ESTIMATING THE PROBABILITY OF FREEDOM FROM PRRS IN SWEDEN: THE QUICK AND DIRTY METHOD OR A MORE THOROUGH APPROACH - DOES IT MATTER?

摘要

After the first outbreak and eradication of porcine reproductive and respiratory syndrome (PRRS) in Swedish pigs in 2007, several surveillance activities were performed and almost 25,000 samples were analysed with a negative outcome. Considering the efforts made and resources spent to control this disease, there was a further need to evaluate the surveillance system and to substantiate disease freedom. In this study, four different approaches were used to estimate the surveillance system sensitivityand probability of freedom from disease: 1. Evaluating the total number of test results using Win Episcope, 2. Evaluating results from structured surveys using FreeCalc, 3. Scenario-tree model of the complete surveillance system, assuming independence ofobservations, 4. Scenario-tree model of the complete surveillance system, adjusting for clustering and overlapping of observations. The approaches were chosen to represent different levels of complexity of calculations and required quality and detail ofinput data. The comparison showed that, in this case, the methods gave very similar estimates of surveillance sensitivities and probability of disease freedom. In fact, the estimated probability of disease freedom only varied between 99.8% and 100.0%. However, the different approaches required different degrees of information. In conclusion, although the choice of approach did not make any major difference to the final estimates, the more advanced approach was considered preferable since it produced amore reliable estimate.