Uncertainty and sensitivity analyse

Uncertainty and sensitivity analyses. As a variable response in the Box-Bhenken design, the percentageof sick animals at the end of lactation, nursery, and fattening period were used. Therefore, three response surfaceswere obtained (Supporting Information). The percentage of infected sows, R 0 and lethality are the parametersthat have a significant effect on the percentage of sick animals at the edge of the nursery and fattening period.However, the probability of transmission of the disease during the lactation phase slightly affects during thisproduction phase. Using the response surfaces obtained, the sensitivity was studied in all three phases of the pigbreeding process (Supporting Information).The increase of infected piglets is very small (from −3‰ to 3‰) during the lactation phase in the range of valuesused for infected sows. However, in the nursery (from 1.5 to 5.9%) and fattening phase (from 0.5 to 5.77%), thisincrease is bigger using low versus high level of infected sows. During the rearing period (nursery and fattening), anincrease of one point (1%) in the percentage of infected sows implied an increase between 3 and 6% of infected pigs.In the latter two cases, the value decreases as the percentage of infected sows increases (Supporting Information).In the case of the R 0 parameter, an increase of one unit implies similar variations in the percentage of sick ani-mals during the nursery (from 2.1 to 5.8%) and fattening phase (from 2.6 to 6.2%) in the range of studied values.However, in the lactation phase, this increase is practically nil (from 0.02 to −0.94%).In the case of variations in the lethality value, the increase in sick animals does not practically vary duringthe nursery (from −0.9 to 0.1%) and fattening period (from −0.7 to 0.1%) in the range of studied values. In thelactation phase, it increases as the lethality increases, but the variations are very small (from −1.2 to 1.3%).