had effusive pleuritis. The feedlot also had a high-risk disease-free group of calves that had been purchased from an auction, were transported several hundred miles to the feedlot, and were intended to undergo initial processing and placement.
The backgrounding cattle scenario involved a calf that was recently weaned and shipped but had not been vaccinated. The calf isolated itself from other calves in its group, was lethargic and febrile (40.6°C), and had a soft and moist cough, increased lung sounds in the cranioventral lung fields, and a decreased rumen fill. This sick calf was in a pen of 50 calves that entered the operation from 3 direct sources around the same time. Signs of respiratory disease were evident in approximately 10 of the 50 (20%) calves. In addition, the operation had a high-risk disease-free group of bull and steer calves that had been purchased from various sources and transported several hundred miles.
The dairy cattle scenario involved a Holstein heifer calf that had been vaccinated and recently weaned. Shortly after weaning, the heifer was transported from a calf ranch to the dairy. After signs of lethargy, isolation, and coughing were noticed, the heifer calf was moved to a sick pen. The heifer calf was dehydrated, febrile (40.6°C), and tachypneic; had increased lung sounds in the cranioventral lung fields; and showed signs of depression. The pen from which the heifer came had 50 heifer calves that were recently weaned and comingled. Of these calves, 5 had similar signs of respiratory disease and 2 others had died. Results of necropsy indicated that one had fibrinous pneumonia and the other had effusive pleuritis. In addition, the facility had a group of weaning-age Holstein heifers that were to be moved from hutches to a grow barn, and the farm consistently had issues with colostrum administration at birth, inadequate housing, and poor stress management. The farm attributed 15% of its annual calf deaths to respiratory disease.
The swine scenario involved a 32-kg (70-lb) grower that been vaccinated a few weeks before leaving the nursery and was then housed in a 1,100-head finisher barn but moved to a hospital pen after a nonproductive cough and thin appearance were noticed. In this finisher barn, 110 of the 1,100 (10%) pigs showed signs of a similar nonproductive, persistent cough and had decreased feed consumption. Mortality rate reached 5% in the last week, and necropsy findings included exudative bronchopneumonia, pleuritis, and occasionally pericarditis. In addition, the facility brought 1,000 pigs from a sow farm with serologic and clinical evidence of porcine reproductive and respiratory syndrome virus to a wean-to-finish barn, and 50 (5%) were small but healthy. At slaughter, a high percentage of pigs from the facility have had pleural adhesions.
The poultry scenario involved a grow house of ten thousand 9-week-old turkeys, of which approximately 1,000 (10%) had signs of respiratory disease, including conjunctivitis, ocular discharge, and swollen paranasal sinuses. Additionally, feed consumption and weight gain were low. Results of necropsies indicated sinusitis and air sacculitis. The farm also had 4 other houses, each containing a group of 10,000 young turkeys. One of the groups was being moved from brooder houses to a single grow house and was considered a high-risk disease-free group. At harvest, turkeys from this farm have had a high condemnation rate because of air sacculitis.
AMD recommendations—Survey participants were asked to answer questions from 1 scenario on the basis of the species that they treated most often. Response options for treatments were multiple choice (eg, AMDs, NSAIDs, fluid therapy, increased nursing and monitoring, nutritional supplements, and vaccination), and multiple treatments could be selected, including an option of other with an open text field in which respondents could describe treatments not listed. If respondents chose AMD treatment for an individual ill animal (yes or no), exposed group (all, some, or none), or disease-free high-risk group (yes or no), alone or in combination, the survey instrument prompted additional questions about prescription information and whether the recommended AMD would be an empirical choice or confirmed with assessment of diagnostic samples.
Perceptions, attitudes, and influencing factors—A 5-point Likert-type scale (1 = strongly disagree, 2 = disagree, 3 = neither agree nor disagree, 4 = agree, and 5 = strongly agree) was used to measure respondents' reported level of agreement with statements pertaining to factors that might influence veterinarians' decisions to prescribe AMDs. The statements were designed to help us assess whether the media's portrayal of veterinary use of AMDs, clients' pressure and expectations for AMD prescriptions, and government policies regulating the use of AMDs in veterinary medicine influenced veterinarians' decisions to prescribe AMDs. In addition, we assessed respondents' reported perceptions and attitudes on AMR with multiple-choice questions that focused on contributing and mitigating factors for AMR. For each of these questions, more than 1 choice could be selected, including a choice of other with an open text field in which respondents provide answers not listed.
Prospective prescription diary
Respondents were invited to prospectively log details of consultations and treatment recommendations over a 4-week period by completing an electronic prescription diary. Data collected through diary entries included species and number of animals affected, clinical signs, and recommended diagnostic procedures and treatments, including whether any AMDs were prescribed. Data that could potentially identify producers or veterinarians were not collected, and all entries were submitted anonymously through an online platform.b Survey respondents who agreed to keep a prescription diary were provided an electronic tablet devicec to be used to record their diary information, then to keep as compensation for their participation.
Statistical analysis
The primary outcome of interest was whether AMD treatment was selected by respondents. The κ2 test was used to compare results for responses reported by respondents grouped according to their selected production setting disease scenarios and reported number of years as veterinarians. Perception and attitude rankings of all respondents were explored with a descriptive analysis and grouped as agreed (ie, strongly agree and agree), disagreed (ie, strongly disagree and disagree), or neutral (ie, neither agree nor disagree). Percentages reported reflected the frequency of agreement out of all responses. Statistical softwared was used for all analyses, and values of P < 0.05 were considered significant.
Results
Survey response
Because the survey invitation was primarily disseminated through 2 email lists that included veterinary students and other animal health professionals in addition to practicing veterinarians, the survey invitation likely reached < 3,000 veterinarians practicing in production animal medicine. The response rate was difficult to calculate because of the mixed membership of the listservs used. Nonetheless, there were 184 respondents who on the basis of their responses were considered to have been veterinarians in the United States. From these 184 individuals, 157 complete and 27 incomplete surveys were received. Incomplete surveys were excluded from analyses.
Demographic information
Of the 157 respondents who submitted completed surveys, 85 (54.1%) reported that they had been veterinarians for ≤ 15 years and 72 (45.9%) reported that they had been veterinarians for ≥ 16 years. The most commonly reported region of practice in the United States was the Midwest region (81/157 [51.6%]), which included the survey response options of East North Central (Ohio, Indiana, Illinois, Michigan, and Wisconsin; 34/157 [21.7%]) and West North Central (Minnesota, Iowa, Missouri, North Dakota, South Dakota, Nebraska, and Kansas; 47/157 [29.9%]) US region options in the survey. Most (152/157 [96.8%]) respondents reported working in private practice, whereas the remaining 5 (3.2%) respondents reported working in industry (2 [1.3%]), academia (2 [1.3%]), or government (1/157 [0.6%]).
Production setting disease scenarios
The most commonly selected production setting disease scenarios were dairy cattle (96/157 [61.1%]), backgrounding cattle (32/157 [20.4%]), and feedlot cattle (20/157 [12.7%]; Table 1). Because few respondents selected the swine (5/157 [3.2%]) or poultry (4/157 [2.5%]) scenarios, those responses were excluded from statistical analysis of AMD treatment.
Demographic characteristics reported and production setting disease scenarios selected by 157 US production animal medicine veterinarians in an online survey administered from March 1 to May 31, 2017, to gather data pertaining to prescription practices for AMDs, perceptions and attitudes regarding AMR, and influencing factors for the use of AMDs.
| Characteristics | No. of respondents | Percentage (95% CI) |
|---|---|---|
| Production setting disease scenario | ||
| Dairy cattle | 96 | 61.1 (54.1–69.1) |
| Backgrounding cattle | 32 | 20.4 (13.3–28.3) |
| Feedlot cattle | 20 | 12.7 (5.7–20.7) |
| Swine | 5 | 3.2 (0.0–11.1) |
| Poultry | 4 | 2.5 (0.0–10.5) |
| US geographic region | ||
| Midwest* | 81 | 51.6 (43.9–59.6) |
| West† | 27 | 17.2 (9.6–25.2) |
| South‡ | 26 | 16.6 (8.9–24.6) |
| Northeast§ | 23 | 14.6 (7.0–22.7) |
| Years as a veterinarian | ||
| ≤ 15 y | 85 | 54.1 (46.5–62.4) |
| ≥ 16 y | 72 | 45.9 (38.2–54.1) |
| Professional sector | ||
| Private practice | 152 | 96.8 (94.9–99.6) |
| Other (eg, industry, government, or academia) | 5 | 3.2 (1.3–5.9) |
The Midwest region consisted of Minnesota, Iowa, Missouri, North Dakota, South Dakota, Nebraska, Kansas, Ohio, Indiana, Illinois, Michigan, and Wisconsin.
The West region consisted of Washington, Oregon, California, Alaska, Hawaii, Montana, Idaho, Wyoming, Colorado, New Mexico, Arizona, Utah, and Nevada.
The South region consisted of Arkansas, Louisiana, Oklahoma, Texas, Kentucky, Tennessee, Alabama, Mississippi, Florida, Georgia, South Carolina, North Carolina, and Virginia
The Northeast region consisted of Maine, New Hampshire, Vermont, Massachusetts, Rhode Island, Connecticut, New York, New Jersey, Pennsylvania, Delaware, Maryland, and West Virginia.
Survey scenario AMD treatment
Individual ill animal—Of the remaining 148 respondents, 147 (99.3%) indicated that they would include AMDs in the treatment recommended for the individual ill animal presented in their selected scenario. This finding was consistent for respondents across their selected production settings (feedlot cattle, 20/20 [100%]; backgrounding cattle, 32/32 [100%]; and dairy cattle, 95/96 [99%]).
Exposed group—Overall, 134 of the 148 (90.5%) respondents who selected any 1 of the 3 cattle production setting scenarios reported they would recommend AMD treatment for all or some animals in the exposed group (Figure 1; Table 2). However, the proportion of such differed significantly (P = 0.01) for respondents grouped by their selected production setting scenario (backgrounding cattle [29/32 {91%}] vs dairy cattle [89/96 {93%}] vs feedlot cattle [16/20 {80%}]). In addition, proportions of respondents who recommended AMD treatment for all versus some versus none of the animals in the exposed group differed significantly (P < 0.01) across production settings in that 13 of the 20 (65%) respondents who selected the feedlot cattle scenario reported that they would treat the entire exposed group, whereas 19 of 32 (59%) respondents for the backgrounding cattle scenario and 23 of the 96 (24%) respondents for the dairy cattle scenario reported that they would treat the entire exposed group. The decision to treat only some animals in the exposed group was reported most commonly by respondents who selected the dairy cattle scenario (66/96 [69%]), followed by the backgrounding cattle scenario (10/32 [31%]) and the feedlot cattle scenario (3/20 [15%]).
Summary of the percentages of 148 production animal veterinarians in the United States who reported in an online survey administered from March 1 to May 31, 2017, that they would recommend AMD treatment of all, some, or none of the animals in exposed groups of calves described in the respondents' selected production setting scenarios (dairy cattle, backgrounding cattle, or feedlot cattle).
Citation: Journal of the American Veterinary Medical Association 257, 1; 10.2460/javma.257.1.87
Results of analysis to identify differences in survey responses regarding recommended AMD treatment of animals in an exposed group (treat all, some, or none) and in a high-risk disease-free group (yes or no) as reported by 148 veterinarians described in Table I, stratified by their selected production setting disease scenario (feedlot cattle [n = 20], backgrounding cattle [32], or dairy cattle [96]) and reported number of years as a veterinarian (≤ 15 years [82; 9 for the feedlot scenario, 17 for the backgrounding scenario, and 56 for the dairy scenario] or ≥ 16 years [66; 11 for the feedlot scenario, 15 for the backgrounding scenario, and 40 for the dairy scenario]).
| Years as a veterinarian | |||||||
|---|---|---|---|---|---|---|---|
| ≤ 15 years | ≥ 16 years | ||||||
| Production setting disease scenarios | Animal groups | AMD treatment | No. of respondents | Percentage (95% CI) | No. of respondents | Percentage (95% CI) | P value |
| Feedlot cattle | 9 | 11 | 0.11 | ||||
| Exposed group | All | 6 | 67 (44–98) | 7 | 64 (45–96) | ||
| Some | 2 | 22 (0–53) | 1 | 9 (0–42) | |||
| None | 1 | 11 (0–42) | 3 | 27 (9–57) | |||
| High-risk disease-free group | Yes | 7 | 78 (67–100) | 8 | 73 (54–100) | 0.07 | |
| Backgrounding cattle | 17 | 15 | < 0.01 | ||||
| Exposed group | All | 8 | 47 (29–75) | 11 | 73 (60–98) | ||
| Some | 8 | 47 (29–75) | 2 | 13 (0–38) | |||
| None | 1 | 6 (0–33) | 2 | 13 (0–38) | |||
| High-risk disease-free group | Yes | 7 | 41 (24–67) | 6 | 40 (20–66) | 0.91 | |
| Dairy cattle | 56 | 40 | < 0.01 | ||||
| Exposed group | All | 10 | 18 (7–29) | 13 | 33 (20–48) | ||
| Some | 40 | 71 (61–83) | 26 | 65 (53–81) | |||
| None | 6 | 11 (0–23) | 1 | 3 (0–18) | |||
| High-risk disease-free group | Yes | 19 | 34 (23–47) | 13 | 33 (20–48) | 0.29 | |
High-risk disease-free group—Overall, 60 of the 148 (40.5%) respondents who selected any 1 of the 3 cattle production scenarios reported that they would recommend AMD treatment of the high-risk disease-free group (Table 2). However, the proportion of such respondents differed significantly (P < 0.01) for respondents grouped by their selected production setting scenario (feedlot cattle [15/20 {75%}] vs backgrounding cattle [13/32 {41%}] vs dairy cattle [32/96 {33%}]).
Years as a veterinarian—When the 148 respondents who selected any 1 of the 3 cattle production setting scenarios were further grouped by their reported years as veterinarians, 82 respondents (9 for the feedlot scenario, 17 for the backgrounding scenario, and 56 for the dairy scenario) reported being veterinarians for ≤ 15 years, whereas 66 respondents (11 for the feedlot scenario, 15 for the backgrounding scenario, and 40 for the dairy scenario) reported being veterinarians for ≥ 16 years. Across production setting scenarios, the proportion of respondents who selected AMD treatment for only some animals in the exposed group was significantly (P = 0.05) higher for respondents who reported being veterinarians for ≤ 15 years (feedlot scenario, 2/9; backgrounding scenario, 8/17; and dairy scenario, 40/56), compared with those who reported being veterinarians for ≥ 16 years (feedlot scenario, 1/11; backgrounding scenario, 2/15; and dairy scenario, 26/40; Table 2). In addition, the proportions of respondents who recommended AMD treatment for the entire exposed group in the backgrounding or dairy scenarios were significantly higher for respondents who reported being veterinarians for ≥ 16 years (11/15 and 13/40, respectively), compared with respondents who reported being veterinarians for ≤ 15 years (8/17 and 10/56, respectively). However, respondents' recommendations regarding AMD treatment for the high-risk disease-free group did not differ substantially on the basis of reported years as a veterinarian.
AMD products recommended—The most commonly recommended injectable AMD treatment reported by respondents for the dairy and backgrounding cattle scenarios was tulathromycin for the individual ill animal, exposed group, and high-risk disease-free group (Table 3). In the feedlot scenario, respondents most commonly recommended florfenicol for the individual ill animal but tulathromycin for the exposed group and high-risk disease-free group. All AMD treatments recommended were according to FDA-approved product labels, and no extralabel AMD use was recommended.
Summary of the top 3 AMDs in injectable formulations most commonly recommended by respondents described in Table 2 for treatment of the hypothetical animals (individual ill animal, exposed group, or high-risk disease-free group, alone or in combination) described in the respondents' selected production setting scenarios.
| Feedlot cattle | Backgrounding cattle | Dairy cattle | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Animals treated | AMD | No. of respondents (n = 20) | Percentage (95% CI) | AMD | No. of respondents (n = 32) | Percentage (95% CI) | AMD | No. of respondents (n = 96) | Percentage (95% CI) |
| Individual ill animal | 20 | 32 | 95 | ||||||
| Florfenicol | 8 | 40 (20–63) | Tulathromycin | 14 | 44 (28–63) | Tulathromycin | 48 | 51 (41–61) | |
| Tulathromycin | 6 | 30 (10–53) | Florfenicol | 10 | 31 (16–50) | Florfenicol | 21 | 22 (13—33) | |
| Enrofloxacin | 4 | 20 (0–43) | Tilmicosin | 5 | 16 (0–35) | Ceftiofur crystalline free acid | 10 | 10 (1–21) | |
| Exposed group (all or some) | 16 | 29 | 89 | ||||||
| Tulathromycin | 7 | 44 (25–72) | Tulathromycin | 10 | 34 (17–53) | Tulathromycin | 46 | 52 (42—62) | |
| Tilmicosin | 5 | 31 (13–60) | Tilmicosin | 5 | 17 (0–36) | Florfenicol | 14 | 16 (6–26) | |
| Florfenicol | 3 | 19 (0–47) | Florfenicol | 5 | 17 (0–36) | Ceftiofur crystalline free acid | 9 | 10 (0–21) | |
| High-risk disease-free group | 15 | 13 | 32 | ||||||
| Tulathromycin | 10 | 67 (47–88) | Tulathromycin | 7 | 55 (31–80) | Tulathromycin | 10 | 31 (16–50) | |
| Tilmicosin | 2 | 13 (0–34) | Tilmicosin | 2 | 15 (0–41) | Florfenicol | 5 | 16 (0–35) | |
| Ceftiofur crystalline free acid | 1 | 7 (0–28) | Enrofloxacin | 2 | 15 (0–41) | Ceftiofur crystalline free acid | 5 | 16 (0–35) | |
Diagnostic samples—Only 38 of the 148 (25.7%) respondents reported that their AMD treatment choice would be confirmed with results of diagnostic sampling (eg, bacterial culture and susceptibility testing, necropsy, blood work, or radiography, alone or in combination). There were no meaningful differences in the proportion of such respondents when grouped by production setting scenario selected or by reported years as a veterinarian.
Perceptions, attitudes, and influencing factors—Overall, most (126/157 [80.2%]) respondents reported that they agreed that government regulations influenced their AMD prescribing, whereas fewer reported that they agreed that client expectations (63/157 [40.1%]) or media perceptions (47/157 [29.9%]) influenced the way they prescribed AMDs. When asked which factors contribute to AMR, respondents most commonly selected AMD use in human medicine (148/157 [94.2%]), followed by AMD use in veterinary medicine (127/157 [80.9%]), environmental factors (116/157 [73.9%]), and evolutionary factors (111/157 [70.7%]). When asked about factors in veterinary medicine that contribute to AMR, most respondents indicated owner and producer compliance (143/157 [91.1%]), followed by client expectations (118/157 [75.2%]), veterinary prescription practices (108/157 [68.8%]), AMDs used for growth promotion (99/157 [63.1%]), and AMDs used for treatment in veterinary medicine (80/157 [51.0%]). When asked about potential interventions to reduce AMR, respondents most commonly selected the survey option to focus on environmental modifications to prevent disease (143/157 [91.1%]), followed by research and development of other disease interventions (130/157 [82.8%]), education of owners and producers on the importance of compliance (126/157 [80.3%]), education of veterinary students on judicious antimicrobial use principles (105/157 [66.9%]), and increased government regulation of AMD use (35/157 [22.3%]). No meaningful differences in results for perceptions, attitudes, and influencing factors were identified when respondents were grouped by their selected production setting scenario or reported years as a veterinarian.
Prospective prescription diary
Of the 157 respondents, 22 (14.0%) volunteered to keep a prospective prescription diary (diary respondents) and recorded 400 diary entries between June 2017 and August 2017. Mean ± SD entries per respondent was 18 ± 7.4 (range, 1 to 79). Of the 22 diary respondents, 13 (59%) identified as primarily dairy cattle veterinarians, 5 (23%) identified as primarily backgrounding cattle veterinarians, and 4 (18%) identified as primarily feedlot veterinarians.
Individual ill animal—All 22 diary respondents recorded ≥ 1 entry pertaining to an individual ill animal. There were 245 entries for individual ill animals, and the most common illnesses represented were respiratory disease (n = 83 [33.9%]), mastitis (38 [15.5%]), and diarrhea (28 [11.4%]). Of these 245 individual ill animals, 211 (86.1%) were reported by diary respondents to have received AMD treatment, which was a lower proportion than reported through the survey (147/148 [99.3%]).
Exposed group—Of the 22 diary respondents, 21 described ≥ 1 case involving a group of animals (≥ 2 animals). Overall, respondents reported 155 cases with exposed groups and AMD treatment of exposed groups in 142 of these 155 (91.6%) cases, with either all (97/155 [62.6%]) or some (45/155 [29.0%]) of the animals treated. The overall proportion of reported AMD treatment of exposed groups was higher in the diary component (142/155 [91.6%]) than in the survey component (98/148 [66.2%]). The proportion of cases in which diary respondents reported that only some animals in the exposed group were treated differed significantly (P < 0.01) for diary respondents grouped on the basis of whether they identified as dairy veterinarians (32/53 [60]), feedlot veterinarians (4/16 [25%]), or backgrounding cattle veterinarians (9/23 [39%]).
Years as a veterinarian—When diary respondents were grouped by reported years as a veterinarian, 11 (50%) reported ≤ 15 years, and 11 (50%) reported ≥ 16 years. Mean number of animals per diary entry was significantly (P < 0.01) lower for respondents who reported being a veterinarian for ≤ 15 years (12.5 animals/diary entry), compared with those who reported being a veterinarian for ≥ 16 years (66.0 animals/diary entry). In addition, the proportion of cases in which diary respondents indicated that only some animals in exposed groups received AMD treatment was significantly (P = 0.05) higher for those who reported being a veterinarian for ≤ 15 years (21/51 [41%]) versus ≥ 16 years (24/104 [23.1%]). In contrast, the proportion of cases in which diary respondents indicated that the entire exposed group received AMD treatment was significantly (P = 0.05) higher for those who reported being a veterinarian for ≥ 16 years (71/104 [68.3%]) versus ≤ 15 years (26/51 [51%]).
The use of diagnostic samples (eg, bacterial culture and susceptibility testing [n = 3], blood work [23], diagnostic imaging [11], or necropsy [13]) to guide AMD use was reported in 58 of the 400 (14.5%) diary entries. The reported use of such additional diagnostic procedures did not substantially differ by diary respondents grouped by reported years as a veterinarian; however, the proportion of such diagnostic sampling was reported less frequently in the diary component (58/400 [14.5%]) than in the survey component (38/148 [25.7%]).
Discussion
In the present study, we used the combination of an online survey component and a prospective prescription diary component to collect data about US production animal medicine veterinarians' perceptions and attitudes on AMR, prescription practices for AMDs, and influencing factors for AMD use. Our results could contribute to the filling of an important data gap in AMD use in production animal medicine and provide baseline information for assessing changes in practices over time. Results indicated that AMD use and usage patterns (eg, to treat all vs some vs none in an exposed group) differed substantially for respondents grouped by their selected production setting scenario and reported years as veterinarians. These differences emphasized that production setting and years in clinical practice are important to consider when assessing AMD prescription practices in production animal medicine.
Antimicrobial drug usage patterns differed substantially across respondents grouped by their selected production setting scenario. For instance, respondents who selected the dairy cattle scenario (vs the backgrounding or feedlot cattle scenarios) more commonly reported that they would prescribe AMDs to all or some animals in the exposed group. Those who selected the feedlot cattle scenario more commonly reported than other respondents that they would not treat the exposed group; however, of respondents who reported that they would treat the exposed group with AMDs, those who selected the feedlot scenario more commonly than other respondents recommended AMD treatment for the entire exposed group. This finding suggested that feedlot cattle are treated as groups rather than as individual animals and that a larger proportion of feedlot cattle in a herd (vs dairy or backgrounding cattle) are treated when AMDs are recommended. In contrast, respondents who selected the dairy cattle scenario more commonly reported than other respondents that they would recommend AMD treatment of some animals in the exposed group, rather than all or none of the animals in the exposed group. Additionally, respondents who selected the feedlot or backgrounding cattle scenario (vs the dairy cattle scenario) more commonly reported that they would recommend AMD treatment for the high-risk disease-free group.
Our findings for differences in recommendations of AMD use reported by respondents grouped by production setting scenarios supported USDA data that shows approximately 296 of 403 (73.4%)16 feedlots used feed-grade antimicrobials, compared with 106 of 582 (18.2%)17 dairy cattle operations. However, the referenced USDA data only describe feed-grade antimicrobials and do not necessarily imply that cattle were not administered AMDs in other forms (eg, injectable). Additionally, because the data were collected prior to 2012, it is not clear whether all AMD use included in these estimates16,17 is covered by the revised veterinary feed directive regulations.10 A recent study18 from Austria found that there was less AMD use in dairy production, compared with other production settings; however, more medically important AMDs (ie, third- and fourth-generation cephalosporins) were used in dairy production, possibly because these AMDs have shorter withdrawal intervals. Results of the present study were consistent with this finding in that third-generation cephalosporins were more commonly recommended by respondents who selected the dairy cattle scenario versus the feedlot or backgrounding cattle scenarios. The differences detected in results for AMD prescription practices reported by respondents grouped by their selected production setting scenarios were likely because of differences in production and herd management and AMD treatment approaches (eg, number of animals treated and AMD class prescribed).
In the present study, respondents across production settings who reported being veterinarians for ≥ 16 years (vs ≤ 15 years) more commonly recommended AMD treatment of all animals in the exposed group, rather than some or none of the animals in the exposed group. This finding suggested that veterinarians with more experience used more AMDs, and we recognized several possible explanations. It was possible that the more recently trained veterinarians were more aware of AMR issues and decided to prescribe fewer AMDs. This argument is supported by a study19 that shows an inverse relationship between years of veterinary experience and knowledge of AMR for bovine veterinarians in Ohio. Similarly, a Dutch study20 shows that less experienced veterinarians were more convinced that AMD use in veterinary medicine contributed to AMR and were more reluctant to prescribe AMDs, compared with their more experienced colleagues. In addition, more experienced veterinarians could have relied more on their personal experience of effective treatment approaches, which may have included prescribing AMDs to all exposed animals. For example, an Irish study21 shows that veterinarians relied heavily on previous experience, resulting in possible inappropriate AMD prescribing practices, and the authors concluded that changing prescribing behaviors in those who rely on previous experience might therefore present additional challenges. Further, respondents who reported being veterinarians for ≥ 16 years in the present study could have overseen larger groups of animals than did those who reported being veterinarians for ≤ 15 years, and larger groups may be less amenable to the intensive nature of partial group treatments. This point was supported by our finding that the mean number of animals per diary entry was substantially smaller for diary respondents who reported being veterinarians for ≤ 15 years, compared with ≥ 16 years.
In the present study, the proportion of survey respondents who reported that diagnostic samples would be used in confirming their AMD choice was low (38/148 [25.7%]) and did not differ substantially across respondents grouped by their selected production setting scenario or reported years as a veterinarian. This finding indicated that respondents would not routinely rely on results of diagnostic procedures to make decisions about prescribing AMDs. A study22 that specifically looked at bacterial culture and susceptibility testing shows that of 203 veterinarians, 155 (76.4%) indicated that they used bacterial culture and susceptibility testing, and only 26 (16.8%) frequently did so. In that study,22 obstacles to the use of bacterial culture and susceptibility testing included cost, owner agreement, and time, whereas nonhealing wounds were a motivating factor for ordering the testing. Similarly, a study23 of 184 veterinarians shows that 62 (33.7%) respondents indicated cost was the major barrier for performing diagnostic testing, specifically bacterial culture and susceptibility testing. We believe that efforts to remove barriers to these tests could be beneficial in reducing inappropriate AMD use and AMR; however, it appears that until cost-effective, time-efficient tests become available and clients accept these recommendations, use of diagnostic testing will remain low.
To validate responses to the hypothetical scenarios in the survey of the present study, data were collected from a subset of participants who voluntarily kept a prospective prescription diary. This approach offered a method for us to compare results between theoretical AMD treatment recommendations reported by survey respondents and real-life practice AMD treatment reported by diary respondents. Although the survey component only included respiratory disease scenarios, any disease affecting production animals could have been included in the diary component. Furthermore, the diary allowed for more variation in group sizes, husbandry practices, and herd health statuses. Although there were absolute differences in results for reported proportions of animals treated with AMDs for the diary component versus the survey component, differences in AMD treatment reported by respondents grouped by their selected production setting or reported years as veterinarians were consistent between the 2 measurement modalities. The differences in results (eg, proportion of respondents who indicated use of diagnostic sample testing and the proportion who indicated AMD treatment of the exposed group) between the 2 measurement modalities (online survey component and diary component) used in the present study could indicate that the survey tool may have underestimated actual AMD use and overestimated use of diagnostic sample testing to confirm AMD choice. However, the differences between the tools may have been secondary to factors related to sample composition of diary respondents. Further, differences in results for the 2 measurement modalities could have been attributed to the variety of production animal disease situations encountered by the diary respondents.
In the present study, most (126/157 [80.3%]) respondents indicated that they agreed that government regulations influenced their AMD prescription practices. This finding was not surprising, given potential consequences of violating these regulations. Few respondents (35/157 [22.3%]) reported that increased government regulation on AMD use could be effective in preventing AMR.
Our findings also indicated that client pressures influence AMD use decisions. This finding was consistent with a previous qualitative study24 that describes that client demand influences AMD prescription practices and that veterinarians would comply with producers' requests to maintain client satisfaction in a competitive environment and to avoid possible conflict in discussing how herd health instead could be improved through better husbandry practices. However, in that study,24 challenges faced by the Dutch veterinarians might have been different from those faced by respondents of the present study. The use of a qualitative approach to better understand how client pressure influences veterinary AMD prescription practices in the United States would be valuable because such results might be used to better mitigate issues and encourage appropriate AMD use. For cattle producers, Beef Quality Assurance and the Farmers Assuring Responsible Management are existing programs that emphasize best practices in areas of cattle production, including AMD use,25,26 and have been proven effective in optimizing herd health management. These programs provide veterinarians with ways to communicate appropriate AMD use and improve herd health management. In addition, our findings indicated that veterinarians point to producer compliance as a contributing factor for AMR, and we believe that continued promotion, evaluation, and adaptation of programs like the Beef Quality Assurance and Farmers Assuring Responsible Management should be emphasized to improve producer compliance.
Although most respondents (127/157 [80.9%]) in the present study reported that they considered AMD prescription practices in veterinary medicine to contribute to AMR, more respondents (148/157 [94.3%]) reported that they considered AMD use in human medicine a contributing factor. Relatedly, a recent German study27 shows that most surveyed veterinarians (52/60 [87%] and 51/60 [85%]) considered AMD use in human outpatient and inpatient medicine, respectively, to be the main contributors to AMR. Our findings that respondents in the present study most commonly reported owner and producer compliance, client expectations, and veterinary prescription practices as veterinary-related factors that contributed to AMR suggested that veterinarians may be most amendable to AMR mitigation efforts designed to improve client compliance, educate clients about judicious AMD use principles and AMR, and educate veterinarians and veterinary students about improving veterinary AMD prescription practices. Veterinarians' recognition and acceptance of AMR as a public health issue is an important factor in reducing AMR, and our findings indicated that veterinarians are concerned about AMR and believe there are strategies to address it. For instance, respondents in the present study commonly indicated that effective efforts to reduce AMR could be focused on environmental modifications to prevent disease. This finding was consistent with a study19 that shows veterinarians agreed that AMD use could be reduced through improved management practices by producers.
There were several limitations to the present study. The sample size was relatively small, and we suspected that the topic itself contributed to a low response because AMD use and AMR can be sensitive issues for veterinarians. In addition, few respondents selected the swine or poultry production setting disease scenarios. The use of a cross-sectional survey design with disease-specific scenarios meant that prescription practices were self-reported and hypothetical, and respondents may have viewed AMD use and AMR differently than nonresponders, resulting in bias. The production setting disease scenarios did not account for the many challenges faced by veterinarians in practice, including client demands, economic barriers, and market factors. Although all respondents were invited to participate in the diary component of the study, only 22 did, and differences between respondents who did versus did not participate in the diary component were unknown.
The present study yielded data on AMD prescription practices in production animal settings, helping to fill an important knowledge gap and providing baseline information for future assessments. Our results provided insight into production animal veterinarians' perceptions and attitudes on AMR, prescription practices for AMDs, and influencing factors for AMD use. Findings of the present study could be used to inform development of trainings, tools, and other resources needed to continue to combat AMR and mitigate its development from AMD use in veterinary medicine.
Acknowledgments
Funded in part by the CDC Epidemiology and Laboratory Capacity for Infectious Diseases Cooperative Agreement (No. CK14–1401).
The authors declare that there were no conflicts of interest.
ABBREVIATIONS
| AMD | Antimicrobial drug |
| AMR | Antimicrobial resistance |
Footnotes
SurveyMonkey, San Mateo, Calif.
Google Forms, Google LLC, Mountain View, Calif.
Fire HD 8 Tablet, Amazon, Seattle, Wash.
SAS, version 9.3, SAS Institute Inc, Cary, NC.
References
1. US CDC. Antibiotic resistance threats in the United States, 2013. Available at: www.cdc.gov/drugresistance/threat-report-2013. Accessed Feb 15, 2018.
2. Lushniak BD. Antibiotic resistance: a public health crisis. Public Health Rep 2014;129:314–316.
3. Barriere SL. Clinical, economic and societal impact of antibiotic resistance. Expert Opin Pharmacother 2015;16:151–153.
4. Holmes AH, Moore LS, Sundsfjord A, et al. Understanding the mechanisms and drivers of antimicrobial resistance. Lancet 2016;387:176–187.
5. Landers TF, Cohen B, Wittum TE, et al. A review of antibiotic use in food animals: Perspective, policy and potential. Public Health Rep 2012;127:4–22.
6. World Health Organization. Critically important antimicrobials for human medicine. 6th revision. Geneva: World Health Organization, 2019.
7. Scott HM, Midgley G, Loneragan GH. Antimicrobials in animal agriculture: parables and policy. Zoonoses Public Health 2015;62(suppl 1):3–9.
8. FDA. The judicious use of medically important antimicrobial drugs in food-producing animals. Guidance for industry #209. Rockville, Md: US Department of Health and Human Services, FDA, Center for Veterinary Medicine, 2012.
9. FDA. New animal drug and new animal drug combination products administered in or on medicated feed or drinking water of food-producing animals: recommendations for drug sponsors for voluntarily aligning product use conditions with GFI #209. Guidance for industry #213. Rockville, Md: US Department of Health and Human Services, FDA, Center for Veterinary Medicine, 2013.
10. FDA. Veterinary feed directive regulation questions and answers (revised). Draft guidance for industry #120. Rockville, Md: US Department of Health and Human Services, FDA, Center for Veterinary Medicine, 2015.
11. Carmo LP, Schüpbach-Regula G, Müntener C, et al. Approaches for quantifying antimicrobial consumption per animal species based on national sales data. a Swiss example, 2006 to 2013. Euro Surveill 2017;22:30458.
12. Bondt N, Jensen VF, Puister-Jansen LF, et al. Comparing antimicrobial exposure based on sales data. Prev Vet Med 2013;108:10–20.
13. USDA. Dairy 2014 health and management practices in US Dairy operation, 2014. Report 696.0218. Fort Collins, Colo: USDA, APHIS, Veterinary Services, Center for Epidemiology and Animal Health, National Animal Health Monitoring System, 2018.
14. USDA. Feedlot 2011 part IV: health and health management on US feedlots with a capacity of 1,000 or more head. Report 626.0313. Fort Collins, Colo: USDA, APHIS, Veterinary Services, Center for Epidemiology and Animal Health, National Animal Health Monitoring System, 2013.
15. USDA. Swine 2012 part II: reference of swine health and health management in the United States, 2012. Report 676.0216. Fort Collins, Colo: USDA, APHIS, Veterinary Services, Center for Epidemiology and Animal Health-National Animal Health Monitoring System, 2016.
16. USDA. Feedlot 2011 part I: management practices on US feedlots with a capacity of 1,000 or more head. Report 638.0913. Fort Collins, Colo: USDA, APHIS, Veterinary Services, Center for Epidemiology and Animal Health, National Animal Health Monitoring System, 2013.
17. USDA. Dairy 2007 part III: reference of dairy cattle health and management practices in the United States, 2007. Report N482.0908. Fort Collins, Colo: USDA, APHIS, Veterinary Services, Center for Epidemiology and Animal Health, National Animal Health Monitoring System, 2008.
18. Obritzhauser W, Trauffler M, Raith J, et al. Antimicrobial drug use on Austrian dairy farms with special consideration of the use of “highest priority critically important antimicrobials”. Berl Munch Tierarztl Wochenschr 2016;129:185–195.
19. Cattaneo AA, Wilson R, Doohan D, et al. Bovine veterinarians' knowledge, beliefs, and practices regarding antibiotic resistance on Ohio dairy farms. J Dairy Sci 2009;92:3494–3502.
20. Speksnijder DC, Jaarsmac DA, Verheij TJ, et al. Attitudes and perceptions of Dutch veterinarians on their role in the reduction of antimicrobial use in farm animals. Prev Vet Med 2015;121:365–373.
21. Gibbons JF, Boland F, Buckley JF, et al. Influences on antimicrobial prescribing behavior of veterinary practitioners in cattle practice in Ireland. Vet Rec 2013;172:14.
22. Fowler H, Davis MA, Perkins A, et al. Survey of veterinary antimicrobial prescribing practices, Washington State 2015. Vet Rec 2016;179:651.
23. Hardefeldt LY, Gilkerson JR, Billman-Jacobe H, et al. Barriers to and enablers of implementing antimicrobial stewardship programs in veterinary medicine. J Vet Intern Med 2018;32:1092–1099.
24. Speksnijder DC, Jaarsma AD, van der Gugten AC, et al. Determinants associated with veterinary antimicrobial prescribing in farm animals in the Netherlands: a qualitative study. Zoonoses Public Health 2015;62(suppl 1):39–51.
25. Beef Quality Assurance. Antibiotic stewardship for beef producers. Jan 2019. Available at: www.bqa.org/Media/BQA/Docs/bqa_antibiotics_final.pdf. Accessed Mar 12, 2019.
26. Farmers Assuring Responsible Management. FARM antibiotic stewardship page. Available at: nationaldairyfarm.com/dairy-farm-standards/antibiotic-stewardship/. Accessed Mar 12, 2019.
27. Schneider S, Salm F, Vincze S, et al. Perceptions and attitudes regarding antibiotic resistance in Germany: a cross-sectoral survey amongst physicians, veterinarians, farmers and the general public. J Antimicrob Chemother 2018;73:1984–1988.
