ABSTRACT
Previous investigations suggest that the prevalence of Mycobacterium avium subsp. paratuberculosis (MAP) in Swedish cattle is low and all recent cases have been linked to imported animals. The aim of this study was to evaluate the surveillance system for MAP infection in Swedish cattle and to estimate the probability that the Swedish cattle population is free from this infection. Calculations of surveillance sensitivities and probability of freedom were made using stochastic scenario-tree modelling, which allows inclusion of information from several different sources, of complex surveillance data including results from non-representative sampling, as well as of documentations of differences in risk of being infected. The surveillance components included in the model were: (1) clinical surveillance, (2) fallen stock investigations, (3) the national surveillance programme (mainly beef herds), (4) a survey involving dairy herds and (5) a risk-based survey targeting herds with imported cattle. Previous or current presence of imported animals and participation in the on-going control programme was specified for each tested herd, in order to adjust for differences in risk. Calculations were made for each year from the start of 2005 to the end of 2008, and this formed the basis for a final estimate covering the whole study period and predictions of future probabilities of freedom from MAP. Results show that when applying a design prevalence of one animal in 0.1% of the herds, the probability of freedom at the end of 2008 was 0.63. At the design prevalence of one animal in 0.5% of herds, the estimated probability is >95% and it is demonstrated that the prevalence of MAP in Swedish cattle is below this level or absent. In order to increase the annual surveillance sensitivity in the future and thereby improve the probability of freedom, new surveillance activities or an intensification of current ones are needed.
Subject(s)
Animal Husbandry/methods , Mycobacterium avium subsp. paratuberculosis/isolation & purification , Paratuberculosis/epidemiology , Animals , Cattle , Dairying , Epidemiological Monitoring/veterinary , Models, Theoretical , Paratuberculosis/microbiology , Population Surveillance/methods , Prevalence , Seasons , Sensitivity and Specificity , Stochastic Processes , Sweden/epidemiologyABSTRACT
In July 2007, PRRS was detected for the first time in Sweden. A total of eight positive herds were identified and various measures were taken to eradicate the disease, including restrictions and slaughter of infected herds. Subsequently, both active and passive surveillance activities were undertaken. This study describes stochastic scenario-tree modelling of all the various surveillance system components, to estimate the current probability that Sweden is free from PRRS. The model includes all actions taken after the first positive herd was detected. The surveillance system components included in the model were as follows: investigations undertaken in association with the outbreak, a serological study based on samples collected at slaughter, samples collected in the national PRRS surveillance programme and passive clinical surveillance. The probability of freedom was estimated in time steps of 1 month, from July to December 2007. After each time step, the calculated posterior probability of freedom from the previous month, combined with the probability of introduction, was used as a prior probability for the next month. The result from the model showed a 99.8% probability that Sweden was free from PRRS at the end of December 2007. The estimated total sensitivity of the surveillance system varied between 81.2% and 94.3% and was highest during the first months after the outbreak. For sensitivity analysis purposes, the model was also applied using higher risks of introduction. However, this did not make considerable difference to the final estimates.
