Factors affecting the use of biosecurity measures for the protection of ruminant livestock and farm workers against infectious diseases in central South Africa

Abstract Biosecurity measures have been introduced to limit economic losses and zoonotic exposures to humans by preventing and controlling animal diseases. However, they are implemented on individual farms with varying frequency. The goal of this study was to evaluate which biosecurity measures were used by farmers to prevent infectious diseases in ruminant livestock and to identify factors that influenced these decisions. We conducted a survey in 264 ruminant livestock farmers in a 40,000 km2 area in the Free State and Northern Cape provinces of South Africa. We used descriptive statistics, to characterize biosecurity measures and farm attributes, then multivariable binomial regression to assess the strength of the association between the attributes and the implementation of biosecurity measures including property fencing, separate equipment use on different species, separate rearing of species, isolation of sick animals, isolation of pregnant animals, quarantine of new animals, animal transport cleaning, vaccination, tick control and insect control. Ninety‐nine percent of farmers reported using at least one of the 10 biosecurity measures investigated (median [M]: 6; range: 0–10). The most frequently used biosecurity measures were tick control (81%, 214 out of 264), vaccination (80%, 211 out of 264) and isolation of sick animals (72%, 190 out of 264). More biosecurity measures were used on farms with 65–282 animals (M: 6; odds ratio [OR]: 1.52) or farms with 283–12,030 animals (M: 7; OR: 1.87) than on farms with fewer than 65 animals (M: 4). Furthermore, farmers who kept two animal species (M: 7; OR: 1.41) or three or more species (M: 7) used more biosecurity measures than single‐species operations (M: 4). Farmers with privately owned land used more biosecurity measures (M: 6; OR: 1.51) than those grazing their animals on communal land (M: 3.5). Farms that reported previous Rift Valley fever (RVF) outbreaks used more biosecurity measures (M: 7; OR: 1.25) compared with farms without RVF reports (M: 6) and those that purchased animals in the 12 months prior to the survey (M: 7; OR: 1.19) compared with those that did not (M: 6). When introducing new animals into their herds (n = 122), most farmers used fewer biosecurity measures than they did for their existing herd: 34% (41 out of 122) used multiple biosecurity measures like those of vaccination, tick control, quarantine or antibiotic use, whereas 36% (44 out of 122) used only one and 30% (37 out of 122) used none. Certain farm features, primarily those related to size and commercialization, were associated with more frequent use of biosecurity measures. Given the variation in the application of biosecurity measures, more awareness and technical assistance are needed to support the implementation of a biosecurity management plan appropriate for the type of farm operation and available resources.


INTRODUCTION
Agriculture plays a key role in the growth of African countries and there is an immense need for stability and improved productivity in this sector (Audibert, 2010). Sustainability issues have been raised in the livestock production subsector as the global demand for protein-rich diets is forecasted to continue to increase (Delport et al., 2017;Department of Agriculture Forestry and Fisheries Republic of South Africa, 2019; Organisation for Economic Co-operation Development and the Food and Agriculture Organization of the United Nations, 2020). The livestock sector is a major employer and contributes substantially to food security in South Africa (Meissner et al., 2013). Livestock farming is a critical economic driver and provides the sustenance for most nonmetropolitan towns and rural communities (Meissner et al., 2013). In many rural areas, the use of pasture is a common herd management practice for ruminants (Palisson et al., 2017). The most limiting factors to pasture-based livestock production in South Africa have been animal diseases, land rights and access, inadequate knowledge of livestock and pasture management and climate variability (Oduniyi et al., 2020). One of the most sustainable ways to protect against threats of infectious diseases and reduce their economic costs is through the implementation of biosecurity (Oliveira et al., 2017).

The World Organisation for Animal Health (OIE) Terrestrial Animal
Health Code (2019) defines biosecurity as 'the set of managerial and physical measures designed to reduce the introduction, establishment and spread of animal diseases, infections or infestations to, from and within an animal population' . Livestock farmers are responsible for and directly benefit from the continued implementation of biosecurity on their farm (Manuja et al., 2014). Prioritizing animal health and welfare improves productivity, enhances resilience to climate change and natural disasters, such as drought and has economic benefits (Vallat, 2015;Lubroth et al., 2017). A robust biosecurity program improves animal welfare by keeping more animals healthy and resistant to environmental factors; it also serves as the first line of defence by detecting diseases early and limiting disease spread within the farm (Kriel, 2018), reduces the costs of treating diseases, helps to ensure production of high-quality, safe, nutritious products and improves the chances of running a successful business and remaining in the agricultural industry (Sinclair et al., 2019). In addition to protecting animal health and its economic benefits, when biosecurity measures are used to limit infectious diseases in livestock, they also directly reduce the risk of zoonotic pathogen transmission to humans and can help to inform specific public health measures (Kimman et al., 2013;Layton et al., 2017). The implementation of farm biosecurity measures should prioritise animal diseases with adverse social, welfare and economic effects and can be designed as multi-hazard mitigation methods (Sternberg-Lewerin et al., 2015;Layton et al., 2017).
Standards and recommendations for a wide range of biosecurity practices for livestock production, either for general disease prevention or to minimise specific infection risks, including zoonotic risks are widely available on the internet (Waage & Mumford, 2008; African Union Inter-African Bureau for Animal Resources, 2015;Windsor, 2017;Robertson, 2019;Van Der Merwe, 2020). Recommended onfarm biosecurity measures include: animal hygiene, sanitation, restrictions on sharing and disinfection procedures for equipment, vehicles and facilities, tick and pest control, vaccination, movement controls and quarantine of newly introduced animals, preventing different groups of animals from mixing, culling of diseased animals, protocols for the handling and treatment of infected animals or contaminated products, feed management, facility and vehicle maintenance and protocols for handling manure and disposing of carcasses.
There are some published studies on the knowledge, attitudes and practices of farmers regarding animal health care and biosecurity in Africa (Simela, 2012;Oladele et al., 2013;Wolff et al., 2019). One recent study examined the determinants of animal health care practices that included vaccination, external and internal parasite control, quarantine of new and isolation of sick animals, restricted access and supplementary feeding in South Africa (Mdlulwa et al., 2021); however, the majority of studies that analyze biosecurity measures used by farmers have been conducted in high income countries Dorea et al., 2010;Sarrazin et al., 2014;Renault et al., 2018;Gunther et al., 2019). The benefits of biosecurity in terms of productivity and profitability have often focused on specific production and management systems, a single biosecurity measure or the prevention of one disease only Laanen et al., 2013;Merrill et al., 2019). Studies on pasture-grazed systems are often neglected.
It is often believed that the choice to implement biosecurity measures by individual farmers is associated with economic constraints and infectious disease awareness (Niemi et al., 2016;Merrill et al., 2019). However, awareness is not necessarily the limiting constraint on practicing biosecurity interventions and some farmers would prefer to pay for treatment rather than prevention and standard biosecurity practices (Dione et al., 2020). Challenges to adopting biosecurity measures for pasture-based systems include financial cost, resulting in fewer and only certain specific measures being implemented, and lack of research (Niemi et al., 2016). Some biosecurity interventions, for example, culling, may cause particular hardship for the livestock owners (Fraser, 2018).
The lack of or inefficient implementation of biosecurity measures and the resultant disease on the farm have downstream effects on all livestock owners who are connected geographically or through market chains. If an outbreak occurs due to an individual landowner's negligence, the surrounding land users are likely to suffer significant financial consequences (Kristensen & Jakobsen, 2011;Knight-Jones & Rushton, 2013).
Data on current, local biosecurity practices are important for veterinary professionals and government in order to support farmers in biosecurity decision-making and identify areas for targeted education to increase the effectiveness of disease control and surveillance. In this study, we examined which and how many biosecurity measures are being implemented by ruminant livestock farmers in the Free State and Northern Cape provinces of South Africa, and evaluated the farm characteristics associated with the likelihood of implementation of biosecurity measures.

Study area
The study was conducted over an area of ∼40,000 km 2 , at 994- Cape is dry savannah, and the southern portion is part of the Great Karoo with grassy shrubland (Köppen & Geiger, 1936). The climate in the study area is mostly temperate semi-arid steppe (Köppen & Geiger, 1936). This area was selected because it had been affected by large RVF outbreaks in the past.

Study design and sampling strategy
Two cross-sectional surveys of farms keeping ruminants were conducted during October 2015-August 2016 and May-November 2017.
The sampling and methodology, including the study farms selection via random geographic points, were described previously by Ngoshe et al.
(2020) and Msimang et al. (2019). We contacted the participating farms from the first survey for the second survey and replaced those that declined to participate again with another farm near the next random geographic point on the list. The sample size of 264 farms provided greater than 80% power to detect a count ratio of 1.2 for a binary predictor of interest, assuming a baseline count of 5 biosecurity measures and an R-squared of 0.25 between the predictor of interest and the other covariates in the model (Supporting Information S2) (Signorini, 1991).

Questionnaire
Participating farmers provided informed, written consent and were The questionnaire was piloted among 17 farmers located just outside the study area in May 2015. The questionnaire was administered to the farm owner or manager in English, Afrikaans or Sesotho using a tablet.
At the time of consent, a farm ID number was assigned and anonymous responses were sent to an internet cloud-based database using the Open Data Kit Application (Hartung et al., 2010). Individual identifying information was not captured by the electronic questionnaire.

Data analysis
The electronic questionnaire data were downloaded and cleaned using RStudio (RStudio Team, 2020). Further data cleaning/structuring was conducted in Microsoft Excel and the data were then imported into Stata 13 (StataCorp, College Station, TX, USA) for the statistical analyses.
For the descriptive analysis, we calculated percentages, medians, ranges and quartiles for each farm and/or animal characteristics, as well as the use of the 10 biosecurity measures. The Fischer's exact test was used to test whether the percentages for each biosecurity measure of private and communal farmers differed significantly. The odds ratios (OR) and 95% confidence intervals (CI) of implementing a biosecurity measure by the farmers with versus without various farm characteristics were estimated using a multivariable maximum likelihood binomial regression analysis, modelling the number of successes (measures implemented) out of the number of trials (possible measures). The number of possible measures was 10 for farmers raising multiple species but only 8 for farmers who raised only one animal species because they did not have the opportunity to implement '(2) keeping different animal species in different/divided areas on the farm' and '(3) having separate equipment for different species' . The assumption of independence in binomial regression was validated by measuring pairwise correlations between biosecurity measures using Phi coefficient calculations where a-d are frequencies of the cells of the 2 × 2 table and  (Table S1) (Simon, 2007). The final binomial regression model was obtained by manual backward elimination, starting with all independent variables that had significant (p < .2) univariable associations using binomial regression and continuing until all remaining variables were significant (p < .05), or if removal of a confounding variable resulted in >10% change in the coefficient of another variable; finally, forward stepwise selection was done to assess previously eliminated variables in order to achieve a final model of significant predictors. were semi-commercial.

Animal movements and purchases
Multiple animal species were permitted to interact at certain or all times on 51% (

Animal losses from abortion
Abortions in farm animals were reported on 25% of farms (67 out

Factors associated with implementation of biosecurity measures
Only relations between the biosecurity measures justified the assumption of independence for the binomial model (Table S1).
The variables 'production system type' and 'land ownership' were collinear as no communal farms classified themselves as having a feedlot or as being commercial, whereas 72% (168 out of 232) of the private farmers reported having a commercial business or feedlot. Therefore, only land ownership was used in the analysis. Similarly, variables regarding rearing of individual species were omitted from the analysis because they were used in the calculation of (and therefore collinear with) number of species reared. In the univariable analysis (Table 2), several farm characteristics were associated (p < .2) with biosecurity measure implementation and were selected for inclusion in the multivariable model ( Figure 4).

F I G U R E 4 Factors associated and odds ratio (OR) values (coloured dots) and
The following factors were associated with a greater odds of implementing of biosecurity measures in the final binomial regression model ( Figure 4)

DISCUSSION
The majority of farmers implemented at least one biosecurity measure on their farm and many used more than six measures, which suggests that farmers in the study had awareness of some biosecurity measures.
Other studies on European private farmers have found that, with the exception of some intensive commercial operations (primarily swine, poultry and ruminant feedlots), most rural livestock producers have a poor understanding of biosecurity (Bellini, 2018 (Mdlulwa et al., 2021) found that 26% of smallholder farmers isolated new or sick animals, and although our findings for quarantine of new animals (21%) agree with that figure, 72% of farmers isolated their sick animals from the rest of the herd.
Nearly all private farmers and most communal farmers reported using fencing around their farm because it provides physical security, such as confining farm animals, facilitating quarantine of new animals and keeping out unwanted animals and people that pose a threat of infectious diseases, predation or theft. We found that the vast majority of adult cattle and sheep were left overnight in the grazing areas on private farms likely due to the safety permitted by fencing, though the condition and maintenance of the fencing were not reported. In contrast, in communal areas, farmers reported keeping their livestock in a corral at night. Some farms did report predation losses. Reducing contact with wildlife is an important biosecurity measure as wild animals are involved in the epidemiology of many livestock and zoonotic diseases and may act as reservoirs for these pathogens (Kruse et al., 2004).
Vaccination was the second most frequently reported biosecurity measure. In contrast to our finding and despite the fact that multipathogen vaccinations are the most cost-effective way to prevent livestock disease, low vaccination rates have been reported amongst smallholder farmers in Africa, Asia and Latin America (Wallace et al., 2013;Donadeu et al., 2019). Rostal et al (2020) found that less than 60% of farmers vaccinated livestock during the RVF epidemic of 2010 in central South Africa, despite vaccination being the most effective measure to prevent loss of livestock and transmission to animal workers (Hartman, 2017).
Low to moderate vaccination levels for brucellosis were also reported by the farmers in our study, and cattle were vaccinated most frequently. Brucellosis is a reportable disease in South Africa  Africa, 2017). Our results suggest that farmers may be aware of the threat of brucellosis, they likely find it difficult to clinically identify. As brucellosis control is a high national priority, the low vaccination rate reported in our study is cause for concern. However, it is possible that farmers who rely on government or private veterinary services may not always be aware of or remember the vaccinations their stock receive.
The findings of low to modest vaccination use against diseases that were highly prevalent and had an impact in the area contrast with findings on increased willingness to invest in vaccination and other biosecurity measures when confronted with a threat (Merrill et al., 2019;Machalaba, 2020). This all suggests that there is still a lot of opportunity to discuss zoonoses control with farmers. More education is required about the risks of general and disease-specific transmission, as well as the prevention of any practices that promote disease spread.
Farmers in our study reported using frequent and routine control measures for ticks. This is an important biosecurity measure with both economic and human health benefits (Jongejan et al., 2020) to prevent common tickborne diseases, such as heartwater in animals, and Q fever and Crimean-Congo haemorrhagic fever in both animals and humans.
Isolation is a measure aimed at preventing the spread of disease from sick to healthy stock, but it must be used in conjunction with disinfection of facilities and working equipment to avoid fomite contamination in order to be effective (Bergström et al., 2012). Our findings show that most farmers are aware that sick animals can spread disease and that farmers have determined that isolating sick animals is an important precaution to protect the health of the remaining herd.
Farmers were much less likely to implement biosecurity measures when introducing new animals into the existing herd, despite the fact that the most common risk factor for the introduction of infectious diseases to a farm is the introduction of new animals (Cuttance & Cuttance, 2014).
Less than 60% of the farmers used vaccination when introducing new animals, compared with the much higher routine vaccination rate discussed earlier (up to 84%). This, combined with farmers reporting that they rarely cleaned and disinfected transport vehicles and did not routinely quarantine new animals (only 13% (16 out of 122) of the farms combined vaccination and quarantine), suggests that these farms are at risk for the introduction of infectious agents. Further, if a third party is used to move the animals and they also do not disinfect the transport vehicles, this may also pose a transmission hazard between farms. Livestock trade at both the local or regional level may contribute to disease spread (Fèvre et al., 2006).
Twelve percent of farmers reported administering antibiotics to animals upon introduction to the farm. Antibiotics are often used among introduced animals when other biosecurity/welfare measures on the farm are insufficient to prevent illness following the stress of transport.
It is important for farmers to incorporate antibiotic stewardship when designing their biosecurity system to ensure their continued efficacy in both animal and human health (Landers et al., 2012;Chantziaras et al., 2014). Although antibiotics are as important for livestock health as they are for human health, this observed prudent use by livestock producers may be due to the farmers' unwillingness to pay for additional biosecurity measures, or on the other hand may be a sign of good husbandry management; this requires more investigation. Vaccination and other non-antibiotic interventions will help minimize the use of antibiotics within livestock populations, but this has not been without its difficulties (Hoelzer et al., 2018). We did not specifically inquire about animal deworming, but some of our farmers mentioned it as another practice they used. Antihelmintic and acaracide resistance, like antibiotic resistance, has become a significant problem in animal production as a result of inappropriate use; Therefore, farmers should be encouraged to invest in developing technical skills or receiving consistent veterinary services, as well as plan biosecurity strategies with veterinary practitioners in order to ensure that the measures implemented are adequate and effective (Hlatshwayo & Mbati, 2005). This also applies when administering any vaccine.
Private landowners likely implemented more measures because farmers who own the land would have more control over the territory as a resource, its condition and access by other farmers, making it easier to implement various biosecurity measures. For instance, only private landowners have a right to fence the farm. In contrast farmers using communal land would have less control over the territory as a resource, its condition and access by other farmers. Using communal land for farming restricts the farmer's ability to diversify their use of the land and expansion of their natural resource base, limiting livelihood improvement (Andrew et al., 2003). Given its history of colonialism and apartheid, South Africa's agrarian structure is still dominated by large-scale farms (Neves, 2020), and because these farmers own the majority of the country's agrarian land, a large number of small-scale farmers have small holdings or use communal land (Aliber et al., 2016). Furthermore, smallholders or communal farmers are often poorer than private large-scale farmers, owing to difficulties in fully participating in formal livestock marketing, which may influence their decision to implement biosecurity measures (Sotsha et al., 2018).
Ruminant livestock farmers may find it more cost effective and beneficial to implement biosecurity measures in large herds, as described in poultry farms. A cost-benefit analysis of poultry farms found that the average cost of biosecurity action is lower per animal in larger operations. This reduction was primarily due to lower labour costs per animal for biosecurity action (Siekkinen et al., 2008). Furthermore, commercial farmers with a larger herd of animals have higher sales returns and hence more financial resources to invest in farm operations, including farm biosecurity.
Farms with two species were likely to employ a broader range of biosecurity measures than single-species farms, which may reduce their vulnerability to species-specific diseases or susceptibility to diseases capable of inter-species transmission. Furthermore, keeping two species may require different types of biosecurity, some of which are more effective with certain animals than others (Kalis et al., 2004;Scagliarini et al., 2012). Biosecurity in farms with more than two versus one species was however not significantly higher which could be because farmers keeping many species are usually less specialized and have poorer overall management, that is, there is confounding due to other unmeasured management variables.
Farmers are aware that introducing newly purchased animals to the herd poses a risk of infectious disease introduction to the farm, since farmers who had purchased animals in the 12 months prior to the study implemented more biosecurity measures than those who had not. However, as previously stated regarding vaccination, chemical use and quarantine by farmers, the combination of biosecurity measures, the varied level of technical skill applied by farmers and the timing and order in which they are used is important for effective prevention (Cuttance & Cuttance, 2014).
We discovered that when farmers were confronted with RVF outbreaks and were aware of the impact on their own farms, they were more likely to implement biosecurity measures, the most important of which is animal vaccination for RVF and that their motivation to engage in biosecurity measures improved, as reported by other studies (Merrill et al., 2019;Machalaba, 2020). and type of biosecurity measures evaluated, both findings suggest that biosecurity application is higher among those with better access to 'resources' . Our descriptive analysis also found that commercial and feedlot farms implemented more biosecurity measures, though this could not be evaluated in the final model as it was collinear with land ownership. These findings were consistent with a study that found that intensive production systems implemented increased levels of biosecurity in Cameroon (Kouam et al., 2020). The variation in the use of biosecurity measures could also be due to a number of factors that were not assessed in our study, such as the costs, time and labour required to implement biosecurity; farmer priorities, education and socioeconomic differences; as well as a lack of evidence for the efficacy or suitability of specific measures for different production strategies (Oladele et al., 2013;Niemi et al., 2016;Denis-Robichaud et al., 2019).
Further research exploring the reasons for the lack of, or variation in, implementation of certain biosecurity measures is imperative. In order to be persuaded of the need for biosecurity, farmers must have access to information on prevalent infectious diseases and their clinical course in animals, and the cost of having an outbreak, as well as on cost benefit of biosecurity measures especially in communal or pasture settings (Minjauw, 2000;Holleman, 2003;Mee et al., 2010;Damiaans et al., 2019). Rather than simply promoting biosecurity on its own, we recommend that this be supplemented by research to understand the prevalence of various diseases and discussions with farmers and veterinarians on how improved biosecurity can reduce exposures and improve productivity and market access. Resource prioritization is critical for farmers that face multiple economic threats. Individual farmers must choose which biosecurity measures they will implement based on the diseases with the greatest economic and/or health impacts, the feasibility of implementing the measures within their production system and the economic cost (Fast et al., 2015). It is important for the farmer to believe that the biosecurity measures are a good investment without perceiving the initial cost to be too high (Siekkinen et al., 2008).

CONCLUSION
Overall, this study provides important insights into current farm biosecurity practices, finding that most farmers do apply some biosecurity measures but also indicating widespread farm-to-farm variation in applied practices as well as some critical deficits. We found that land ownership, herd size and keeping two species of livestock, past outbreak experience, animal purchase all increased the likelihood of a farmer implementing more biosecurity measures. However, more research is required to gain a comprehensive understanding of what other characteristics are influential in determining or discouraging farmers from adopting biosecurity, and similar studies could be implemented in other areas. Increased biosecurity support would likely benefit all farmers, and the limited use of vaccination, tick control and isolation among communal or less commercialized farms implies that increased outreach and improved access to interventions are needed.