Management Practices and Reported Antimicrobial Usage on Conventional and Organic Dairy Farms¹

Study farms (n = 131) were recruited from dairy herds located in the states of Michigan, Minnesota, New York, and Wisconsin between May 2000 and March 2001. Recruitment methods differed slightly among states and between farm types [conventional (CON) and organic (ORG)]. Conventional herds were identified by obtaining a list of dairy farms licensed to ship milk from the department of agriculture of each respective state. Organic herds were required to be certified by an organic certifying agency for at least 3 yr. There is no central database of ORG herds, so lists of ORG farms were obtained through personal contacts, organic milk cooperatives, and independent organic certifying organizations. In some states, all known ORG herds within each state were contacted by phone to determine eligibility and interest in participation in this study. In Minnesota and Wisconsin, letters were sent to ORG herds to solicit participation in this study.

Conventional herds were contacted by letters (500 to 571 per state for a total of 2102 letters) mailed to randomly selected herds located in counties near each university. Nonresponders received multiple mailings in order to increase participation rate. The letters included a postage paid return postcard. Farms that indicated willingness to participate in the study (n = 295) were screened to ensure that they met predefined criteria for eligibility: preweaned calves and heifers raised on the farm property, a herd record system, unique identification for all cows and calves, >90% of the herd population was Holstein, the farm shipped milk year round, and the farm owner agreed to participate in the study. The absence of any single eligibility criterion resulted in exclusion from the study. The four herd size categories of 30 to 49 cows (n = 21), 50 to 99 cows (n = 42), 100 to 199 cows (n = 33), and 200 or more cows (n = 35) were filled by random selection from eligible herds.

A 64-question survey instrument (available on request) including inventory and herd size (3 questions), herd expansion status (2 questions), housing (5 questions), feed and water system (10 questions), calf management and feeding (12 questions), production and health (10 questions), manure management (9 questions), and antimicrobial use (13 questions) was administered. The questionnaire was adapted from a previously published survey (). For many questions, producers could select all options that applied specifically to their farms. An individual animal area was defined as an area where animals could not commingle with other dairy animals. A multiple animal area was defined as an area where animals did have opportunities for direct contact with other dairy animals. Diets of dry cows were recorded separately from diets fed to lactating cows. An additional 10 questions (addressing the separation and sale of animals that received antimicrobials and type of products used for treatments) were included on the questionnaire administered to ORG dairy farms. To reduce recall bias, all questions referred to practices used in the previous 60 d (unless otherwise noted). For one question, we asked study participants to estimate the percentage of cattle that had received at least one antibiotic injection or oral dose of antibiotics within the previous 60 d for adult cows (milking or dry), bred heifers, and heifer calves (preweaned or weaned). This question did not include the use of intramammary products or topical administration of antibiotics but did include injectable and oral treatments given by farm workers or veterinarians. A separate question addressed the use of intramammary dry cow therapy. The questionnaire was pretested by administration to several farmers and 2 faculty members located in Minnesota. The questionnaire took approximately 45 min to 2 h to complete and was administered during a farm visit. Data were collected during a meeting with the farmer or herdsperson and included an interview, review of product labels (i.e., milk replacer and antibiotics), and information obtained during a farm tour.

Data from all states were entered into a central database for analysis. Data were stratified by herd size (≤120 total cattle = small, >120 total cattle = large) for initial data analysis. There was no significant difference in the number of cows between ORG and CON herds for small herds (89.9 cows [CON], 84.0 cows [ORG]; P = 0.41). Large CON herds contained more cattle compared with large ORG herds (398.0 cows [CON], 262.9 cows [ORG]; P = 0.02). The data were stratified by herd size (small or large) with herd type (ORG or CON) forming the rows and the response variable (yes or no for specified antibiotics) forming the columns and were analyzed using the Cochran-Mantel-Haenszel option in PROC FREQ (). Herd size was not associated with antibiotic usage so data were combined for further analysis. The PROC FREQ chi-square analysis and Fisher's exact test () were used to identify associations between herd type and the use of selected antibiotics or other treatments. Continuous variables were analyzed using PROC TTEST and PROC MEANS ().

Herd enrollment was distributed among the participating states: 30 (CON) and 2 (ORG); 23 (CON) and 10 (ORG); 25 (CON) and 9 (ORG); and 21 (CON) and 11 (ORG) for Michigan, Minnesota, New York, and Wisconsin, respectively. The distribution of enrollment by herd type was: 12 (ORG) and 9 (CON); 12 (ORG) and 30 (CON); 4 (ORG) and 29 (CON); and 4 (ORG) and 31 (CON) for herd size categories of 30 to 49, 50 to 99, 100 to 199, and >200 total cows, respectively.

Organic herds were smaller and produced less milk than CON herds (Table 1). One ORG herd was allowed to remain in the study even though it contained only 26 adult dairy cows on the day of the herd visit. Higher SCC values were reported for proportionally more ORG herds compared with SCC values reported by CON herds (Table 1). A larger proportion of cows in CON herds were milked in a pit parlor, whereas a larger proportion of cows in ORG herds were milked in a tie stall or stanchion barn (Table 1).

Preweaned calves in ORG herds were housed in areas that permitted less contact with other animals, but there was no significant association between herd type and specific housing types such as the housing of preweaned calves nor with the housing of weaned calves (Table 2). However, significantly more CON herds housed lactating cows in free stalls and had maternity housing separate from lactating cows compared with ORG herds (Table 2). Calves in more CON herds were fed individually (i.e., no access to group feeding such as shared troughs) compared with ORG herds. The use of separate facilities to house sick animals separately from lactating cows and the washing of calf pens were not influenced by herd type.

Nutritional management was affected by herd type. A significantly greater percentage of CON herds fed lactating cows a TMR (79.8 vs. 31.3%, P < 0.001), a transition ration (71.7 vs. 31.3%, P < 0.001), and anionic salts (33.3 vs. 3.1%, P = 0.001) to close-up cows. Whole soybeans and soybean meal were the only purchased feeds reported on ORG dairies, but significantly less ORG (46.9%) compared with CON (90.9%) herds used them (P < 0.001). The use of whole cottonseed or hulls, brewer's byproducts, and tallow or animal fat was not reported for ORG dairies, whereas these products were used on 50.5, 48.5, and 60.6% of CON farms, respectively (P < 0.001). No organic farmers reported the use of meat and bone meal or blood meal, whereas these products were used on 27.3% of CON herds (P < 0.02).

Individual automatic water bowls were used on significantly more ORG herds (Table 3). Occasional use of surface water by milking cows and primary use of surface water by dry cows occurred in significantly more ORG herds (Table 3). Chlorinated drinking water was used on significantly more CON dairies. There was no significant association of herd type with group access to individual automatic water bowls, the use of water tanks for multiple cows, occasional use of surface water by dry cows, and use of surface water as main water sources for milking cows. Similarly, there was no significant association between water source (well water, municipal source, or surface water) and herd type.

Physical contact between poultry and dairy cattle was reported on significantly more ORG herds (37.5%) compared with CON herds (11.1%; P = 0.001). Herd type was not associated with physical contact with beef cattle, horses, pigs, sheep, goats, dogs, cats, or wild geese (P > 0.20).

A treated cow was defined as a cow that had received antibiotics that required a withholding period that had not yet ended. Whole milk obtained from treated cows was fed to calves on significantly more CON herds (38.4% [CON] vs. 0.0% [ORG], P < 0.001), whereas whole milk obtained from untreated cows was fed to calves on more ORG herds (25.3% [CON] vs. 93.8% [ORG], P < 0.001). The majority of farms fed ≥3.78 L of colostrum to newborn calves (80.0% [CON] vs. 63.4% [ORG], P = 0.29).

Diagnosis of Johne's disease was more commonly reported on CON farms (48.5%) compared with ORG farms (25.0%; P = 0.02). There was no significant association between recall of previous diagnosis of Salmonella or bovine viral diarrhea and herd type (P > 0.11). The use of coliform mastitis vaccines was more common on CON (47.5%) compared with ORG dairy farms (15.6%; P = 0.001).