Identification of a One Health Intervention for Brucellosis in Jordan Using System Dynamics Modelling

Abstract

Brucellosis occurs because of a complex multisector system involving the interplay between animal production and human behavior. A better understanding of transmission dynamics helps pinpoint the most effective interventions to reduce human and animal cases. Modeling methodologies have not been applied extensively to brucellosis. This study applies system dynamics modeling to identify the interplay between the different sectors that drive disease transmission and identify One Health interventions to control brucellosis. The study applies a quantitative system dynamics modeling methodology based on a qualitative in-depth analysis of the Brucella transmission system in Jordan. The data are analyzed manually to establish causal pathways to develop a stock and flow (SF) model. The structure is examined and reviewed by key informants. Several intervention scenarios to control Brucella transmission is assessed. Brucella transmission among sheep and between farms and markets are the main drivers of human incidence. Farmers’ visits to veterinary clinics are a critical intervention point for control associated with access to vaccination. The vaccination of sheep and health education activities alone might not be an effective strategy. Testing and culling sheep is the most efficient individual control strategy for animal incidence and a combination of public and animal health strategies (One Health) is the most effective strategy to reduce human incidence. The synthesis of current knowledge through the system dynamics model enables better understanding, visualization, and interpretation of the sectors involved in Brucella transmission. There is a strong synergy between sectors; therefore, greater control might be provided by utilizing multi-sectoral relationships embedded in the system.

1. Introduction

Brucellosis is the most common zoonotic disease in the world that is caused by infection with multiple species of Brucella [1]. Brucella infection in humans is associated with direct or indirect exposure to infected animals or through the consumption of uncooked meat or dairy products [2]. Brucellosis is particularly prevalent in the Mediterranean and the Middle East region [3]. The epidemiology of brucellosis is associated with the infecting species/livestock host relationship with occupational exposure to animals the most important risk factor [4]. Brucellosis is also a cause of significant economic losses related to reduced fertility in livestock [5].

Jordan has one of the highest annual incidence rates for brucellosis in the world (300 cases per million/year) and it shares a border with Syria that reports the highest incidence in the world (1603 cases per million/year) [3]. Brucellosis remains a significant issue for the livestock sector in the country, which accounts for 55% of the agricultural output [6]. In Jordan, people traditionally keep sheep to produce meat and milk. Vaccination is the only control measure systematically employed in Jordan [7]. In Jordan, most of the livestock market trade, particularly local trade, is not regulated by the government and does not follow any animal or public health guidelines to prevent disease transmission. Visiting the governmental veterinary clinics is the main source of information and education for farmers about brucellosis and other disease control practices.

Brucella transmission occurs as part of a complex interacting system that includes livestock, food production systems, human and animal health systems, and the cultural and social factors that drive human behavior [8,9]. Understanding Brucella transmission at a multidisciplinary scale requires a holistic research approach to better understand the nature of the complex interactions embedded in the system. This requires stakeholders’ participation and involvement to enrich the model building and to understand the underlying problems in the system that facilitates the consensus on interventions and scenario planning and implementations. Previous work has identified the qualitative feedback dynamics and the emerged qualitative themes that interplay in brucellosis transmission system in Jordan [10]. This study expands to present and develop the quantitative dynamic model of the identified brucellosis transmission system in Jordan.

System dynamics modeling (SDM) enables a better understanding of complex health conditions [11]. The use of approaches that include stakeholders’ participation enriches the model-building process [12] and enables a better understanding of a problem at different disciplinary scales (i.e., the population, behavioral, and health services) [13]. Systems dynamics modeling is used in this study to determine the interdisciplinary complexity of Brucella transmission [11] and explore the intervention scenarios to achieve effective control [14]. This is achieved by expanding beyond the traditional susceptible–exposed–infected–recovered (SEIR) model structure to include the impact of feedback loops and embedded relationships found at other different multidisciplinary scales.

2. Materials and Methods

The model was developed in five steps: (1) problem articulation to define the dynamic problem and the timeframes related to it; (2) forming a dynamic hypothesis and mapping out the structures; (3) building the simulation model structure and estimating the parameters; (4) testing the model was tested; and (5) designing and evaluating the interventions [13,14].

2.1. Problem Articulation

The main qualitative data sources included literature and document reviews, governmental reports, and interviews with stakeholders. Key stakeholders were identified through the systematic approach proposed by Elias et al. (2001) [15] and the WHO [16]. This involved listing all possible eligible stakeholders prioritized by a local governmental expert. Snowball sampling was used to recruit stakeholders not listed previously. An interview guide was developed, and fourteen interviews were conducted. The obtained data were used to build, develop, and inform the stock and flow model (Appendix A).

Qualitative data were analyzed using the progressive approach of Halcomb and Davidson [17] that involved iterative and repetitive reviews of the fourteen audio-recorded interviews to identify all themes expressed by the respondents. Inconsistencies identified in the recordings were resolved by seeking further elaboration from the respective respondents [18]. Data saturation was confirmed by comparing the provided information with the information collected from previous interviews.

2.2. Dynamic Hypothesis

A dynamic hypothesis was formulated by identifying the main sectors that showed non-linear causal relationships that drove the dynamics of brucellosis transmission [10]. The dynamic relationships embedded in the system were categorized based on their causal contribution to brucellosis transmission in the system. Causal-loop diagrams were used to formulate the dynamic hypothesis and the causal non-linear relationships between the identified sectors [10]. The dynamic relationships produced were used to inform the systems dynamics model structure.

2.3. Model Building

The model was built in Stella Architect (Version 1.9.1 iSee Systems) [19]. The software presented model variables and relationships mathematically as integral and algebraic equations and offered a user-friendly graphical interface [14,19]. The dynamic hypothesis was used to inform the model structure. The parameters were estimated based on the primary data collected from stakeholder interviews and secondary data from publicly available government data and published papers on brucellosis in Jordan and worldwide. Whenever possible, parameters that represented a local measure in Jordan were used; otherwise, parameters were based on globally acceptable estimates identified from the literature (Appendix B).

2.4. Model Testing

Model boundaries were checked [14] to ensure all relevant and endogenous variables in the scope of the complex problem were included. The model was tested for dimensional consistency, mass balance, structural assessment, integration error, and extreme error tests according to the method of Sterman and Maani and Cavana [13,14]. Additionally, the overall fit of the simulated behavior to the reference behavior (annual incidence of brucellosis in Jordan) was checked (Appendix C).

A sensitivity analysis was conducted according to the method of Sterman and Maani and Cavana [13]) to identify the variables most likely to influence the model behavior by examining the greatest change produced in the system (main outcome) by a minimal change in the system’s parameters. The parameters used in the sensitivity analysis included variables, such as “buy and sell restriction”, “vaccination rate”, and “illegal livestock entry restrictions”. The outcome of interest in the study was the number of confirmed cases of human brucellosis per year. Secondary outcomes were the prevalence of brucellosis in sheep and the number of disease-caused abortions. The sensitivity analysis was conducted by introducing a percent increase or decrease to all the sensitivity parameters, and then examining the effect on the main outcome, cumulative over time.

2.5. Intervention Designs and Evaluation

The scenario design and analysis (interventions) were conducted to simulate interventions to reduce the incidence of human and/or animal brucellosis based on the results of the sensitivity analysis and the feedback provided by the key stakeholder reviewers’ expert opinions. The following intervention scenarios were designed.

First Scenario: Farmers visits intervention

This intervention aims to encourage farmers to visit veterinary clinics, the primary contact point in Jordan, to increase the frequency of visits per year. The initial visit rate was estimated based on previous studies [20]. A moderate increase (+ 25%) in the visit rate was included based on local expert opinion.

Second Scenario: Market trade restrictions intervention

In Jordan, most of the livestock market trade, particularly local trade, is not regulated by the government and does not follow any animal or public health guidelines to prevent disease transmission. This intervention aims to introduce restrictions on sheep trade in local markets to reduce the introduction of infectious sheep into farms. The base-case scenario value for this parameter is zero because there are currently no restrictions on the local trade of sheep in Jordan. A modest increase in the overall market trade restriction of 15% was chosen based on local expert opinion and acknowledging the predicted resistance from local governments that operate sheep markets.

Third Scenario: Awareness of farmers intervention

In Jordan, visiting the governmental veterinary clinics is the main source of information and education for farmers about brucellosis and other disease control practices. This intervention involves activities to raise awareness of brucellosis to reduce contact (improve hygiene) and to reduce the consumption of unpasteurized dairy products, particularly by farmers in rural areas. The current awareness is estimated based on local studies with the scenario assuming a moderate increase in the proportion of farmers’ awareness (+30%).

Fourth Scenario: Test and cull infected sheep intervention

This intervention is designed to eliminate sources of infection through a test and cull policy. The base-case scenario is set to zero because there is currently no test and cull, of infected animals, policy in Jordan. A 20% increase in test and cull was chosen based on the likelihood of significant barriers to implementation due to challenges in the logistics and the need to provide financial compensation for farmers.

Fifth Scenario: Slaughter aborted/suspected sheep

This intervention aims to increase the removal of sheep with clinical signs of brucellosis (abortions). This practice is an existing ad hoc approach used by farmers in response to reproductive loss. The proposed intervention increases the removal of sheep with clinical signs of brucellosis (aborted sheep) by 25%, based on a communication program by the Ministry of Agriculture (MoA).

Sixth Scenario: Enhanced protective immunity

This intervention is based on the current government policy to introduce a new brucellosis vaccine that provides protective immunity for 36 instead of 9 months. The base-case immunity duration was extended from 9 to 36 months for this scenario.

Seventh Scenario: Combined vaccination and immunity

This intervention is designed to combine extended protective immunity with an increase in vaccination rate. A conservative increase of an absolute 17% in the current 1% vaccination rate was observed.

Eighth Scenario: One Health

This “One Health” intervention combines interventions across multiple sectors, including a 30% increase in farmer’s awareness (Scenario 3), a 25% increase in the slaughter rate for aborted/suspected sheep (Scenario 5), a longer duration of vaccine immunity (36 months) (Scenario 6), and a 15% restriction for market trading (Scenario 2).

2.6. Evaluation of Interventions

The SD model was run first using existing parameters to establish the base-case or “business as usual” outputs. Then, each scenario was simulated, and the “effectiveness” was evaluated by comparing the change in four output measures: the cumulative number of confirmed cases of human brucellosis, the intensity (amplitude) of human brucellosis, brucellosis prevalence in sheep, and cumulative sheep abortions to the base-case outputs. The cumulative number of confirmed cases was the total number (sum) of human brucellosis cases over the time of the model. The intensity (amplitude) of the human brucellosis outbreak was the number of cases that were reported each year, which indicated the height of the epidemic curve. The Brucellosis sheep prevalence was the total number of reported Brucellosis cases in sheep for a given year. Each intervention was then ranked by the overall reduction of brucellosis in sheep and humans calculated as the sum of the four output measures.

3. Results

A modified age-sex structured compartmental model of the susceptible–infected–recovered (SIR) [21] was developed to simulate Brucella transmission in sheep with a spillover infection of humans because of the direct and indirect exposure to infected animals and animal products. The model was structured using a monthly time-step over 12 years beginning in January 2004 and extending to January 2016 (a total of 144 simulation months) to match the time for which the data were obtained. Figure 1 presents a high-level model schematic of the whole developed model sections.

The reference mode, Figure 2, is the annual number of human cases of brucellosis and the corresponding incidence rate in Jordan from 2004 to 2016 [22]. The preferred behavior for this system is to decrease the annual incidence of human brucellosis in Jordan. The pattern of human brucellosis observed in the data shows a cyclical pattern with two peaks 3–4 years apart with a recent substantial increase in incidence from 2012.

The model incorporated Brucella transmission among sheep via venereal and oral (milk-feeding) pathways, which were represented through separate sub-system models for ewes (female sheep), rams (male sheep), and pregnant ewes. The human infection model was developed to allow the transmission of Brucella from sheep by including a dairy production section (infected ewes). Public (medical and veterinary) health sectors and the impact of local sheep markets and the transboundary movement of sheep were also incorporated. Factors, such as the economic and environmental (weather and climate) aspects, or other transmission pathways were considered exogenous to the model.

3.1. Model Sub-Systems

The model structure was informed by the diagrams developed as the dynamic hypothesis that included the most influential eight non-linear causal relationships in each sub-system driving the transmission system [10]. Each causal relationship represented a sub-system in the model. The model comprised eight interconnected sub-systems, including human, ewe and ram, sheep mating and pregnancy, venereal transmission probability, food safety, public health reporting, and animal health. The complete structure for each sub-system model is presented in Appendix A.

3.2. Model Testing

Model testing showed that boundaries were fit for study purposes (boundary adequacy test); a boundary adequacy test was conducted by regularly matching and referring to the developed CLDs [10,19]. The developed CLD was the map that provided the desired scope and borders reference and the main tool to test the boundary adequacy of the model. Additionally, the boundary adequacy was established by allowing stakeholders (semi-structured interviews) to propose the unidentified inputs to incorporate in the final model structure. This participatory model building process ensured an adequate model boundary for the scope and purpose of the model.

The model was checked so that the parameters were adequate, realistic, and meaningful (i.e., dimensional consistency test [14]). Dimensional consistency was tested by the dimensional analysis tool incorporated in the Stella Architecture program [19]. This tool checked the logic and consistency of all the included equations and units and found no inconsistencies or faulty units. Furthermore, the dimensional consistency was robust because the dummy variables were not incorporated, and each equation was individually analyzed.

The model did not construct or destruct flows (i.e., what entered into the model was either saved as a stock or exit as an output) unintentionally (mass-balance test) [14]. Mass balance means that all mass values can be accounted for, that is, the mass flows into the system must either stay in the system (by being stored in stocks) or it must flow out. This was tested by calculating the difference between the inflows and outflows for each section of the model to confirm that it was zero. Three different variables were created: all inflows, all outflows, and the net mass balance variable. The model structure did not violate any natural laws (structural assessment test [14]) (i.e., rams did not produce progeny). This was tested by checking if the model produced unrealistic values, such as negative stocks of naturally non-negative values (stocks).

In addition, the model was not sensitive to the changes created for time-dependent calculations or the integration method, which showed model stability (i.e., integration error test). This was performed by the model settings tool incorporated in Stella Architect, which was used to test for errors by changing the delta time to half its original value. Additionally, different integration methods were executed (Runge–Kutta 2 and 4).

The model was evaluated for extreme conditions (i.e., extreme conditions test). For example, the model’s main output—in the base-case scenario—was 394 reported human cases and a prevalence in sheep of 0.119. The outputs of each simulation run with extremely low (0.0001), and high (0.95) values of sheep prevalence showed the expected decrease and increase, respectively, in reported human cases. The expected behavior of the model was to produce more human cases if the number of infected sheep increased. The model predicted an increase in human cases when the fraction of re-stocked infected ewes was increased, which increased the prevalence of sheep brucellosis. Conversely, it predicted a decrease in human cases with a reduced prevalence of brucellosis in sheep.

3.3. Sensitive Variables

The results of the sensitivity analysis are presented in Figure 3. These variables were chosen based on their expected potential change to the number of reported human brucellosis cases as advised by the opinions of the consulted experts and the proposed interventions identified in the literature. The changes in “farmers visits”, “market restrictions”, “awareness”, “test and cull infected sheep”, “slaughter aborted ewe”, “vaccination”, and “immunity duration” were associated with the greatest changes in the number of reported cases of human brucellosis.

3.4. Simulation of the Proposed Intervention Scenarios

The model simulations produced four outcome measures that showed the impacts on animal and human populations, including the cumulative number of reported human cases, amplitude of human case, cumulative number of sheep abortions, and the prevalence of brucellosis in sheep compared to the base case.

3.5. The Cumulative Number of Reported Cases of Human Brucellosis

Figure 4 shows the actual (base case) and predicted numbers of reported cases of human brucellosis throughout the study period. The results of the simulations show that each proposed intervention reduces the cumulative number of reported human cases with the most decreases achieved using the One Health (37%) and “test and cull” (29%) interventions. The remainder of the interventions had modest to little impacts on the predicted number of reported cases of human brucellosis over the study period.

One Health Outcomes

Table 1. The combined outcomes of interventions for brucellosis in Jordan using a system dynamics model.

Intervention Scenarios	Combined Reduction in Brucellosis
Test and cull	−154%
One Health	−135%
Market trade restrictions	−58%
Slaughter of aborted or suspected	−48%
Awareness	−20%
Farmers visit intervention	−17%
Combined vaccination and immunity	−16%
Enhanced immunity	−14%

shows the values and percent changes in the combined model projections of the four outcome measures.

The radar chart (Figure 5) shows the four output measures for each simulated intervention compared to the base case (outermost line) of the observed number of reported cases of human brucellosis in Jordan and the corresponding projected cases in sheep.

The interventions designed to directly reduce Brucella transmission in sheep, “test and cull”, “market trade restrictions”, and “slaughter of aborted/suspected”, were the most effective compared to other scenarios having an estimated overall reduction (total change) of 154%, 58%, and 48%, respectively. The “test and cull” intervention was the most effective in reducing sheep prevalence (48% reduction) and the intensity of human cases (61% reduction). The combined intervention (“One Health”) was the second most effective intervention that reduced brucellosis across all output measurements by 135%.

Interventions designed to modify human behavior, “awareness” and “farmers’ visits intervention” had a lower-case reduction ability, 20% and 17%, respectively. The least effective interventions involved changes in sheep vaccination programs (“standard intervention” and “enhanced immunity”) that reduced brucellosis by 16% and 14%, respectively.

4. Discussion

This was the first study to develop an SDM to understand Brucella transmission and simulate interventions to reduce the incidence of brucellosis in any context. It was also the first time the potential benefits of a One Health approach were quantitatively demonstrated. The benefits of understanding the whole system [11] are critical because the model provides policy makers with a greater understanding of the connections between brucellosis system structures, their behavior, and how the whole system changes over time [23]. In addition, the SDM enables the iterative development of real-world interventions through a rapid simulation of potential scenarios that enables decision makers to quantify the likely outcomes and identify points of cooperation or collaboration [24]. The model approach helps overcome the real-world limitations related to the data and uncertainty about specific elements of the brucellosis system by allowing the integration of several forms of evidence, including expert opinion and information provided in government reports [25].

Previous epidemiological studies (including those from Jordan) focused on the direct relationship between the disease agent, vector, and host [7,20,26,27] to identify risks [28,29,30]. These studies do not explain the underlying complex relationships that drive transmission dynamics, particularly the feedback links between the disease and the characteristics of the sectors involved. This study demonstrated the non-linear relationship between the bacterium, sheep, and people as a continuous complex interaction between Brucella sectors, such as food safety [31], animal infection [32], other risk factors [26], and health systems (animal and human).

The reported and simulated patterns observed in the reported cases of brucellosis reflect its cyclic (oscillatory) behavior with peaks in the incidence every 3–4 years. This pattern has not been documented in the past [22] with any cyclical patterns attributed to the annual impacts of seasonal patterns in sheep breeding [33]. The model structure suggests that the multi-year cyclical behavior may arise from longer-term effects due to delays in the impact of sheep vaccinations and changes in human behavior as a result of the changes in their knowledge of brucellosis. This information is useful in the development of more accurate risk assessment studies that account for such delays and sectoral associations.

4.1. Intervention Scenario Discussion

4.1.1. Scenario One: Farmers’ Visits Intervention

The important link between farmers’ visits to veterinarians, vaccination, and human Brucellosis cases has not previously been explored in the context of Jordan [34]. This is important because attempts to move veterinary services beyond animal and food production to control zoonotic diseases were unsuccessful [34,35]. In our research, we highlight the importance of this link because, in Jordan, it is a focal point for access to veterinary services, such as vaccinations and disease control.

We found that improving the number of farmer visits to veterinarians resulted in a predicted modest reduction in human incidence. However, this intervention relies on changes in farmers’ behavior that are difficult to achieve because of the delays between farmer visits and follow-up disease investigation and vaccination activities [6]. These delays lead to a loss of trust in the veterinary services by farmers [36], which affects the visit frequency and rate, eventually leading to vaccination failure. Additionally, seeking veterinary services is not typically a priority for most farmers [35], particularly when they suspect brucellosis, due to a fear that their animals will be culled without compensation, which is the current situation in Jordan [37]. The observation that the scenario will only have a negligible (<1%) impact on sheep abortions, despite a marked reduction in human cases, shows that this scenario trades off an animal health expense that benefits human health.

Service delivery, particularly in veterinary health, is a worldwide public benefit that requires a sustainable investment. In general, poor financial support is associated with an insufficient capacity to deliver suitable levels of veterinary services [38]. Therefore, the main implementation challenges to achieve improved levels of veterinary service delivery are the availability of financial and logistics capabilities required to create cooperation between the stakeholders. This will eventually gain the farmers’ trust and confidence and encourage them to seek and visit veterinary services.

4.1.2. Scenario Two: Market Trade Restriction

Unlike other studies in Jordan, this study explored local markets’ involvement in brucellosis. Markets may facilitate local (Jordan) and international (cross-border) sheep movement and mixing [5]. The control of trading infected sheep significantly reduced the consumption and contact risks for humans by reducing the Brucella load in the flock. This dual effect led to greater reductions in sheep abortions, brucellosis prevalence (sheep), and human incidence cases (16%, 19%, and 11%, respectively) compared to the farmers’ visits scenario.

Sheep trade and husbandry practices are fundamental sources of income and food security for most farmers in Jordan. The significant productivity impacts of brucellosis threaten food security [5]. The World Organization for Animal Health (OIE) considers that trade surveillance and restrictions are valuable strategies to control brucellosis and that market surveillance is an important factor in disease elimination and control campaigns [39]. This is important because a recent study identified that the uncontrolled, formal and informal, sheep trade was widespread, and a major factor associated with high rates of brucellosis in Jordan [20]. The proposed “market restriction” scenario in our study successfully reduced brucellosis by 58% and was ranked the third best scenario. These results highlight the importance of controlled trade in limiting the risks of irregular sheep markets through market trade restrictions, as currently there are no restrictions on formal or irregular sheep trade in Jordan. A recent report from Colombia where brucellosis outbreaks were attributed to unrestricted livestock movements [40] showed the effectiveness of this intervention. A policy that used testing to establish disease freedom prior to the movement of animals was associated with a substantial reduction in the number of outbreaks [40].

4.1.3. Scenario Three: Awareness of Farmers Intervention

The observation that improving farmers’ personal hygiene and use of personal protective equipment through improving their awareness leads to a modest reduction in human incidence cases of brucellosis, but not the prevalence in sheep, is understandable given that it does not directly reduce transmission in sheep. This scenario was modeled because it is a standard approach to reduce human cases advocated by public health authorities [6,30]. According to previous studies, only 38% of farmers in Jordan have adequate knowledge gained mainly through advertisements and doctors’ and veterinarians’ advice [20]. The awareness and knowledge of infectious diseases have been shown to reduce infection risk and be associated with improved health-seeking behaviors in people at risk or infected [30,41], which reduces the human incidence of disease. This is important because the simulation results are congruent with the WHO recommendations of preventing occupational exposure and the consumption of unpasteurized dairy products through public health education and awareness programs to control brucellosis [29].

4.1.4. Scenario Four: Test and Cull Intervention

The results of the “test and cull” intervention showed the greatest overall reduction in brucellosis in both human and animals (61% and 48%, respectively). This intervention reduces disease intensity because it is designed to cull infected ewes only, which directly reduces brucellosis prevalence, and eliminates consumption and contact risks. These results are plausible given that culling infected sheep to eliminate brucellosis is considered a highly effective control strategy [42,43].

In Jordan, there are no national campaigns or strategies to cull infected animals [37]. However, most sheep owners voluntarily culled diseased sheep that aborted fetuses due to the disease, by at-home slaughter, or they sold them to a butcher [20]. The absence of suitable compensation programs for farmers is one of the main implementation and enforcement challenges of such programs, as many farmers do not approve the culling of their sheep without compensation, particularly if illness is not apparent [20,44]. Moreover, forced culling without compensation may lead to increased illegal animal movements that would increase the rate of disease transmission [45]. Therefore, although culling is the most cost-effective strategy to control brucellosis, several economic, social, and other factors should be evaluated before selecting this strategy for implementation [43]. Since evidence is needed to justify animal culling as a control strategy, this study provides a reasonable justification that this program can be considered as an option to control brucellosis in Jordan; however, there is a need to consider financial compensation.

4.1.5. Scenario Five: Slaughter of Aborted/Suspected Sheep

This scenario is less effective compared to other intervention scenarios because the scenario relies on voluntary changes in the behavior of farmers without governmental or professional supervision. In addition, farmers often do not necessarily assess the risks posed by keeping infected sheep compared to the financial losses they perceive if they slaughter their sheep [36]. The lack of knowledge of farmers with regards to brucellosis [28] also reduces the effectiveness of this scenario.

Previous studies have found that, in most brucellosis cases, farmers prefer to manage the disease occurrence independently rather than call a veterinarian [46]. This is because of the lack of trust [36] associated with the fear of slaughter of infected sheep without compensation, and the perception that the brucellosis vaccine causes abortions [20]. Therefore, despite half of the sheep owners in Jordan being aware of brucellosis risks, there is no systematic approach to in-house slaughtering practices [20].

4.1.6. Scenarios Six and Seven: Vaccination and Immunity Effectiveness

Sheep vaccination is the main governmental control strategy in Jordan [37]. Although considered a successful brucellosis control strategy, previous studies found an extremely low vaccination coverage (<2%) and low efficiency of these vaccination programs [47]. The results of these scenarios indicate that vaccination and enhanced immunity scenarios are unlikely to be effective, even when projected for the new vaccination strategy adopted by Jordan that involves extending the duration of immunity of the vaccine Rev-1. This can be explained as sheep vaccination in Jordan is performed at veterinary clinics only and there is no regular national vaccination campaign [37]. Therefore, vaccination is the farmers’ responsibility and dependent on their compliance and commitment. This dependency creates problems with the implementation because it relies on farmer behavior, like scenario five, the slaughter of aborted/suspected sheep. Consequently, this dependency affects the vaccination rates, even though vaccines are available and provided free of charge [6].

The efficacy of control programs by Rev-1 immunization varies between countries [48]. For example, in Spain, brucellosis prevalence remained high (6.5%) following 6 years of a government funded vaccination program that aimed to vaccinate the whole sheep population, because of the poor vaccine quality and low vaccination coverage [48].

The compliance of farmers to interventions is a fundamental requirement for any successful control strategy, like vaccination (also see Scenario 1). Our simulation results for this scenario re-emphasize the previous discussion regarding a farmers’ roles in brucellosis control and emphasize the importance of their compliance to the effectiveness of such strategies. Therefore, unless there is a good compliance of farmers, sheep vaccination and immunity strategies may remain inefficient.

4.1.7. Scenario Eight: One Health

Most of the local and international organizations, like the WHO, provide a continuous and increasing interest in health programs that apply the One Health approach [49]. One of the most significant findings of this study was the inclusion of a One Health control strategy. The results of the One Health scenario show that it results in the greatest reduction in human brucellosis and is the second most effective scenario across all intervention measures (135% reduction). This was mainly because it included interventions from several sectors of the system that had a synergistic effect reducing human and sheep brucellosis cases.

These results are consistent with several previous studies that recommend the “One Health” management approach to prevent epidemics and epizootics [50]. Additionally, other studies concluded that the application of holistic and multiple-scale approaches, like One Health, enabled the control of several infectious diseases, such as schistosomiasis and avian influenza [50,51]. This One Health intervention involved the main pillars of the One Health approach: human (awareness), animal (aborted slaughter and enhanced immunity), and the environment (markets) [51]. These results emphasize the significance of the One Health approach and the importance of the multi-sectoral relationships embedded in the system that drive its behavior and propose the achievement of better outcomes in both human and animal sectors. Therefore, stakeholders should be directed toward a new paradigm that shifts their attitudes from considering one sector’s responsibility (Ministry of Agriculture) as the responsibility of all involved sectors. This may create a shared interest and a common goal to be pursued by all the relevant stakeholders.

Recently, in Jordan, efforts were made towards the One Health approach to map out the current networks and explore the points of communication, coordination, and decision making that permitted several sectors and stakeholders to act interactively and investigate the priorities and gaps, which limited the sharing of information [52]. Despite the existence of the informal communication across sectors, particularly during emergencies, formal routine coordination was absent. Therefore, these findings set the foundations and directions for the coordination, potential benefits, and outcomes of adopting the One Health approach. Additionally, these results provide theoretical evidence of the usefulness of the One Health approach for public health interventions.

5. Conclusions

The results of the individual focus interventions clearly demonstrate the trade-offs that occur between the competing goals of reducing the impacts on sheep vs. reducing the incidence in humans.

The One Health intervention was the most effective method because it leveraged the previously unrealized synergies within the system. Further work is required to validate this model and explore the actual efficacy of combined One Health interventions.