Computer simulation to support policy-making in Aujeszky's disease control

Aujeszky's disease is a contagious viral disease that affects the central nervous system of pigs. Several eradication programs or measures are available, each of them providing different results. Determining the preferred strategy is to a large extent a matter of economic consideration.Under the EU rules, countries or regions that are Aujeszky-free can ban imports of breeding animals carrying antibodies of the disease; movements to Aujeszky-free areas from other areas of both breeding and rearing pigs are subject to strict conditions and controls, which differ depending on whether or not the area of origin has an EU-approved eradication program. If important import destinations achieve disease-free status, exporting countries that have failed to eradicate the disease will be severely penalized. Therefore, sterner demands are to be expected considering control and eradication of Aujeszky's disease in the Netherlands in the future. To meet these demands, the objective of this study was to develop a computer simulation environment in which "what-if' scenarios can be performed to explore the epidemiological and economic effects of different Aujeszky's disease control programs. The model can be used to support the choice of the optimal eradication program under various conditions, in particular from an epidemiological and economic point of view.First, a flexible economic framework to evaluate Aujeszky's disease eradication programs was developed, and illustrated with an example (Chapter 2). The framework has four elements: changes in percentage of infectious herds, changes in product quantities, changes in product prices and economic integration. Each of these elements is defined as a separate module in the simulation model and has its own input and output data, depending on the control strategy under consideration. With these elements all epidemiological and economic aspects of the disease can be monitored over time.In an illustrative example, probability distributions of the number of infectious herds corresponding to each control strategy were compared and the optimal strategy was chosen, according to the risk attitude of the decision maker. The framework can be considered a standardized approach in comparing and selecting animal health control strategies by integrating technical and economic data and principles.To obtain epidemiological information with respect to the control of Aujeszky's disease virus, an epidemiological state-transition simulation model was constructed to evaluate the spread of the virus (Chapter 3). In the model, the population of herds in the Netherlands is subdivided into four herd types: great-grandparent stock+multiplier, rearing, farrowing and fattening. Every time step, each herd is in one of 32 states per herd type. The states are based on (1) the reproduction ratio R ind , which is the number of individuals infected by one infectious individual, (2) the prevalence for each value of R ind and (3) the expected number of infectious animals in an infectious herd within each prevalence range of the herds. The different values Of R ind are based as much as possible on field data and experiments, where different vaccination strategies were applied.The transition matrix with the probabilities of every possible transition from one state to another was calculated on a weekly base. With this matrix the distribution of herds over states from week to week was derived. To include the non-linearity of the transmission process, the transmission probabilities from non-infectious to either non-infectious or infectious were developed such that they depend on the state vector itself The fraction of herds that becomes infectious equals one minus the fraction of herds that has not been infected by the virus emitted by infectious herds.Calculations revealed that infection in the Dutch pig population would not disappear without vaccination, nor with a vaccination scheme in which sows were vaccinated less than 3 times per year and fattening pigs once per cycle (Chapter 4). The infection, however, would be eradicated within 2 to 3 years, if sows were vaccinated 3 or 4 times per year and fattening pigs twice per cycle. The outcome turned out to be sensitive to the impact of other than animal contacts on the number of new effective virus introductions per time unit.The structure of the production pyramid and herd density in the affected regions were other important factors which influenced the course of infection. To examine the impact of these factors the total number of herds in the Netherlands were further subdivided into four regions (North, East, West-Middle, South).Outcomes showed that the percentage of infectious herds in equilibrium was highest for rearing herds (76.3%) and lowest for great-grandparent stock+multiplier herds (20.0%) if no vaccination was done. The herd type "fattening" had more impact on the effectivity of the different vaccination strategies than the herd type "farrowing". This difference is becoming less if more intensive vaccination strategies are applied. Besides the difference in herd type, also herd density and the percentage of non-vaccinated herds were an important factor in the eradication process.After simulating these epidemiological characteristics of Aujeszky's disease virus, market outcomes and pig producers' returns were simulated under different scenarios with respect to closure of export markets for live piglets and fattened pigs (Chapter 5). If the Netherlands fails to eradicate Aujeszky's disease before its trading partners in these markets, live piglet exports would be banned, reducing industry revenue and export earnings by about 9% and 10% respectively in the medium term. If exports of live fattened pigs are also banned, the reductions are 26 and 32% respectively. The piglet-producing sector would be moreseverely affected than the fattening sector. The model also showed that, if export markets for carcasses were also to close for an unspecified food safety reason, capacity of the industry would fall over 50%.Lastly four control strategies to eradicate Aujeszky's disease virus in the Netherlands were compared epidemiologically and economically (Chapter 6). Vaccination decreased the number of cases per production loss. The decrease was largest if vaccination strategy changed from "no vaccination" to the less intensive vaccination. Extra vaccinations under more intensive vaccination strategies, however, still had impact. The attendant costs were highest per dead animal (especially for gilts) and per abortion. Growth delay of gilts and piglets turned out to be of minor importance.The sales distribution on the piglet markets (import, export and on the domestic markets) was particularly influenced by vaccination, but the decreases in revenues were only less than 4.3%. The only exception was the number of piglets and live animals that were imported into the Netherlands, which decreased by more than 15% and about 9% respectively. The accompanying revenues from piglets and fattened pigs were highest if "no vaccination" was done. Compared with the revenue in this strategy, this difference is greatest on the piglet market, as the decrease in revenue was about 3.6%, while the decrease was about 0.55% on the market of fattened pigs.According to the resulting present values over a period of 10 years, "no vaccination" is economically the best solution only if no trade restrictions are to be expected. Economically speaking, however, the most intensive vaccination strategy should be applied, if an export ban of two years on live animals to, for instance, Germany is expected within 10 years after the start of the vaccination strategy. A prolonged export ban makes this strategy even more favourable. From an economic point of view intermediate vaccination strategies are never preferred.The main conclusions of this thesis are:- State-transition simulation proves to be an appropriate method to evaluate transmission of Aujeszky's disease virus. The epidemiological information obtained can well be used in economic evaluation of different control strategies.- Aujeszky's disease is only eradicated in the Netherlands if the most intensive vaccination strategy (≥3 times per year) is applied for breeding sows, and fattening pigs are vaccinated at least once per cycle.- If applying the most intensive vaccination strategy, it takes about 200 weeks for an average herd to become non-infectious.- The relative impact of other than animal contacts on the number of new virus introductions increases from 4% to 98%, if the vaccination strategy is changed from "no vaccination" to the most intensive vaccination program.- Subdivision of the total population into herd types and regions is important to enhance insight into transmission of infection in the pig population and to support decision making at regional level.- Price equilibrium models can well estimate the short-term changes in prices as well as those in the medium term. To accomplish this, it is of importance that sufficient historical data about quantities, prices and the infections occurred are available to estimate the required parameters accurately. A monthly data-set of about 10 years turned out to be sufficient.- Direct production losses from Aujeszky's disease virus are less than 6% of the vaccination costs when vaccination is carried out. More than 80% of these losses are caused by growth delay of fattening pigs.- The most intensive vaccination strategy (i.e., sows are vaccinated 3 or 4 times per year and fattening and rearing pigs twice per cycle) is economically preferred if an export ban on part of the live animals is expected during at least 2 years within 10 years after the start of the vaccination program. If this is not to be expected then "no vaccination" turns out to be the best strategy. The risk of an export ban on live animals should justify the eradication of the virus from the population.- For the current situation in the Netherlands it is economically preferred to start blood sampling all sows and remove the gE-positive animals instead of continuing vaccination, provided that the additional risk of new introductions of the virus is sufficiently limited.