Farm staff wrong movements affects PRRSv prevalence and viremia

Background Biosecurity is known as the implementation of measures to reduce the risk of introduction (external biosecurity) and spread (internal biosecurity) of disease agents. One of the most common diseases in the porcine industry is the porcine reproductive and respiratory syndrome (PRRS), which has a huge negative impact on the well-being of the animals and consequently, on their productivity. Nonetheless, most of the biosecurity evaluation tools are based on scored systems. A new digital biosecurity system was designed to help control PRRS virus (PRRSv) infection status throughout an objective tool for the evaluation of internal biosecurity based on a system of control of the flow of internal movement of personnel in commercial farms. Movements, routes and health data were combined to classify the staff movements into three categories including “Risky” (From PCR(+) to PCR(-) barns), “Unsafe” (between PCR(+) barns) and “Safe” (From PCR(-)). Therefore, the main aims of the present work were to evaluate the efficacy of this new tool, its relationship with PRRSv incidence as well as to demonstrate the importance of biosecurity education to help farm workers to adopt safer daily practices. Results The observed results showed an overall smaller number of monthly movements (p < 0.05) and a significant increase in the Safe movements percentage (p < 0.05), concomitant with a decrease in the Risky movements percentage (p < 0.05) after the training session. In regards the relationship between staff movements and PRRSv presence, neither the percentage nor the total amount of both Safe and Unsafe movements were significantly different between the PCR(+) and PCR(-) groups of PRRSv status (p > 0.05). Nonetheless, both the total number and the percentage of Risky movements were significantly lower in the PCR(-) group (p < 0.05) compared with PCR(+) group. These results show a clear relationship between the total amount of Risky movements and the probability of a PRRSv outbreak in the farms.

Conclusions Our results support the notion that staff movement patterns within the different farm areas are a major factor in its internal biosecurity. The new tool described in the current work showed a significant relationship between staff movements and the probability of PRRSv outbreak and demonstrate the importance of biosecurity training to help farm workers adopt safer daily practices.

Background
Biosecurity is known as the implementation of measures to reduce the risk of introduction and spread of disease agents [1] and can thus be divided into two aspects. External biosecurity relates to the prevention of pathogens entering a herd, while internal biosecurity prevents the spread of disease within a herd, mainly from older to younger animals [1]. In this regard, biosecurity is an important aspect of preventing the transmission of diseases, thus improving health and reducing the need for antimicrobials [2]. Moreover, most diseases have a negative impact on the well-being of the animals and consequently, on their productivity. Thereby higher levels of biosecurity lead to also improved the economy of the farmer.
One of the most common diseases in the porcine industry is the porcine reproductive and respiratory syndrome (PRRS). This disease impairs swine health and is responsible for huge economic losses in the swine industry worldwide [3]. Infection with PRRS virus (PRRSv) is characterized by reproductive failures in pregnant sows, high pre-weaning mortality in piglets infected in utero and respiratory signs in both growers and finishers pigs [4][5][6].
Numerous studies [7][8][9][10][11][12][13][14][15][16] have described the routes which are involved in PRRSv transmission between and within herds, including the introduction of positive animals or semen, management of the quarantine for the newly introduced animals, as well as vehicles, aerosols, insects or contaminated fomites. Moreover, the marked genetic and antigenic heterogeneity of the virus, combined with its immune evasion strategies, inhibit the full efficacy of current commercial PRRS vaccines [17]. Therefore, PRRS control based only on the use of vaccination has often provided limited efficacy under field conditions [18]. Hence, it is of paramount importance the implementation of good biosecurity measures to prevent the introduction of the virus into a farm but also to slow down its transmission within a herd once infected.
Nonetheless, most of the developed programs of biosecurity measures are based on scoring systems or survey forms. For instance, researchers from Ghent University developed a scoring system called Biocheck.UGent™ [2,19] as a risk-based scoring tool to evaluate the biosecurity quality of pig herds. Another scoring system has been developed by the University of California-Davis (Disease Bioportal®) for the dynamic risk assessment and farms benchmarking also based on surveys [20]. In this line, Sternberg-Lewerin et al. [21] developed a risk assessment tool for Brachyspira hyodysenteriae and Mycoplasma hyopneumoniae considering the frequency of contacts, but it was only focused on external biosecurity. All these tools are based on values obtained through expert opinion panels; however, the perception of experts may vary depending on different circumstances, therefore, scoring systems based on perceptions should be adapted to each situation.
The aim of the current study was to evaluate the efficacy of a new digital biosecurity system, an objective tool for the evaluation of internal biosecurity based on a system of control of the flow of internal movement of personnel on pig farms as well as its relationship with PRRSv incidence.

Material And Methods
In order to evaluate the relationship between PRRSv viremia and risk level of farm staff movements in several commercial farm scenarios, eight Spanish farms with different PRRS disease status, facilities, ratio sows per person, size and biosecurity (internal and external) score were chosen. PRRS status was defined by barn (either Positive or Negative) in all farms. After this health evaluation, digital biosecurity system (Biorisk®, PigCHAMP Pro Europa, Segovia, Spain) system was installed, and workers started using the personal tracking devices and behaved as they usually did in order to measure the real workflow and risk level of movements on the farm. This period was considered as the control period.
After the first control period of one month, the collected data were used to develop a customized internal biosecurity training for farm staff based on their movements and their PRRSV situation, setting up the main goals to reach and steps to follow to control the disease.

Experimental farms
Main inclusion criteria: Each farm was included based on two main inclusion criteria: PRRS classification and biosecurity score. 1.
PRRS classification. It defines the initial PRRSv status according to shedding and exposure conditions [22] (Table 1).

2.
Biosecurity score. Every selected farm was initially evaluated based on both internal and external biosecurity using a standard questionnaire which gives a score from 0 (the lowest risk) to 100 (the highest risk) as Table 2 shows. Table 1 Breeding-herd classification for porcine reproductive and respiratory syndrome virus according to shedding and exposure status [22].  Table 2 Farm classification according to their biosecurity score (internal and external biosecurity) made by using a standard questionnaire.  Table 3 shows the characteristics of the eight selected farms including the type of farm management, the number of sows, ratio sows/person, biosecurity score and historical PRRS herd category.

Evaluation of PRRSv status
Before the implementation of this new system, some of the farms used their own diseases status evaluation system, while other farms even had not health evaluation methods.
Nonetheless, taking into account this scenario, the implementation of a digital biosecurity system is also linked to health assessments to determine the disease status, being PRRS evaluation one of the most relevant in this procedure. This process had two steps which included the initial evaluation and the disease monitoring. The number of samples depended on farm size and virus prevalence.
Initial evaluation. PCR (pools of 5 samples) and ELISA (individual samples) tests were run on each group of age on the farm to understand the infection dynamics. Defined groups of age were suckling piglets, weaned piglets, middle and ending nursery piglets, early and ending fattening pigs.
PRRSv infection monitoring. Every two months, a basic PCR profile (pools of 5 samples) was developed to ensure the PRRS status. Samples from suckling piglets, nursery and fattening pigs were taken. Every two-monitoring sampling, an initial profile was done to deepen the PRRS situation.

Farm barns classification
Based on laboratory results, every barn of the farm was defined at the beginning and during the process as: 1.
3. Sanitary area. Every changing clothes/boots or shower facilities.
This barn/building classification was used to define the risk level of every movement on the farm.
Movements and routes description.
Movements, routes and health data were combined to classify the staff movements into three categories, including "Risky", "Unsafe" and "Safe". Both "Risky" and "Unsafe" routes were those movements that could potentially spread the virus within the farm.
"Risky" movements corresponded to the most dangerous routes.
Movements going from a hazardous barn into a sensitive one.
Movements between two different kinds of hazardous areas. For instance, from infected fattening to the infected nursery.
"Unsafe" movements corresponded to those movements between two different hazardous areas barns inside the same kind of area -for instance, movements between two different fattening infected buildings.
"Safe" movements. All the movements between sensitive areas and from sensitive to hazardous areas.
Risky and Unsafe movements could be changed into Safe ones by taking a shower or changing clothes in the lockers. If the system detected the minimum time spent in locker rooms (defined farm by farm regarding the specific characteristics) in this route, the movement was classified as "Safe".

Hardware used
The new digital biosecurity system was based on two on-farm hardware pieces including beacons and readers. Each farm worker was given small Bluetooth™ transmitters called beacons, which were required to wear all the time while they were within the farm facilities. Readers were installed and fixed at every access of every barn, including lockers and showers. These devices can detect beacon signals by proximity. Whenever a beacon was within a device's detection range, the device registered the beacon identity as well as the detection time and uploads the record to a database.
Data collection and processing records from readers were sent to the cloud and processed, so the movements and routes of the farm's workers were computed. Each movement represented a route made by a farm worker from an origin zone to a destination zone. Thus, the system allowed the real-time monitoring of the farms' staff movements patterns.

Datasets
Data collected during this project was processed into two datasets: 1.
Dataset 1 contained the monthly amounts of Safe, Unsafe and Risky movements at each of the eight farms. As it has been commented, the first monthly record for each farm corresponded to the movement pattern before the biosecurity training. This dataset was used to compare the safety of the movements' pattern before and after the training across all farms.

2.
Dataset 2 contained the results of the bi-monthly PCR analytics at the three age groups for each farm, as well as the percentages of Risky and Safe movements observed on the farms during the two months preceding the analytics. The PCR records were classified into two groups: PCR(+), for those analytics with at least one age group PCR positive, and PCR(-), for those records with all three age groups PCR negative. This dataset was used to compare the percentage of Risky movements between the PCR(+) and PCR(-) groups.

Statistical and data analysis
Movements comparison before and after training

Results
The numerical results of the comparison of staff movements between farms are shown in Table 4. The Wilcoxon test determined a statistically significant decrease (W = 2; p = 0.025) in the total amount of movements after the training session in 7 out of the eight farms. After the examination of the total amounts of movements of each type, a significant decrease in total Safe (7 out of 8 farms, W = 3, p = 0.036) and Risky movements (7 out of 8 farms, W = 2, p = 0.025) was observed, while Unsafe movements tended to be lower after the training session (6 out of 8 farms, W = 6, p = 0.093).  However, the percentage of Risky movements was significantly smaller in the PCR negative group as determined by both the Kruskal-Wallis (K = 9.08, p < 0.005) and the ANOVA (F = 4.94, p < 0.05) tests. The group difference was more evident when comparing the total amount of Risky movements as determined by the Kruskal-Wallis total amount (K = 10.7, p < 0.005) as well as the ANOVA (F = 4.94, p < 0.05) tests. These results show a clear relationship between the total amount of Risky movements and the probability of a PRRSv outbreak on the farms.

Discussion
The present work tries to demonstrate the effectivity of a new digital biosecurity system based on objective data to control internal biosecurity in commercial pig farms. With this purpose, the system based on the control of the flow of internal movements of farm workers was implemented in different farms.
In contrast with tools previously developed, the system described in the present work is based only on objective parameters. In other words, the real movements that workers do in farms are recorded.
As it has been mentioned, all tools previously developed regarding biosecurity (both external and internal) are based on values obtained through expert opinion panels (farmers themselves and veterinarians) [2,[19][20][21]. For instance, Biocheck.UGent™ test, originally developed by Laanen et al. [2,23], consists of a total of 109 questions grouped in different subcategories for external and internal biosecurity. Subcategories related to internal biosecurity are: "Disease management"; "The farrowing and suckling period"; "The nursery unit"; "The fattening unit"; "Biosecurity measures between compartments and the use of equipment"; and "Cleaning and disinfection", and questions have to be completed by the experts.
Nevertheless, it is important to keep in mind that the perception is a subjective aspect, and therefore, expert opinion will always introduce some bias that might be influenced by different factors, such as knowledge, previous experiences or personality. For this reason, scoring systems used as a tool to control biosecurity are not the best option.
Allepuz et al. [24] recently pointed the desire for the development of more complex models to providing quantitative risk assessment. Emphasizing that this kind of model could be more precise in mimicking the reality and might provide a more accurate estimation of the probability of virus spread within the herds by the different routes. The same authors noted that these quantitative models could not be developed easily because of the lack of relevant long-term data series. Moreover, according to Sternberg-Lewering et al. [21] is extremely difficult to obtain quantitative data from field studies.
Nonetheless, as it has been presented, the development of this new tool allows the collection of a huge data series and their real-time processing. At the time in which the present study was carried out, other 19 companies have installed the system and are working with it worldwide.
It is already known that compliance with biosecurity measures varies depending on personality traits, such as responsibility, work experience and education [25].
Furthermore, in this sense, veterinarians play a key role in training and educating farmers with the final purpose that they carry out the application of the correct biosecurity measures on their farms [26]. Hence, the system developed in the current research allow both veterinarian and consultants acting at three different levels: 1) Follow-up of the health plans recommended and its degree of compliance 2) Customized training to the workers that need more support, coaching or time dedicated 3) Provide easy remote monitoring and consultancy.
As McCaw [27] already described, adequate management practices help to control mortality rates in both suckling and nursery pigs during acute outbreaks of PRRS. In this sense, McREBEL (management changes to reduce exposure to bacteria to eliminate losses) management has demonstrated an improvement in preweaning and nursery mortality related to PRRS. The McREBEL management must achieve the following four steps: 1) Cross-fostering of piglets between litters will not occur after 24 hours of age. 2) Suckling piglets and nursery pigs will be moved strictly all-in-all out by room. 3) Piglets will not be moved among different rooms to "nurse sow" (especially poor-doing piglets to younger age groups attempting to save them). 4) Piglets without a prognosis for recovery will be euthanized to minimize the exposure of other pigs in the litter or room to secondary bacteria and PRRS.
In that regard, and has already demonstrated McREBEL management, thanks to the installation of the new tool to control personal flow inside a farm as is describing in the current work, it has been proved that little variations of farm management, can reduce the spread of diseases and improve the global sanitary status of farms.
One of the most important conclusions obtained in the current work was that after training, farm workers were more conscious of the risks associated with movements and they usually avoid unnecessary routes within the farm. This observation is directly associated with a lower probability of spread any disease, and consequently an increase of productivity of farms.

Conclusions
According to the results, it can be concluded that the present work supports the notion that staff movements patterns within the different farm areas are a major factor of its internal biosecurity. This new digital biosecurity system confirmed a significant relationship between staff movements and the probability of PRRSv outbreak incidence.
Moreover, the movements observed before and after training demonstrated the importance of biosecurity knowledge and training to help farm workers adopt safer daily practices. This system opens new possibilities to make objective those factors that were impossible to detect or quantify so far. Furthermore, it allows focusing the efforts on those areas or people that show more important deviation from the recommendations. This new digital tool allows for collected data which combined with classic tools enable to generate information helping decision making.

Declarations
Abbreviations PRRS: porcine reproductive and respiratory syndrome.
PRRSv: porcine reproductive and respiratory syndrome virus.

Acknowledgements
The farmers and their stockmen are greatly acknowledged for agreeing to engage in the trial and for working in good cooperation with the investigation team.

Funding
This study was co-financed by the Institute for Business Competitiveness, which depends on the Regional Government of Castilla y León (Spain) and the European Regional Development Fund (ERDF) aligned with Thematic Objective 1: "Strengthening research, technological development and innovation".

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on request.  Average of percentages of Safe (green), Unsafe (orange) and Risky (red) movements before (■) and after (□) the training session considering the eight farms in which the system control movement was installed.

Figure 2
Comparison of percentages and totals of Safe, Unsafe and Risky movements between the PCR analytics positive and negative groups. The boxes extend between the first and third quartiles of the data for each group, and the whiskers extend between the minimum and maximum values. Group average is shown by the red triangles, while median lines are shown with solid green lines.