Animals and sampling
According to the data published by the Gansu Province Bureau of Statistics, there were 6.89 million swine in 2020 (Table 1). To obtain a sample size representative of the Gansu swine population, the number of pigs to be tested was calculated with EPITOOLS (https://epitools.ausvet.com.au/oneproportion). The minimum sample size was calculated to be 664 with an assumed STB prevalence of 50%, a confidence level of 99%, and desired precision of 5%. According to the ratio of the number of swine in each city to the number of swine in the province, the sample size was calculated and presented in Table 1.
Then, a cross‐sectional multistage sampling strategy was used for selecting the study population. Firstly, scaled pig farm (intensively raised three-crossbred pig number ≥ 200) was the primary sampling unit, and at least three farms were randomly selected in each city. It is worth noting that the pigs raised in Gannan Tibetan autonomous prefectures are free-range Tibetan pigs. Secondly, swine were the secondary sampling units, and at least 30 swine were randomly selected from each selected farm. Pigs were selected using systematic sampling, in which the first pig was randomly selected, and subsequent pigs were selected with a fixed sampling interval. In addition, there were 180 serum samples collected from Guangxi province.