Foot-and-mouth disease (FMD) is a viral disease primarily of cloven-hoofed animals that can profoundly affect animal husbandry by evolving into severe epidemics that reduce the productivity of susceptible livestock.
There are several potential risk factors associated with both introduction and spread of the FMDV infection in the region. A possible mechanism for the introduction of FMD into the EU might therefore be an undetected epidemic in wild boar in countries adjacent to Member States. Further transmission within the EU could also be related to susceptible wildlife, notably wild boar.
Hunting activities are considered valuable for surveillance in the hunting areas, especially at the forest edges, and are a possible option for surveillance particularly in summer.
The SAS unit was requested to provide an assessment on the performance, in terms of early detection, of different monitoring and sampling schemes, including hunting, in case of an incursion of FMD virus in a free wild boar population area.
Two main output parameters for the comparison between the surveillance strategies under investigation are considered: 1) the Probability of detection overtime, i.e. the weekly probability of detecting at least one sero-positive animal and 2) the Number of weeks needed to first detection, i.e. the minimum number of weeks needed to detect at least one sero-positive animal. Both output parameters were obtained by simulating a one year time period.
The first step of the assessment estimated the sample size needed to detect at least one positive animal for a given prevalence of sero-converted individuals, with a given level of confidence, using the formula of Cannon (2001). Various prevalence values and different testing system sensitivities were considered, including the parameters used in Council Directive 2003/85/EC, i.e. 5% prevalence to be detected with a 95% Confidence Level.
The second step consisted in the development of a simulation model able to generate the true sero-status of the population. From this simulated population it was possible to virtually implement different sampling strategies and calculate the probability of detecting at least one sero-positive animal for a given time period, after assumed exposure to infection. To obtain the number of weeks needed to detect the first case, the difference between the week of first detection (recorded in the simulation) and the week of incursion was calculated.
Simulation results suggest that with a sampling strategy based on hunting it takes a considerable amount of weeks (13 to 39) to detect the first sero-positive animal. On the contrary, a sampling strategy based on a regular weekly sampling over time obviously performs relevantly better as independently from the time of the incursion and the probability of spreading, on average the maximum number of weeks needed to detect the first sero-positive animal is 5. When simulating a regular sampling a system based on a prior fixed sample size over time performs better than a system based on a constant proportion in terms of variability of the probability of detection over time and time needed to first detection (~ 1 week less). This is due to the fact that Cannon’s formula used to calculate the number of samples needed to detect a disease takes into account a possible lack of sensitivity.
The simulation results show that, in the investigated sensitivity range (i.e. 85% - 100%), the sensitivity of the test/surveillance system has almost no impact on the number of weeks needed to detect the first case. On the contrary, the specificity of the testing system, must be 100% in order to avoid the reporting of false-positives which could produce misleading results.