Various rubrics of surveillance modalities, ranging from targeted surveillance [ 2 ], risk-based surveillance [ 1 ], and participatory disease surveillance [ 3 ] have emerged in response as a way of making such programs more cost-effective; recent research has synthesized these terminologies [ 4 ]. Much of the focus in the veterinary epidemiology literature, on risk-based surveillance in particular, has been on technical criteria associated with risk and risk factors that underpin disease, and the means by which disease searching and monitoring better account for such criteria [ 1 ]. By contrast, analyses that focus on the resource allocation side of the surveillance equation are much more limited.
Recently, Cannon [ 5 ] reviewed different metrics of surveillance resource optimization problems based on different objectives by decision-makers e. Such optimization approaches to surveillance have been applied in other related areas including fisheries surveillance [ 6 ] and invasive species management in the ecology literature [ 7 - 9 ], but have not generally been utilized in the veterinary literature. Rather, surveillance programs are generally analyzed in more broad terms, such as a component of a simulation analysis of disease mitigation options [ 10 - 15 ] or in benefit-cost analyses using economic welfare indicators producer and consumer surplus [ 16 , 17 ].
Rather than focusing solely on the tradeoffs between surveillance and intervention, the authors highlighted the need to jointly consider both types of expenditures together to maximize the net benefits associated with avoided disease losses. In the context of risk-based surveillance, Prattley et al. Earlier work by Mariner et al. Farm-level predictors included characteristics such as the production system employed, monitoring scheme membership, frequency of veterinary contact, herd size, and farm-level record keeping.
Resource Allocation Theory Applied to Farm Animal Production
However, none of these risk-based approaches looked explicitly at how best to allocate resources given different risk factors. An additional, and important, gap in all of these analyses is couching the optimal allocation of resources however defined , with the risk factors that might significantly modulate their effectiveness in mitigating disease itself.
Hauser and McCarthy [ 8 ] come closest in this regard by paying particular attention to the spatial allocation of resources based on the efficacy and presence of disease over a landscape. Despite this, their analysis was static, and did not link the allocation and efficacy of surveillance resources to subsequent disease incidence in future periods. However, in an animal disease setting, where the behavior of agents can actively influence the epidemiology of disease [ 25 ], it is critical to evaluate resource allocation in its appropriate dynamic systems setting. Put differently, decisions by policymakers in surveillance programs will have an effect on the behavior of producers and other actors in the agri-food chain, which in turn can influence the epidemiology of disease, and consequently, the cost-effectiveness of surveillance systems over time.
These feedbacks can be complex and not intuitive, requiring more nuanced approaches to their economic analysis. Related to this, and a crucial factor which has rarely been considered, is that the nature of surveillance itself is not static. Surveillance programs will differ in nature not only based on the context of the disease, but also on the efficacy of disease control programs and the objectives and goals of decision makers themselves.
They considered three stages of disease: i sustainment, in which the surveillance objective is either to maintain disease freedom or detect disease; ii investigation, in which the objective is to obtain more information about an endemic or epidemic disease; and iii implementation, in which surveillance serves as an information source for mitigation options. Such a conceptual framework highlights the dynamism inherent within surveillance programs and strongly suggests a need for system-based empirical approaches to address them.
In this paper, we provide a more robust conceptual framework for the allocation and composition of surveillance resources, overlaying the socio-economic drivers of risk and disease response alongside the biological and spatial dimensions of disease. In this manner, the paper builds on two recent analyses that examined resource allocation issues in a disease setting. First, it extends the work of Cannon [ 5 ] by adding the systems dimension regarding the behavioral aspects, attitudes, constraints, and practices of producers and other recipients of surveillance resources [ 24 ].
The approach is akin to the recent analysis of Duintjer Tebbens and Thompson [ 28 ] that analyzed alternative decision rules for resource allocation in its dynamic epidemiological context. However, our approach highlights the interface of the disease epidemiology with different types of actors in the agri-food chain based on their risk profiles [ 25 ], while maintaining the alternative decision rule metrics of past analyses.
An advantage of this framework is that it can accommodate significant heterogeneity and feedback mechanisms in this socio-economic overlay, based on data availability and the nature of the disease in question. Our analysis first presents a series of general principles that underpin this approach, followed by a generic example of an application of these methods to disease control in Scotland.
A starting point for our analysis is to first place surveillance efforts in their epidemiological and socio-economic contexts. One way to conceptualize surveillance efforts is in their contribution to the reduction of disease.
The smooth shapes of such decay curves are more prominent in diseases in which wide-scale eradication programs are implemented, as illustrated in the examples of polio, smallpox, and malaria [ 28 ]. At different points along this decay curve, one will utilize different types and mixes of surveillance and mitigation strategies to bring disease to lower and lower levels. In endemic settings, the decay curve might level off at some non-zero level or oscillate in regular intervals on the basis of a variety of agro-ecological, climatic, socio-economic, or other factors Figure 2.
This latter point is particularly salient in many disease instances, but also illustrates a fundamental point often overlooked in the literature; namely, surveillance activities do not take place in a vacuum. As different programs are put into place, they will have a direct influence on the epidemiology of the disease through control measures , and an indirect effect on behaviors taken by recipients of surveillance services that will modulate positively or negatively those efforts.
Meanwhile, other changes take place that are not directly associated with the decay curve or its surveillance programme but nevertheless have an important influence on them. These might include progressive changes in the agricultural industry e. Figure 3 illustrates these interactions in a causal loop diagram, a commonly used tool in the system dynamics literature to illustrate the feedback mechanisms present in complex systems [ 29 ].
Note that Figure 3 is an abstraction of many of the behaviors implicit in this system and necessarily excludes a number of key influences to simplify the analysis. The inner loop of the diagram provides the standard conceptualization of the role of surveillance — greater intensity in disease surveillance leads to more detections, and correspondingly more control measures, which over time should have the effect of bringing disease down to its desired level possibly zero.
However, as the diagram also illustrates, both surveillance programs and control measures impose costs on producers and others in the agri-food chain.
This might directly influence the level of production less animals produced , modulating the incidence of disease downward. On the other hand, it could also lead to more risky behaviors that increase the risk of disease and lead to more trade to meet consumer demand from areas with more or less disease risk. These risk factors can be further segmented based on the type of disease considered endemic vs.
- Contact Info;
- Resource Allocation Theory Applied to Farm Animal Production.
- INIA & ITACYL | SusPig ERA-Net?
The importance of various factors will necessarily be disease-specific, with some factors much more important than others depending on the disease. Countervailing this further are the pressures surveillance and control programs face from resource constraints on budgets, which place limits on the level of efforts that can be made from surveillance. An important aspect of Figure 3 is in directly illustrating the relationships between the disease, its socio-economic risk factors, and the role surveillance efforts play in influencing this feedback structure. Correspondingly, any optimal allocation of surveillance resources needs to account for these dynamic impacts as additional constraints.
An important related issue further illustrated in Figure 3 involves distinguishing who carries out surveillance activities themselves i.
ØªÙØ§ØµÙÙ Ø§ÙÙ ÙØªØ¬
These motivations might be rather different. From the standpoint of the public sector, regulatory mandates including compensation schemes in place might shape protocols for surveillance, while for the private sector, these may be driven by competitive factors, such as the need to differentiate products in the marketplace. Among private sector actors themselves, and in different agri-food chains e.
Furthermore, surveillance activities conducted by the private and public sectors may overlap. There are a couple of potential methods that more explicitly model the complexity of the problem discussed above. Such methods move away from optimization approaches [ 5 ], though these methods can represent a first approximation to the complexity of the surveillance problem, albeit without feedback effects implicitly considered.
Resource Allocation Theory Applied to Farm Animal Production
One method of setting up this type of problem is to model the system illustrated in Figure 3 directly as a system dynamics or SD problem. In this manner, the allocation of surveillance and mitigation resources can follow various decision rules established within the model [ 28 ], as can various strategies for disease control itself [ 30 ]. In this manner, Homer and Hirsch [ 31 ] examined the tradeoffs between diagnostic and therapeutic interventions in a generic model of public health.
As will be seen shortly, such an approach can be easily adapted in a model of animal disease surveillance. A further advantage of a systems approach is the ability to overlap relevant socio-economic drivers that might influence the allocation of resources themselves. For instance, Ulli-Beer et al.
Rich [ 25 ] proposed a way to link economic decisions with biological drivers of disease, but cost implications of alternative strategies from the epidemiological side were not considered. The manner in which economic agents are modeled can take a number of forms, depending on data availability, level of analysis, production system, and spatial diversity. In this model, a simple S-I-R model of disease spread is developed that traces the evolution of animals or herds between different disease states of nature.
In the diagram, the rectangles represent stocks of animals or herds at any given period of time, i. In the mathematical language of S-I-R models, these would be the differential equations that would underpin the movement of actors between states. The small circles and thin arrows that connect them to stocks, flows, or other circles are parameters that relate stocks and flows and parameters.
In the model, transitions between susceptible and recovered via vaccination and recovered to susceptible due to waning immunity are included, while other disease states included latency and incubation periods could also be added [ 30 ]. A powerful advantage to modeling in iThink is the graphical representation of complex, non-linear systems of differential equations. Indeed, behind the graphical interface are functional forms that relate stocks, flows, and parameters in line with standard epidemiological theory. At the top and bottom of Figure 4 are two diamond shapes that denote decision processes.
His research suggested that feed intake could be broken into two parts:. Through the use of these two items, animals that deviate from the expected feed intake, either by eating more than or less than what's expected of them, can be identified and managed as the farmer sees fit.