Antibiotic use in pig farming and its associated factors in L County in Yunnan, China

Abstract China has a long history of pig rearing, and it currently raises and consumes approximately half of the pigs in the world. Major improvements have been made in pig farming in China in the last four decades with the growing application of new livestock farming technologies. Among the new improvements, the use of antibiotics in pig farming is a common but not well‐documented practise. In order to understand the behaviour of the farmers regarding antibiotic use in pig farming, we conducted a household survey in four townships of L County in Yunnan Province, China, during August 2014 and April 2015. In this survey, 404 farmer households were interviewed using a questionnaire. Among the farmers interviewed, 89% reported easy access to antibiotics, 83.7% reported experience of self‐purchasing antibiotics, and 40.3% reported that they often used antibiotics in pig farming mainly for the prevention and treatment of pig diseases. These farmers identified 20 antibiotics that they had used in pig farming 6 months before the survey. Of these, 11 and 8 antibiotics have been categorised under ‘critically important’ and ‘highly important’ antimicrobial groups, respectively, by the World Health Organization (WHO), and 12 and 8 have been categorised under the ‘Watch’ and ‘Access’ groups, respectively, as per the 2019 WHO AWaRe classification of antibiotics. Factors associated with the behaviour of self‐purchasing antibiotics included types of farms, sources of antibiotics, and previous experiences of pig diseases: those who were smallholders, buying antibiotics from veterinary drugstores and village vets, and whose pigs had suffered diseases previously were more likely to self‐purchase antibiotics for their pigs. Farmers who cleaned their pigsties less frequently and those whose pigs had suffered from diseases used antibiotics more frequently as compared to their peer farmers.


| INTRODUC TI ON
China is the world's biggest meat producer by far, and it currently raises and consumes approximately half of the pigs in the world (Elliott, 2015). Moreover China has a long history of pig rearing (Cucchi et al., 2016;Kuo, 2013). In the last four decades, major improvements have been made in pig farming in rural China with the growing application of new livestock farming technologies, such as new breeds, feed, vaccination and veterinary drugs. The new technologies have greatly enhanced the productivity of pig farming by increasing the supply of pork to meet the increasing market demand.
However, these practises have had profound implications, both positive and negative, for public health. The positive aspects include improved nutritional status of the population and increased income of pig farmers, whereas negative aspects include, but are not limited to, environmental pollution caused by unutilised pig manure and potential threats of zoonoses. The application of antibiotics in pig farming is an increasing but not well documented and regulated practise.
Studies have revealed that samples of pig manure and soil collected from large swine farms in China contained diverse and abundant antibiotic resistance genes (Mu et al., 2015;Zhu et al., 2013), many of which were found in chicken and human faeces as well (Ma et al., 2015). While it is well acknowledged that antibiotics are widely used in pig farming in China (Wang et al., 2017), the behaviours of pig farmers regarding antibiotic use, particularly smallholder farmers, are neither well documented nor understood due to insufficient research. In order to regulate the use of antibiotics better in animal husbandry and to limit antimicrobial resistance, we undertook re-

| Study site
L County in the Yunnan Province of China was selected as the study site for several reasons. First, it is a poverty-stricken county with a long history of pig rearing that has been promoted by the local government to alleviate poverty. Second, this county presents diverse forms of pig farming practises ranging from large-scale modern pig farms to traditional smallholder pig rearing due to its mountainous geography and presence of ethnic minority groups. Third, the research team has better geographic access to this county because of its distance from Kunming (100 km), the capital city of Yunnan Province. In 2014, there were 13 townships and 160 administrative villages with a total population of 411.6 thousand, and 32% of them belonged to ethnic minority groups. Among the 13 townships in this county, four townships and eight administrative villages with two administrative villages per township were selected for this research. The selection was made after consulting with the County Animal Husbandry and Veterinary Bureau (CAHVB) to identify the townships and villages with the greatest number of pig farmers and highest density of pig population in the county. Thus, the selected townships and villages were not random, but purposefully chosen samples. All four townships and eight villages showed similar characteristics in terms of economic development level, culture, and veterinary services, and they were all located in the central part of this county. This county cannot be claimed as the representative of all counties in China, but it definitely represents some counties in the Yunnan Province showing similar geographic, socioeconomic and cultural conditions.

| Household survey
The household survey was the main research method employed in this study. Interviews with local farmers, vets and staff members from the CAHVB were conducted as a supplementary method, with the findings used to design the survey questionnaire. Pig rearing requires inputs such as pigsty, feed, cleaning and disease prevention, and these factors affect the incidence of pig diseases and the subsequent treatment involving the use of veterinary medicines, including antibiotics. Therefore, we designed questions to investigate the types of pigsties, methods and frequency of pigsty cleaning, frequency of disinfection, sources of pig procurement, sources of drinking water, vaccinations and sources of feed. The main contents of the questionnaire included general demographic information of Impacts • This survey, conducted in a county in the Yunnan Province of China, showed that 83.7% of the farmers reported self-purchasing antibiotics for their pigs and 40.3% expressed that they often use antibiotics in pig farming mainly for the prevention and treatment of pig diseases.
• These farmers reported 20 antibiotics that they had used in pig farming in the last 6 months before the survey. Of these antibiotics, 11 and 8 have been categorised under 'critically important' and 'highly important' antimicrobial groups, respectively, by WHO, and 12 and 8 have been categorised under the 'Watch' and 'Access' groups, respectively, as per the 2019 WHO AWaRe classification of antibiotics.
• Factors associated with the behaviour of self-purchasing antibiotics included types of farms, sources of antibiotics, and the experiences of previous pig diseases, and the factors associated with the farmers often using antibiotics for their pigs included the frequency of cleaning pigsty and previous pig diseases. the pig farmers, pig farming size and rearing practises, previously encountered pig diseases, and the knowledge, accessibility and use behaviour of the farmers regarding antibiotics.
The sample size of the household survey was calculated using the statistical formula provided below, and the result was 384 farmer households. Considering incomplete questionnaires or missing data, we decided to expand the sample size to 450 households with 110-113 per township and 55-56 per administrative village. Prior to the formal survey, we pre-tested the questionnaire between 17 and 19 july 2014. We investigated 50 pig farmers from two other villages of this county and found that many farmers could not recollect the name of the antibiotics that they had administered to their pigs in the last 6 months. Therefore, we consulted with local veterinary service providers and farmers to identify antibiotics that were most commonly used in the county, and selected 18 antibiotics to list in the questionnaire as an index. When we undertook the formal household survey, our trained investigators read the names of these 18 antibiotics one by one to the respondents and asked them whether they had used any of those antibiotics in the last 6 months, or if they had used any antibiotics apart from the 18 listed ones. Thus, we collected relatively accurate information from each surveyed farmer on antibiotic use.
We planned to survey all pig farmers in the selected villages of the four townships, but not all farmers were available when we visited their houses for the survey. Hence, the surveyed farmer households formed a convenience sample, but they accounted for more than 80% of all pig farmers of the selected villages because we revisited those farmer houses at another time to capture those who had been missed previously. When starting the survey, the investigators explained the purpose and process of the survey, promised confidentiality of the participants' identity information, and obtained oral informed consent from the farmers. Approximately 40-50 min were spent to complete one household questionnaire, with some even taking over an hour. A plastic washbowl costing around 1.5 USD was given to the surveyed farmers as a small gift to thank him/her for their time when the interview was concluded.

| Data cleaning, entry and analysis
The completed questionnaires were brought to Kunming Medical University. Data were dual entered into EpiData 3.1 and cross- Large-scale farmers were those who kept more than 30 heads, and smallholder farmers were those who kept fewer than 30 heads. We then compared the behaviours of antibiotic use of the farmers and the associated factors between the two groups using the chi-square test or non-parametric test, and analysed the factors that affected the farmers' antibiotic use behaviours in pigs using binary multivariate logistic regression. Two behaviours were selected as dependent variables for the multivariate logistic regression analysis: self-pur-

| Basic demographic information of the respondents
Of the 404 surveyed farmers, 215 were female (53%) and 189 were male (46.8%). More than 70% (77.2%) of the surveyed farmers were aged 40 years and above, and 18.1% were aged 60 years and above.
Han ethnicity, the majority group in China, accounted for 88.6% of the surveyed group, and the remaining (11.4%) belonged to ethnic minority groups. Approximately two-thirds (59.9%) of the farmers had an education level of primary school or below, 34.2% attended middle school, and merely 5.9% had an education level of high school and above. Male farmers had a better education level than their n = Z 2 P(1 − P) 2 female counterparts (χ 2 = 12.784, p < 01), and majority (93.3%) of the farmers were married (Table 1).

| General situation of pig rearing of the surveyed households
Among the 404 surveyed farmers, 86 (21.3%) were considered large-scale farmers and 318 (78.7%) were smallholder farmers.

| Farmers' accessibility to antibiotics
Among the surveyed farmers, 89% acknowledged that it was easy for them to buy antibiotics for pigs, and 93.3% of them reported that they could buy antibiotics over-the-counter without a prescription issued by a vet (Tables 3 and 4).
Regarding sources of antibiotics, 54.8% and 26.5% of the farmers (total, 81.3%) reported that they mainly purchased antibiotics from local veterinary drugstores and village vets, respectively ( clearly prohibited the use of human medicine, including human antibiotics, in animals. The qualitative interview data revealed that the main motivations for using human medicine for pigs included the perceived better quality of human medicine and dual use of human medicines for both human beings and animals after purchase. In some cases, farmers used the leftover antibiotics from a human disease treatment course of family members to treat the sick pigs.

Farmers' behaviours regarding self-purchasing antibiotics
In this study, self-purchasing antibiotics was defined as farmers purchasing antibiotics for pigs over-the-counter, without consulting a vet or presenting a prescription issued by qualified vets. Of the 404 surveyed farmers, 83.7% reported self-purchasing antibiotics, suggesting that this behaviour was common among the surveyed farmers.
There were no statistically significant differences in the behaviour of self-purchasing antibiotics between the farmer groups with respect to sex, age, education level, ethnicity, training on livestock rearing and years of pig raising (data not presented). Table 6 shows that more large-scale farmers (93%) reported this experience as compared to smallholder farmers (81.1%).
Furthermore, farmers who used hygienic pig houses and chemical disinfectants for pig house disinfection reported this behaviour more than those who used traditional pigsties and quicklime for disinfection. However, these differences could have been caused by confounders.
We performed binary multivariate logistic regression to identify factors that affected the farmers' antibiotic self-purchasing behaviours. The results showed that farm types, sources of antibiotics and previous pig diseases were linked with self-purchasing behaviour. Smallholder farmers, farmers who bought antibiotics from veterinary drugstore/village vets, and those whose pigs had previous diseases were more likely to self-purchase antibiotics for pigs as compared to the others (Tables 7 and 8). Table 9 shows 40.3% of the surveyed farmers expressed that they often used antibiotics in pig farming. Farmers who received training on livestock raising provided by local animal husbandry and veterinary systems, cleaned their pig house more frequently, used chemical disinfectants for pig house disinfection, and had encountered pig diseases were more likely to report the frequent use of antibiotics than farmers who did not. However, these differences might be caused by confounders. There were no differences in the frequency of antibiotic use among the farmers of different sex, age, education level, farm types and years of raising pigs (Table 9). Using the frequency of antibiotic use as a dependent variable, we performed binary multivariate logistic regression to identify factors that affected this behaviour. The results showed two independent variables were included in the model at p < .05: frequency of cleaning pig house and previous pig diseases. Farmers who cleaned their pig houses less frequently were more likely to use antibiotics than those who cleaned them more frequently. Similarly, farmers whose pigs had suffered from diseases were more likely to use antibiotics than those whose pigs had not (Tables 7 and 10).

TA B L E 3 Farmers' accessibility to antibiotics
In this study, 62% of the farmers stated that the purpose of using antibiotics was to treat pig diseases, 35.7% claimed that their use of antibiotics was to prevent pig diseases, and only 2.3% of the farmers said they used the drug to promote pig growth (Table 11).
Thus, treatment and prevention of pig diseases were the two major purposes of antibiotic use reported by the surveyed farmers. There was no statistically significant difference between the large-scale and smallholder farmers regarding the purpose of using antibiotics.
However, there was a statistically significant difference between the male and female farmers (χ 2 = 3.962, p < .05). More male farmers used antibiotics for disease prevention, whereas more female farmers used antibiotics for disease treatment. There were no differences in the purpose of using antibiotics among the two farmer groups with respect to age, education levels and years of raising pigs (Table 11).
Moreover we used three hypothesised conditions to explore the farmers' use of antibiotics further, namely when pigs had cold, fever and diarrhoea, and whether the farmers would use antibiotics to treat these conditions. The results revealed that the majority of smallholder and large-scale farmers (78.3%-84.9%) reported that they would use antibiotics for the aforementioned conditions, and there was no statistically significant difference between the two groups (Table 12).
Although it is difficult to evaluate the appropriateness of using antibiotics for those pig symptoms by asking a simple question, cold, fever and

| Major antibiotics used by farmers in pig rearing
Using the household survey, we investigated the antibiotics used by  (Table 13). These antibiotics were the most commonly used ones by the surveyed farmers in pig rearing. The other 10 antibiotics used are also presented in Table 13.
We sorted the 20 antibiotics into nine different classes based on their chemical structures.

| Antibiotic knowledge of the surveyed farmers
We designed 15 questions in the household survey questionnaire to test the farmers' antibiotic knowledge. knew the regulation that buying antibiotics needs a prescription issued by a vet. We divided the 404 farmers into two groups based on their antibiotic knowledge level: higher knowledge group (correct answers for 9 or more of the 15 questions, 60%), and low knowledge group (correct answers for 8 or less of the 15 questions). Next, we compared the behaviours regarding 'self-purchasing antibiotics' and 'frequency of antibiotic use' of the two groups using the chisquare test. There were no statistically significant differences found between the two groups (p > .1 for both; data not shown). Thus, the surveyed farmers' antibiotic knowledge was not included as an independent variable in the multivariate logistic regression analysis.

| Unrestricted antibiotics use in pig farming and weak regulation
Our survey data showed that both large-scale and smallholder farmers in this county had easy access to antibiotics, and 40% of them often used antibiotics in pig farming, which is similar to that in other developing countries (Dang et al., 2013). The data showed that they also did not comply with the regulations; instead, certain actively recommended drugs, including antibiotics, were sold to farmers who were looking for treatment for their pigs without asking for a prescription. Some pig-feed sellers sold antibiotics to farmers in order to prevent diarrhoea and other diseases that may occur after eating the feed, particularly diarrhoea that may occur at certain production stages of pigs, for example, moving from weaners to growers. While pig farming has been controlled or even banned in some areas of China in order to control environmental pollution caused by animal farming (Li, 2013;State Council, 2016), pig farming is encouraged in this county by the government as a way of income generation and poverty alleviation, and the pig population density has increased as com-

| Purpose of antibiotic use
The main purpose of using antibiotics in pig farming as reported by farmers in our survey sample was to prevent and treat pig diseases to avoid the economic loss caused by pig deaths, which is different from the behaviour of using antibiotics as a growth promoter. The results of multivariate logistic regression also suggest that farmers whose pigs had diseases in the past were more likely to self-purchase which is similar to our findings in this study. Since most farmers in our study were older than 40 years with little school education and training on livestock husbandry, along with easy accessibility to antibiotics, their use and misuse of antibiotics in pig farming were very likely. This finding suggests that if we want to reduce the farmers' use and misuse of antibiotics in pig farming in this setting, we need to find alternative ways other than antibiotics to help them in the TA B L E 11 Purpose of using antibiotics in pig raising * prevention and control of pig diseases and the associated economic loss.

| Classification of antibiotics most commonly used by the surveyed farmers
Farmers identified 20 antibiotics that they had used in pigs in the 6 months before the survey. Although the magnitude of using these antibiotics, including frequencies and dosages, was unknown due to the limitations of this survey, it is worth noting that 11 and 8 of those antibiotics have been categorised as 'critically important' and 'highly important' antimicrobials, respectively, by the WHO in 2013.
Moreover 10 were categorised as 'prioritisation of critically important antibiotics' (WHO, 2013). Among the nine classes of antibiotics, the quinolones and macrolides classes to which four of the 20 listed antibiotics belonged, were listed by the WHO using three criteria in 2013 as the 'highest priority critically important antimicrobials'

Antibiotics
No. of farmers who reported using % and 'Access' groups, respectively, as per the 2019 WHO AWaRe classification of antibiotics (WHO, 2019). Therefore, further research is warranted to confirm the antibiotics used and to monitor the antimicrobial resistance associated with their use.

| Use of human antibiotic products to treat pigs
Both animals and humans need antibiotics to treat bacterial infections, but certain antibiotics usage is indicated for human be- products to treat sick animals may be a common practise among farmers (Li, 2008;Ning et al., 2018;Ren, 2015;Wu et al., 2013).
Although the farmers' main purpose of using human antibiotic products to treat sick pigs was pursuing better efficacy, this behaviour not only violates the Act, but also increases the risk of antimicrobial resistance to human antibiotic products. This is because bacteria with antimicrobial genes developed in animals consumed as food could be transferred to humans via food chains and other channels (Landers et al., 2012). In fact, both human and animal health requires antibiotics, and many antibiotics are common between the human healthcare and veterinary sectors. Therefore, we need to develop strategies to find a good balance between the two sectors, and more importantly, contain the development of antimicrobial resistance.

| Limited knowledge of antibiotics and antimicrobial resistance among farmers
The surveyed farmers exhibited poor knowledge of antibiotics, and they used antibiotics indiscriminately like the other veterinary medicines. More than two-thirds of the farmers did not know the concept of antimicrobial resistance. Given that most farmers had only primary school education, it is not surprising that they had poor knowledge of antibiotics and low awareness regarding antimicrobial resistance. Notably, a survey conducted in Switzerland revealed that Swiss pig farmers were less aware of the risks of antibiotic usage in pig husbandry (Visschers et al., 2014). This suggests that farmers in both developed and developing countries need to be educated on the risks and consequences of using antibiotics in animal husbandry.
Alternatively, it suggests that education does matter as it is overridden by economics. Further research is needed to explore ways that can more accurately capture the behaviours of farmers, antibiotics and their quantities used in such settings, and role of veterinary drugstores and vets at the grassroots level, and to monitor the antimicrobial resistance associated with such using behaviours.

| CON CLUS ION
It has been increasingly acknowledged that antibiotics used in To what extent the antibiotic use in this pig farming system contributes to antimicrobial resistance and antibiotic residues in pork products and the environmental pollution is unclear. Further research is warranted to explore these questions.

ACK N OWLED G M ENTS
We thank International Development Research Centre (IDRC),

Canada and Innovative Research Team of Yunnan Province
(2019 (6)) for providing financial support to undertake this research. We also thank the students of Master Degree Program of Public Health in Kunming Medical University and the staff members of the County Animal Husbandry and Veterinary Bureau at the study site who participated in the household survey of this research.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest. Investigation.

PEER R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1002/vms3.390.

Jing Fang
https://orcid.org/0000-0002-7357-8976 E N D N OTE 1 Pig farming has a long history in this county. In the past, almost every farmer household raised several pigs, and they kept pigs in simple shelters that were usually a part of the farmhouse located close to the toilet in order to collect pig faeces together with human waste for fertilizer. Hence, the pig rearing condition was unhygienic, and this type of pigsty was called a traditional pigsty. In order to reduce pig morbidity/mortality and promote productivity, the agricultural extension workers have promoted 'hygienic pigsty', that is a pigsty with better hygienic conditions, such as a separate pig house with cement flooring and walls that is easy to clean.

R E FE R E N C E S
Chen, X., Wu, L., Xie, X., Zhu, D., Wang, J., & Zhang, X. (2016). Influence of pig farmer characteristics on improper use of veterinary drugs.