The Oklahoma Cow-Calf Biosecurity Survey was developed by Oklahoma State University through funding from the United States Department of Agriculture, Animal and Plant Health Inspection Service (USDA APHIS), National Animal Disease Preparedness and Response Program (NADPRP). The survey was administered through a contract with the USDA National Agricultural Statistics Service (NASS). The stratified proportionate random sample of 5,000 producers was obtained from the USDA NASS beef cattle list frame and represented a statistically valid sample of the state’s beef cattle producer population. In January 2022, producers received a study selection postcard, followed 2 weeks later by a mail survey with a postage-paid return envelope. Two weeks after the mail survey, producers were contacted via phone with the option to complete the survey over the phone rather than return the paper survey. The data collection process was completed in February 2022. Of the 5,000 producers contacted, 1,466 surveys were completed, either by mail or via phone, resulting in a 29 percent response rate and a 95 percent confidence level associated with the sample.

Screening questions were used to filter out producers who did not actively manage cattle in the 2021 calendar year and those not involved in the cow-calf¹ production stage. The resulting pool of 981 producers actively managed beef cows and produced calves in 2021. To put this sample in context, the 2022 USDA Census of Agriculture reported 4.5 million beef cattle (2.0 million beef cows) on 43,223 ranching operations (39,338 cow-calf operations) in Oklahoma (24). Oklahoma is a top 5 beef cattle production state, with an average beef cow herd size of 51 beef cows in Oklahoma as compared to a national average of 47 head. Respondents shared producer and cattle operation demographics, current herd management regarding disease testing and vaccinations, knowledge and use of biosecurity elements and the Secure Beef Supply Plan, and disease familiarity and threat perceptions.

Means² for binary variables in Table 1 (e.g., herd size categories) can be interpreted as the percent of the respondents who responded ‘yes’ to the question, and means for continuous variables are the average response to the question across respondents (e.g., the portion of respondents who have not heard the definition of biosecurity). Continuous biosecurity practice variables in Table 2 are mean rates of biosecurity adoption. The local N—or the number of observations remaining after removing respondents that left the question blank when filling out the survey are also included for each variable. All summary statistics are unweighted sample means; no population means were calculated.

Probit analysis is used when the study population makes a discrete choice. In this case, the discrete choice is the choice to vaccinate against respiratory disease by Oklahoma cow-calf producers. The left-hand side, or dependent variable, of a probit analysis takes the form of 0/1 where 1 indicates that the respondent chose to vaccinate. The coefficients of the right-hand side variables, or independent variables, give insight into the factors that increase or decrease the probability that cow-calf producers would choose to vaccinate. The data were tested for normally distributed errors, and we failed to reject the null hypothesis that errors were normally distributed. In combination with our large sample size, we concluded that the probit model was appropriate for this discrete choice analysis.

This analysis explored independent variables associated with various herd characteristics, production practices, and alternative herd health management practices like biosecurity that may influence vaccination according to the literature. Because different factors may influence respiratory vaccination of breeding cattle in comparison to calves, two separate regressions were developed. The dependent variable for the calf regression is “respiratory vaccination of calves” ( C a l f RespVac ) which takes a value of 1 if the respondent vaccinated and 0 if they did not. Equation 1 does not include breeding herd variables, except the variable indicating whether the breeding herd receives respiratory vaccines. A second regression has the dependent variable “respiratory vaccination of the breeding herd” ( Breedin g RespVac ) which takes a value of 1 if the respondent vaccinated and 0 if they did not. Equation 2 includes no calf variables, except the variable indicating whether calves receive respiratory vaccines. The estimated equations are:

To simplify Equations 1, 2 visually, independent variables are clustered into vectors shown in Table 3, with variable definitions in Tables 1, 2. The vector of variables associated with education and producer knowledge (X_ED) included variables based on producer knowledge of the disease. The administration and herd management (X_Mgmt) vector included variables in which a producer was asked about the administration of vaccinations, testing, and record keeping. The biosecurity vector (X_Bio) has variables based on biosecurity elements and familiarity with the definition of biosecurity and recommendations of the Secure Beef Supply (SBS) Plan. Finally, a vector of control variables other than geographic region was included (X_Control). Regional variables (X_NE, X_SE, X_SW) control geographic variability with the Northwest region in the intercept. We included regional control variables for differences in weather and ecosystems across the state. The intercept is denoted by β₀.

A probit model regression gives estimated coefficients ( β 1 , β 2 , β 3 , β E D , β Mgmt , β B i o , β control ) in the form of z-scores, which can be hard to interpret into something meaningful. Transforming coefficients of the regression into marginal values using the margins function in R resulted in coefficients that are easier to interpret. Marginal values are partial derivatives of the regression with regard to the other variables. In other words, marginal effects illustrate how a small change in one independent variable, holding other independent variables constant, affected the probability of a producer vaccinating either calves or the breeding herd. These margins were interpreted as predicted probabilities (25). Results discussion focuses on the sign of the marginal value (i.e., did it increase or decrease the predicted probability of vaccination) and the magnitude (i.e., did it increase the predicted probability more or less as compared to other independent variables). Error terms (ε) were assumed to be normally distributed in probit models.

Respiratory vaccination for calves was utilized by 73.5 percent of respondents (Figure 1). Most producers in the state vaccinated calves twice (37.3 percent of all respondents), and the most common timing was “at branding or 1 to 3 months old” (43.5 percent) and “at weaning” (49.6 percent) (Figure 2). The breeding herd was vaccinated by 47.9 percent of respondents. Of those who vaccinated the breeding herd 34.4 percent used a killed vaccine and 21.3 percent used a modified live vaccine (26.5 percent of respondents did not specify vaccine type). Variations in weather, stocking rates, and cattle movements across the state likely influenced the risk of respiratory disease.

Differences in vaccination rates by vaccination type (RVX for respiratory vaccines and CVX for clostridial vaccines) and animal type (‘calves’ and ‘breeding’ for all beef cows, bulls and weaned replacement heifers) reveal two patterns (Figure 3). First, calves were vaccinated at a higher rate than the breeding herd for both respiratory (75.6 percent RVX calves and 53.8 percent RVX breeding) and clostridial diseases (89.1 percent CVX calves and 62.8 percent CVX breeding). Second, clostridial vaccinations for calves and the breeding herd were implemented at a noticeably higher rate than respiratory vaccinations. Regionally, the northern half of the state had higher rates of respiratory vaccination. Within this context, factors affecting calf and breeding herd vaccination rates were explored.

A somewhat unique characteristic of the state is that some cow-calf producers retain calves after weaning to gain value by grazing on small grains—primarily hard red winter wheat—in the winter. Selling calves at weaning was still the most common (55.2 percent of respondents), but 35.2 percent grazed for 30 days post-weaning and 19.4 percent retained calves through the stocker/backgrounding stage. Retaining calves through finishing and selling direct beef to consumers (freezer beef) was becoming more common, with 14.8 percent of respondents routinely participating in this activity in the last 5 years. Biosecurity needs may be different for producers bringing additional calves in to graze in the winter as compared to only grazing their own calves or selling calves at weaning.

Testing and vaccination results indicated that testing rates for BVD-PI animals are relatively low for cows (16.8 percent), bulls (22.7 percent), and non-breeding stock (8.2 percent) (Table 1). Vaccination can be combined, in a variety of ways, with testing before new herd additions enter the farm. Producers were asked to report whether they test for Johne’s disease, bovine tuberculosis, brucellosis (males and adult females), and whether they use testing, vaccination, or both for brucellosis (heifers), BVD, respiratory disease (IBR, PI3, BRSV), trichomoniasis, and leptospirosis. There were a total of 15 testing or vaccination combinations. The Vac-Test variable in Table 1 was the average number of vaccination and testing practices for respiratory disease that the producer required. The average value of 45 percent indicated that a producer desired between 6 and 7 vaccines and/or tests from the list above. Record keeping for herd medical treatment was reported in 47.3 percent of operations for calves and 49.0 percent of operations for the breeding herd (Table 1).

Perceived risk associated with disease has been another reason to adopt vaccination programs. Producers were asked to assess the threat of BRD, and more specifically BVD on their own operation. BVD was anticipated to be the most currently discussed viral disease of the BRD complex by producers in Oklahoma. Results were groups into those that felt BRD was a moderate to large threat (BRDp threat), a low threat or no threat (BRDp no threat), or an “unknown” threat level (BRDp uk). Similar questions were asked regarding specific producer perceptions about perceived BVD threat levels (BVDp threat, BVDp no threat, BVDp uk). For both BRD and BVD, a higher percentage of producers did not perceive them as a threat to their operation (47.6 percent BRDp no threat and 45.8 percent BVDp no threat).

A second series of questions asked about the threat of BRD, and BVD specifically, to the beef industry in general. Producers perceived disease to be a greater threat to the industry (58.0 percent BRDi threat and 49.6 percent BVDi threat) in general than to their own cow-calf operation, which may have reflected the large losses experienced in the feedlot sector.

Understanding that both calf and cow herd vaccinations are a part of comprehensive biosecurity plans, producers were provided with the definition of biosecurity. As stated in the introduction, for this survey, biosecurity referred to everything that was done to keep diseases and the pathogens that carry them (including viruses, bacteria, fungi, parasites, and other microorganisms) away from livestock, property, and people. Statewide (Table 1), most producers had either not heard of the definition (30.7 percent, Bio not heard) or had heard of it but had not implemented biosecurity into their operation (56.0 percent, Bio not used). Only 13.3 percent of producers had implemented some level of biosecurity into their operation (Bio implemented). Producers were also asked about their familiarity with the recommendations of the Secure Beef Supply Plan. Only 15.4 percent of the producers had heard of the Secure Beef Supply Plan (SBS heard used) and had some level of implementation of it in their operation. Adoption rates represent what producers’ reported adoption of specifically described biosecurity practices for the breeding herd and for the calf crop. It is noted that some producers potentially incorporated management practices with biosecurity benefits for other reasons such as profitability, without necessarily acknowledging the biosecurity benefits fully. For instance, producers may have vaccinated cattle for cost reduction or to capture market premiums without full understanding of the biosecurity implications.

The SBS or Beef Quality Assurance (BQA) guidelines for a written biosecurity plan were broken down into 22 individual activities for this survey. Producers were asked to indicate biosecurity activity implementation (yes = 1, no = 0) on their operation. A percentage of biosecurity plan activities adopted by each respondent was calculated (Table 2). Across all respondents, only 4 biosecurity plan activities were adopted by more than half of the producers; this corresponds to an average proportion of biosecurity plan adoption of 18.2 percent (Bioplan Elements). Further, if a producer answered “no” for a particular activity, they were asked to select one of 9 reasons why they did not adopt it. The most common reason an activity was not adopted was due to lack of familiarity with the activity (BP not familiar), followed by the producer feeling they did not have enough cattle (BP cattle) to make it worthwhile.

Finally, a typical producer in the state was between the age of 65–74, received 1 to 20 percent of their household income from the cattle operation, and had an education level higher than a high school degree (Table 1). Demographics were used to control for inherent differences between producers.

Probit analysis was completed using the statistical software R. Due to the large number of independent variables, statistical checks were incorporated. First, model fit was examined using the Bayesian Information Criterion (BIC) and Akaike Information Criterion (AIC). The final calf vaccination regression included 638 local observations and 41 variables. While this is a large number of variables, it resulted in the best model specification fit statistics. The breeding herd regression included 526 local observations and 45 variables. Second, variance inflation factors (VIF) were examined to assess the correlation between independent variables, with no variables included in the final regression exhibiting VIF values over 4. Both models were statistically significant based on Likelihood Ratio Tests and Wald tests.

As noted earlier, our focus in this results discussion will be on the sign and magnitude of each marginal effect relative to other marginal effects. There were only five variables of statistical significance (Table 4, in bold), all with a positive effect on a producer’s decision to vaccinate their calves for respiratory disease. Clostridial vaccination of calves, respiratory vaccinations of the breeding herd, and larger herd sizes were all significant, positive influencers on a producer’s decision to vaccinate calves against respiratory disease. The use of clostridial vaccinations in the calves increased the predicted probability that a producer vaccinated their calves for respiratory disease by 0.2421. The use of respiratory vaccination on the breeding herd increased the predicted probability that a producer vaccinated their calves for respiratory disease by 0.2131. The probability of vaccinating calves increased with herd size. A herd size of 50 to 99 head increased the predicted probability of calf respiratory vaccination by 0.1244 as compared to those with very small (1 to 25 head) herds. Likewise, a herd size of 100 to 249 head increased the predicted probability that calves received respiratory vaccination by 0.1367, and herd sizes of 250 head or more increased the predicted probability by 0.1789 compared to a herd of 1 to 25 head.

Shifting to breeding herd results, respiratory vaccination of calves, clostridial vaccination of the breeding herd, keeping medical records on the breeding herd, education, producer perception of BRD in the industry, and a producer’s decision to not adopt biosecurity plan elements due to cost were all significant influences on a producer’s decision to vaccinate their breeding herd for respiratory disease (Table 5). The use of respiratory vaccinations in calves increased the predicted probability of vaccinating the breeding herd for respiratory disease by 0.3783. The use of clostridial vaccinations in the breeding herd increased the predicted probability of also vaccinating them for respiratory disease by 0.1714. A producer who kept written medical records on the breeding herd had an increased predicted probability of vaccinating their breeding herd for respiratory disease by 0.0919. Producers with a secondary educational degree had a higher predicted probability of vaccinating their breeding herd for respiratory disease by 0.0899.

However, if a producer did not know what BRD was, the predicted probability of vaccination declined. If a producer responded “unknown” when asked to what extent BRD is a threat to the industry, the predicted probability of vaccinating their breeding herd for respiratory disease decreased by 0.1895. In addition, if a producer found biosecurity plan adoption to be too expensive, the predicted probability of vaccination adoption in the breeding herd declined by 0.6463. Overall, this suggests that a producer who participated in some good herd health management practices, including biosecurity practices, was more likely to vaccinate their breeding herd for respiratory disease.