Science of biosecurity: understanding the mechanisms of pathogen entry into pig populations, and how to implement effective biosecurity improvement programs

The global swine production industry has experienced a revolution in the last 30 to 40 years. Forty years ago, production was geographically centralized (i.e., all production activities were centrally located), small scale and outdoors. Producers were knowledgeable in many areas, but a specialist in none and few had employees. Swine production has now transitioned to where many of the production activities are geographically decentralized at specialized facilities. The stages of production are geographically segregated where the breed-to-wean, nursery (or rearing) and finishing (or fattening) stages of production occur at separate locations. Other examples include boar studs where semen is collected, processed and delivered to sow farms and large-scale feed mills that mill and deliver feed to many farms. Small scale, outdoor production has been replaced by large-scale systems where pigs are raised entirely indoors. There are now over 31 production companies globally that have at least 100,000 sows (Hess, 2019). The size of individual farms has also grown. In the United States, for example, 27 percent of swine farms were considered large (>5,000 head) in 1994 (USDA, 2015). By 2014, just 20 years later, over 93 percent were considered large. As the size of farms and production companies increased, they became more specialized and more reliant on employed labor. On many swine farms, the owner(s) no longer regularly works on the farm. The larger production companies now hire very specialized labor, including veterinarians, nutritionists, reproductive physiologists, and animal welfare specialists. The larger production companies have opportunities to share equipment, such as livestock trailers and feed trucks, to use these resources more efficiently (i.e., not sitting idle) and capture significant economies of scale. What have these changes meant for biosecurity? All have resulted in major increases in the size of losses when pathogens are introduced into a herd and the frequency of opportunities to introduce those pathogens. It is not uncommon in herds with at least 4,000 sows to have losses that exceed US$1,000,000 after an outbreak of PRRS. The various forms of specialization and geographic segregation of production and ancillary production activities (e.g., feed mills, boar studs and gilt multiplication farms) have led to a major increase in the frequency of events where pathogens can be introduced into a herd. A list of examples, that is by no means all-inclusive, includes feed, semen and replacement gilts that are routinely delivered to sow farms, and with them the drivers, vehicles and trailers that deliver them. 

Swine diseases and disease management have also evolved in the last 30 to 40 years. The emergence of PRRSV in the 1980s was arguably the most significant swine disease event to occur in most of our lifetimes. The recent spread of African swine fever virus (ASFV) may alter our collective opinions about that in the next decade. The adoption of all-in-all-out production of growing pigs and the emphasis on eliminating swine pathogens, such as porcine reproductive and respiratory syndrome virus (PRRSV), porcine epidemic diarrhea virus (PEDV) and Mycoplasma hyopneumoniae (Mhp) from sow herds have also increased the value of biosecurity. When groups of growing pigs are negative for certain pathogens and lacking immunity at placement, the value of keeping those pathogens out is immense. When investments are made to eliminate pathogens from sow farms, the value of preventing the reintroduction of the pathogens is enormous. In the case of PRRSV, where heterologous immunity is often sub-par, the value of not introducing new isolates into sow herds is large, even if the herd is already infected with other isolates. 

For countries that are presently free of specific pathogens, such as ASFV, foot and mouth disease virus (FMDV) and pseudorabies virus (PRV) in the United States, the need to prevent these transboundary pathogens from entering and spreading rapidly has also increased in importance as the world has become more connected and global trade more important. The recent spread of ASFV in Southeast Asia and Eastern Europe has put the world pork industry on high-alert. 

While the need for biosecurity on swine farms has increased, the industry has been slow to respond to the new calculus. Why has progress been slow? For the last 15 years, I have focused much of my research and outreach on biosecurity, biosecurity risk assessments and outbreak investigations. During that time, I have observed swine producers and veterinarians use a "ready-fire-aim" approach to biosecurity. I have been guilty of it myself. However, is "ready-fire-aim" the right order? The answer is not readily apparent. Entrepreneurs are frequently celebrated and rewarded for approaching things in this order, especially with disruptive technologies where being first-to-market is everything. However, is it the right order to approach biosecurity? I believe that it has led to misallocation of resources and very slow progress on improving biosecurity and reducing the frequency of outbreaks of PRRS and other diseases.

The ready-fire-aim approach has been guided by experimental research. When a new study was published demonstrating that, for example, livestock trailers, boots, insects or aerosols are capable of carrying pathogens from herd-to-herd, the industry has opened fire with only a rudimentary understanding of how much livestock trailers, or boots, or insects or aerosol were contributing to the herd-to-herd transmission of pathogens. Just because they can, does not mean that they frequently do. Were resources being spent on these when they would have been better spent addressing other things that can carry pathogens from herd-to-herd? 

The better order is "ready-aim-fire." What does the aim step in ready-aim-fire look like? Very simply, it is trying to identify the most frequent ways by which pathogens are being introduced into herds and then prioritizing improvements in biosecurity to address them. It is taking time to identify the most significant vulnerabilities on farms to determine what should be done next. Experimental research studies are no help for identifying the most frequent ways that pathogens are introduced. To study this question, we have to observe what is happening. However, ad-hoc approaches to biosecurity risk assessments and outbreak investigations are not sufficient. Learning faster from observing requires a more systematic, comprehensive and consistent approach. 

My experience with the Production Animal Disease Risk Assessment Program (PADRAP) (Holtkamp et. al., 2012) in the early 2000s left me frustrated. The survey was useful for identifying specific risk factors that were present or absent and specific biosecurity practices that were done or not done. That may have provided some sense of what to fire at; however, it offered little in the way of prioritizing (i.e., aiming). In PADRAP, a score was assigned to each response to the questions which conveyed a sense that some risk factors or practices were more important than others. The aggregated risk scores in PADRAP did prove to have some value for predicting herds where introduction of PRRSV, causing outbreaks, occurred more frequently. But the risk scoring system had several shortcomings. First, the score assigned to a specific response was independent of the response to other questions in the survey. For example, the score for not washing livestock trailers between every load of weaned pigs was the same whether the trailer was dedicated to the farm or used to haul pigs from 20 other sow farms. The score for any given response was also independent of a host of other information that was not addressed in the survey. For example, if one of the trailers that hauled weaned pigs was driven by someone who also worked at a market where culls sows were commingled. There is no survey that does, or ever will, account for all of the important information for every farm under every circumstance. That is also why large observational studies where surveys are used to collect data on biosecurity practices and risk factors give conflicting answers about the most significant practices and risk factors on farms (Holtkamp et.al., 2010; Holtkamp et. al., 2012; Bottoms et. al., 2012). PADRAP was also poorly organized according to a haphazard combination of pathogen carrying agents, such as trucks and people, and events, such as entry of semen and entry of employees, which occur on swine farms. PADRAP was not very helpful for identifying the most frequent ways by which pathogens were being introduced into herds. It did not help prioritize where to spend time and other resources next.

Experience with PADRAP also taught me that producers and veterinarians were most interested in assessing biosecurity and taking action to improve it in the aftermath of an outbreak. It is human nature to be most responsive in the face of a crisis or threat. It makes a great talking point to say that we should not wait for an outbreak to improve biosecurity, but it rarely works that way. With funding from the Iowa Pork Producers Association, my students and I developed an approach to conduct epidemiological investigations of outbreaks in a systematic, comprehensive and consistent manner. The resulting approach, and forms used to guide the approach, has also proven valuable for conducting risk assessments prospectively. 

The approach developed was based on the basics of epidemiology and risk assessment and considered how swine pathogens are transmitted from one herd to another. Since bacterial and viral pathogens are not capable of locomotion, something must carry the pathogens from one herd to another. The term “pathogen carrying agent” was applied to any object, animal, or person that may carry a virus into the breeding herd. A pathogen carrying agent may directly transmit the pathogen by being infected or indirectly by being contaminated with the pathogen. For epidemiological investigations of outbreaks and prospective risk assessments, the objective is to identify vulnerabilities on the farm being investigated or assessed. Therefore, the investigation or assessment is made from the perspective of the farm being investigated or assessed. Imagine standing in the middle of the farm, observing the potential pathogen carrying agents entering the farm. The entry of pathogen carrying agents, delineated by when they cross the outer perimeter of the farm, can be characterized as events. A "carrying agent entry event” is defined as an event where one or more potential pathogen carrying agents crosses the outer perimeter, physical or abstract, of the swine farm. Carrying agent entry events that typically occur on sow farms include delivery of semen, entry of breeding replacements, removal of cull animals, removal of weaned pigs, disposal of carcasses, delivery of feed, collection of garbage, delivery of supplies, entry of on-farm employees, entry of repair/service personnel working inside barns, entry of repair/service personnel working outside barns, entry of other visitors, entry of food, removal of manure, entry of other animals (non swine domestic and wild), entry of insects, and entry of air and water. Some carrying agent entry events, such as entry of air and water, occur continuously. For discrete events that are easily observed and recorded, such as entry of semen and removal of manure, a date(s) or frequency over some period of time can be determined. For others that are discrete but not always observed, such as entry of rodents and other wild animals, a subjective assessment of how frequently or intensively they occur can be made.

Organizing epidemiological investigations and risk assessments by carrying agent entry events has proven to work very well. PADRAP and many of the biosecurity risk assessment surveys available today are confusingly organized by combinations of pathogen carrying agents and carrying agent entry events. It is essential to be consistent. The alternative is to organize them by pathogen carrying agents consistently. However, there are too many potential pathogen carrying agents that enter farms; over one-hundred on sow farms depending on the level of resolution used to define them. Organizing epidemiological investigations and risk assessments by pathogen carrying agents becomes overwhelming. 

Concepts of Hazard Analysis Critical Control Points (HACCP) were also incorporated. HACCP is a system that was initially developed in the 1950s by a team of scientists in the United States to ensure food safety for the manned space program. In HACCP, critical control points (CCP) are steps in a process where failures lead to adverse outcomes, such as the introduction of a pathogen into a herd. In the context of outbreak investigations and risk assessments, identifying the steps in a process, such as entry of semen, and the opportunities for failure is necessary to identify the vulnerabilities in biosecurity on the farm being investigated or assessed. A CCP is also defined as a step where a biosecurity practice (i.e., control) can be applied to prevent a failure or reduce the probability of failure. In the swine industry, the logistics of production are highly variable. For example, if asked to describe the steps involved in collecting, processing, delivering, entering semen into a farm and using it to inseminate animals, most veterinarians would have to think about how it is done on a farm single with which they are familiar. If asked to describe the steps for another farm, even if it is in the same company, they would likely be different. Therefore, to conduct effective epidemiological investigations and risk assessments, where the purpose is to identify vulnerabilities for a farm, the steps involved with each carrying agent entry event, and opportunities for failure must be understood and documented for that farm. That is the basic principle on which the approach we developed to conduct epidemiological investigations of outbreaks and risk assessments was built. 

One thing that is consistent for every farm is that the same series of failures is required for a pathogen to be introduced into a herd by a pathogen carrying agent. A biosecurity failure at a single step is necessary but not sufficient. Instead, a series of 3 failures is required for a pathogen to be introduced into a herd by a pathogen carrying agent (Fig. 1). First, the pathogen carrying agent must be contaminated by, or infected with, the pathogen before entering the breeding herd. Second, the biosecurity measures in place to mitigate the infection or contamination of that carrying agent must fail. Third, the infected or contaminated carrying agent enters the farm and the pathogen is transferred from the carrying agent to a pig within the herd. 

Figure 1. The series of failures required for a carrying agent to carry a pathogen into a herd.

The epidemiological and HACCP principles were used to develop a standardized investigation form. The investigation form for sow farms contains a section for each of the carrying agent entry events that typically occur on sow farms. There is also a section for characteristics of the site, characteristics of the area surrounding the site (up to 7.5 km) and characteristics of the herd. 

The information collected for epidemiological investigations of outbreaks and risk assessments includes the frequency (and dates for outbreak investigations) with which each carrying agent entry events occurs. Information on all of the CCPs or steps where the failures may occur resulting in the introduction of a pathogen is gathered through an open-ended discussion. Closed-ended questions are included in the form to gather relatively standard information about the steps involved with each carrying agent entry event and the likelihood of failures occurring at each step. The closed-ended questions also help guide the open-ended discussion. For example, closed-ended questions in the section on entry of semen about testing of boars in the boar stud serve as a reminder to ask about all the relevant details related to how the testing is done and whether the results of the testing are always known and communicated before the semen is used. The closed-ended questions will never capture all of the relevant information, so it is important to get a good narrative of the steps and the likelihood of failures occurring in order to assess the biosecurity vulnerabilities fully. Responses to the closed-ended questions can also be aggregated for analysis of data from multiple investigations and assessments.

As an example of how epidemiological investigations of outbreaks conducted in a systematic, comprehensive and consistent manner, can help to identify carrying agent entry events that are most frequently making farms vulnerable to the introduction of pathogens, nineteen cases of PRRS outbreaks in swine breeding herds were investigated. The herds were located in the Midwest US and all of the investigations were completed between January of 2015 and September of 2017. To assess how the virus may have been introduced into each herd, the investigations were conducted, using the form developed, for a 4-week investigation period ending on the date clinical signs were first recognized. Diagnostic results confirmed that each case occurred as the result of a new isolate of PRRSV in all of the cases, as determined by the herd veterinarian. 

For every case, the investigation interview was conducted by a facilitator and an assistant facilitator who took notes and captured all relevant information in the investigation and report form. The same facilitator (Holtkamp) conducted all 19 of the investigations, with three different assistant facilitators. In addition to the facilitators, the herd veterinarian and the farm manager were present at the investigation interview for all of the cases. 

The investigation and report form served as a guide for an open-ended discussion of each case and all of the carrying agent entry events that occurred in the four weeks preceding the first clinical signs. The facilitators subjectively assigned each carrying agent entry event a rating low, medium, or high for the likelihood of PRRSV introduction into the herd. The assessment was based on how frequently and when each carrying agent entry event occurred during the investigation period, the steps associated with each event where failures may occur (i.e., CCPs) and how likely the series of failures required for a pathogen carrying agent to introduce a pathogen into a herd occurred (Figure 1). All pathogen carrying agents associated with each carrying agent entry event were assessed. 

For the 19 cases investigated, the number of times each risk event was rated medium or high for the likelihood it was responsible for the introduction of PRRSV is summarized in Figure 2. Entry of on-farm employees was the entry event rated high most frequently; rated high 9 times and medium 5, out of the 19 cases. Removal of cull sows and entry of repair personnel working inside the barns were rated high next most frequently. 

For the cases where on-farm employee entry was rated high, significant vulnerabilities were identified. On-farm employee entry was the most frequent risk event on all of the farms investigated, creating many opportunities for failures to occur. The average number of discrete carrying agent entry events that occurred during the 28 day investigation period was 369 events per farm, of which 170 (46%) were on-farm employee entry events. A single employee entering the barns was considered one risk event. All of the farms had a shower, but 4 lacked a well-defined clean-dirty line and only 7 had a bench in addition to the shower. Biosecurity practices that were recommended to address the biosecurity vulnerabilities in these cases included the addition of well-defined clean dirty lines and installing benches. The bench adds a second clean-dirty line and serves as an additional layer of biosecurity. It is impossible to fully know or control where employees go while away from the farm, but attempts at some control were observed. In 7 of the 19 cases, a rule that employees were not allowed to visit or work on other swine premises was in place at the time of the outbreak. In 9 of the 19 cases, at least one on-farm employee did visit another swine farm during the investigation period. Only 8 of the farms required downtime for employees after visiting or working on other swine farms and none required employees to wash and disinfect their vehicles. Rules prohibiting employees from living with people that worked on other swine farms were in place at 10 of the farms at the time of the outbreak, but 6 of the farms allowed employees to work in other swine related activities, such as hauling pigs, loading market pigs, delivering gilts, managing a feed mill or delivering feed. Other frequent observations were towels on the dirty side of the shower, shoes worn into the shower area and dirty entryways and showers. High employee turnover, disgruntled employees and managerial transitions were also factors in 4 of the cases. One case, in particular, exemplified how employee issues can put farms at risk. At the time of the outbreak, the farm manager was transitioning to a new job; several employees had been observed intentionally violating biosecurity rules; two were suspected of having left in the middle of the workday, without permission, to interview on another swine farm and employee turnover exceeded 100% in the year before the outbreak. Because it is difficult to know or control where employees go while away from the farm, attempts to reduce the probability that employees or their vehicles become contaminated (i.e., the first failure) are often less productive. Most producers and veterinarians feel they have more control over practices to mitigate the contamination (i.e., the second failure) by adding more layers of biosecurity, such as benches, hand washing stations, and showers, and focusing training efforts on improving compliance with those practices.

Figure 2. Number of cases, out of 19, that each carrying agent entry event was rated medium or high for the likelihood it was responsible for the introduction of PRRSV.

The dramatic transformation of the pork industry in the last 30 to 40 years has changed the need for, and value of, biosecurity on swine farms, but progress in slowing the transmission of pathogens from herd-to-herd has been slow? In the United States, this was very evident in 2013 and 2014 when PEDV was first introduced into the country. Within 18 months of the initial case, nearly half of the sow farms in the country became infected (University of Minnesota, 2020). The ad-hoc approaches to biosecurity risk assessments and outbreak investigations have not been sufficient. While it is true that mistakes are opportunities to learn, the learning part is not guaranteed. Learning faster from our mistakes requires a more systematic, comprehensive and consistent approach like the one described in this proceedings paper. The approach and form are being used as part of a Rapid Response Program funded by the Swine Health Information Center (SHIC) in the United States. The Rapid Response Program is a nationwide network of veterinarians, state animal health officials or representatives and epidemiologists who are trained, prepared and committed to act within 24 hours of contact to conduct epidemiological investigations when a new transboundary or emerging disease threat occurs with a known etiology. Resources developed for the Rapid Response Program, including the form for conducting an epidemiological investigation of outbreaks, are available at the Swine Health Information Centers website (Swine Health Information Center, 2020). 

Development of the epidemiological investigations of outbreaks and risk assessments forms partially funded by the Swine Health Information Center (Project #19-149 SHIC) and the Iowa Pork Producers Association (Project #14-283).