Assessment of drivers of antimicrobial use and resistance in pig and poultry farming in the Msimbazi River Basin in Tanzania

Uncontrolled use of antimicrobials both in humans and animals coupled with environmental contamination by euents, toxic chemicals, and heavy metals exacerbate the development and spread of antimicrobial resistance. This paper assessed the drivers of antimicrobial use and resistance in poultry and domestic pig farming and the environment in a densely populated area, with wide-ranging human activities in Tanzania. This was a cross-sectional study conducted from June to September 2019 in the Msimbazi River basin in Eastern Tanzania, which covers an area of 271km2 and harbours a population of approximately 1.2 million. Questionnaires, in-depth interviews, and focus group discussions (FGDs) were used to collect information regarding demographic characteristics, knowledge, practices, attitudes, and perceptions of the drivers of antimicrobial use and resistance in animal farming and the environment. Descriptive statistics and the chi-square tests were used to analyse data from the survey and p<0.05 was considered as statistically signicant. The FGDs and interviews were transcribed, coded into categories, and analysed manually. reliance on antimicrobials, as well as a review of government strategies, policies, and regulations on AMU. Descriptive percentages generate the results in tables and gures. Also mean, median and the standard deviation was computed for continuous variable. The Chi-square test was conducted to identify the association between the outcome and explanatory variables. The value of p < 0.05 was considered signicant. Information from the focus group discussion and the interview was subjected to transcription, creating categories, and then coded into categorical variables as per (31) and were analysed manually. The results for the overarching themes were presented. research to evaluate the active ingredients of antimicrobials consumed in specic animal species in the study area and across the country in order to get the true reection of the magnitude on AMU in food animals.


Introduction
In Tanzania, the increased demand for short-cycle animal stocks such as poultry and pigs has led to intensive animal production of these animals (1)(2)(3)(4). The poultry and domestic pigs are in most cases managed by women as essential elements for their income and empowerment (1,5,6). According to national projections, the annual chicken meat and pig production was expected to increase from 130,000 tonnes in 2017 to 465,600 tonnes in 2020 and from 22,000 in 2017 to 37,200 tons in 2022, respectively (7). This increase has been attributed to a number of factors including increasing urbanization rate and increased trade of live animals and animal products (4).
The increasing animal production has been associated with several challenges (8)(9)(10) including a high number of animals being con ned into poor quality shelters with limited space (1,11) overstretched veterinary extension services and lack of implementation of disease control strategies (12)(13)(14). Expectedly, the frequency and magnitude of infectious diseases are high (15), compelling farmers to excessive antimicrobial use (AMU) for prophylaxis and/or treatment of diseases in order to reduce management costs and maximize returns on investment (2,10,12). The sources of antimicrobial agents used by farmers vary widely but are mostly from hawkers and informal drug dealers who have little prescription knowledge (16,17), and often leading to their misuse (12,18). Over the counter sale of antimicrobial agents is widespread in low-income countries (19,20). This situation is compounded by a lack of coordinated animal surveillance system, weak enforcement of food safety regulations, lack of basic knowledge on AMU and resistance among the livestock keepers, and infection prevention and control (IPC) strategies in animal production (14,16,21,22). smartphone. Interviews were conducted at the household/farming premises. The heads of the households were the main respondents.
In the absence of the head of households, information was sought from any other adult (> 18 years old) occupants who were engaged with livestock keeping. The questionnaire contained close and open-ended questions covering four main themes: (i) sociodemographic information such as sex, age, marital status, education level, and experience in animal farming; (ii) knowledge, attitude, practices and perceptions on AMU and AMR in poultry and/or domestic pig farming; (iii) drivers of AMU and development of AMR in animal farming; and (iv) environmental contamination resulting from animal wastes. Additionally, observation of the antimicrobial used and stored, type and hygiene of the housing and personnel, type of feeds, waste management, and disposal were done and recorded into a different sheet. The closed-ended questions were coded into categorical variables and the open-end questions and observed information was organised into subcategories before analysis.

Focus Group Discussions and in-depth interviews
The FGDs and in-depth interview participants were both male and female selected from farmers, poultry and pig buyers, and community members. The aim was to collect complementary information from different categories of respondents who are directly or indirectly involved in the AMU and are at risk of exposure to AMR. The time taken for each FGD was 2-3 hours and the in-depth interview was 30-45 minutes. The FGDs and interview guide questions were prepared to capture their views about drivers of AMU and AMR, availability and accessibility of veterinary drugs, animal treatment practices, environmental contamination, waste management practices, and the role of the government concerning policy, regulation, and control of veterinary antimicrobials.

Data Management And Analysis
The collected quantitative information was sent to a server, integrated and mapped to produce information that was transferred to the Microsoft Excel spreadsheet and analysed using the SPSS version 20.0 for Windows (IBM Corp., Armonk, NY, USA) software. The outcomes concerning the drivers that were assessed using the 4-point Likert scale were initially described with numbers and percentages. They were then dichotomized as "True" versus "False" such that the value of "strongly agree" and "agree" (as True) versus "Strongly disagree" and "disagree" (as False). Descriptive statistics such as frequency and percentages for categorical variables were determined to generate and summarize the results in tables and gures. Also mean, median and the standard deviation was computed for continuous variable. The Chi-square test was conducted to identify the association between the outcome and explanatory variables. The value of p < 0.05 was considered signi cant. Information from the focus group discussion and the interview was subjected to transcription, creating categories, and then coded into categorical variables as per (31) and were analysed manually. The results for the overarching themes were presented.

Ethical Considerations
The Medical Research Coordinating Committee of the National Institute approved this study for Medical Research of Tanzania Informed verbal consent was obtained before the commencement of the study. Participants were informed of the opportunity to withdraw from the study at any time without prejudice.

Results
A total of 113 farmers responded to the questionnaire, out of which 59 (52.2%) were females. Their age ranged between 18 and 69 years (mean age was 46.2 years and standard deviation of 11.8). The majority had secondary school education (46.9%, n = 53) and were married (63.7%, n = 72), and most of them (96.5%, n = 109) had been farming for more than six months (Table 1).
Knowledge and practices regarding AMU and AMR in poultry and pig farming About 88.5% of the respondents ever heard about antimicrobials, and 69.9% indicate to know how to use them for treatment of their animals. However, the majority (92%, n=104) did not know AMR (Table 2). A signi cant relationship was detected between the respondent's education level (p=0.04) and hearing about antimicrobials. Farming experience has a signi cant relationship with the knowledge of antimicrobial use (p=0.03) and the knowledge on the source of antimicrobial use (p=0.04). cases (95.6%, n=108) diagnosis was based on clinical signs. The majority (84.1%, n=95) had access to veterinary services, and most of them (89.4%, n=101) bought antimicrobials from veterinary centres. The level of education was signi cantly associated with the frequency of using antimicrobials (p=0.04) and performing group treatment (p=0.04) irrespective of disease conditions. Stocking of antimicrobials was reported by many respondents (62.8%), mainly for 1-2 months. Group treatment of animals irrespective of disease condition was very common (88.5%, n=100). There was a signi cant relationship between farming experience and the method used for disease diagnosis (p=0.04), access to veterinary services (p=0.01), storage of veterinary antimicrobials (p=0.03) and stocking veterinary antimicrobial at home (p=0.03).
Perception and attitudes on antimicrobial use Over half (59.3%; n=67) of the respondents indicated farming must be accompanied by AMU, mostly for prophylaxis (97.3%, n=110) than treatment (80.5%, n=91) and the majority (66.4%) knew how to administer them ( Table 4). The level of education was signi cantly related to the possibility of reducing antimicrobial use while attaining maximum production (p=0.01). Many (66.4%, n=75) admitted that AMU in animal farming may pose a risk to human health, and some of them (53.1%, n=60) agreed that it was possible to reduce antimicrobials use in animal farming and yet achieve maximum production. Farming experience has a signi cant relationship with the drugs not being effective (p=0.03) in treating animal diseases.
Perception and attitudes on the drivers of antimicrobial use and resistance About two-thirds of the respondents reported that the use of a combination of drugs is necessary for effective treatment, and nearly a half (48.7%) indicated to using the human medicines, and private drug sellers in uence uncontrolled AMU. The majority of the respondents (92.9%, n=105) acknowledged having inadequate knowledge on infection prevention and control of animal diseases, and most of them (95.6%, n=108) indicated that veterinary and extension o cers were inadequate. The use of human antimicrobials in animals (p=0.004), inadequate veterinary extension o cers (p=0.000), and inadequate knowledge on infection prevention and control of animal diseases (p=0.005) were signi cantly associated with the level of education (Table 5). About two-thirds indicated that pro t maximization necessitates the misuse of antimicrobials to shorten the period of farming, and 72.6% were aware of withdrawal periods.
Factors associated with the development and spread of antimicrobial resistance in the environment Respondents indicated the following to be the main factors associated with the spread of AMR in the environment; disposal of solid wastes from the household (78.8%, 89), agricultural activities that involve use of animal manure (92.9%, n=105), uncontrolled disposal of human and veterinary drugs (85%, n=96) and use of river water for irrigation (92.9%, n=105). As shown in table 6, the farming experience was signi cantly related to the presence of agriculture activities that use animal manure (p=0.001); the use of river water for irrigation (p=0.001); and disposal of wastes from the household (p=0.004). Surprisingly, many (82.3%) did not think that pharmaceutical industries, which discharge e uents directly into the river, contribute to the spread of AMR.
Overarching themes from in-depth interview and FGD Analysis of qualitative ndings from interviews and FGDs generated several themes regarding drivers for Antimicrobial use and resistance. The main emerged themes were poultry and domestic pig diseases, access to antimicrobials, exposure to animal disease diagnosis and treatment, unavailability of government extension o cers, waste management and environmental contamination, and the adherence to withdrawal periods.
Diseases affecting poultry and pigs, the reason for diseases occurrence and drugs used for treatment Interviewed farmers reported that the most common diseases affecting poultry and pigs were typhoid, respiratory diseases, and diarrhoea in poultry and worms, skin diseases, diarrhoea and African Swine Fever (ASF) in pigs. Our study respondents mentioned different groups and names of the drugs used. Upon detailed analysis of the mentioned drug varieties, the research team grouped them into eleven classes with the main one being tetracycline and quinolones. The major causes for a disease outbreak that were mentioned included, poor animal management practices, temperature changes, and contaminated water, feeds and the environment ( Fig. 2).

Antimicrobial Use And Accessibility
The use of veterinary antimicrobials for treatment and prevention of diseases was familiar to all respondents. The ndings from FGDs revealed that high usage of antimicrobials was attributed to the high frequency of disease occurrence and easy accessibility of antimicrobials from private veterinary drug sellers. It was reported that drug sellers were insisting and encouraging farmers to use drug combinations for effective treatment. Most of the FGDs participants reported that the disease burden was a result of the poor management systems. One of the FGDs participants reported "It is close to impossible to keep poultry or pigs without using antimicrobials. Animals get sick frequently and the drugs are easily accessible. The sellers insist on the use of a variety of drugs to get the best production. Instead of helping us on appropriate use, they just want to sell more. Another FGDs participant had a similar comment stating, "Drugs are used for treatment but mainly for prevention of diseases that frequently attack animals. The use of drugs save the life of the animals and consequently retain the investment capital"

Exposure To Diagnosis And Treatment Of Diseases
Respondents admitted that the exposure they had in animal farming has enabled them to be familiar with the diseases; therefore they were capable of detecting and treating their animals. They explained that they even collect samples from sick and dead animals for laboratory investigation. One of the focus group discussants had this to say: "The exposure in animal farming enables one to tell the type of disease based on the signs and symptoms. The treatment instituted saves the lives of animals and the costs of consulting private extension o cers. There is a possibility of either under/over-dosing because the amount of drug to be administered is guessed! Misdiagnosis is also likely to occur". Furthermore, study respondents indicated that private veterinary/animal health practitioners were expensive and were mainly involved in the treatment in pig diseases such as mastitis, coccidiosis, and severe respiratory infections.

Availability And Accessibility Of Government Extension O cers
Most of the respondents were not aware of the existence of the government veterinary services within their streets/wards. The private veterinary/animal health workers who owned most of the veterinary drug selling centres were working with the farmers. "It is like the government does not employ livestock extension o cers anymore and if they do, they just stay in o cers probably because they don't want to, or they lack facilities for visiting farmers" (Key informant (KI) number 2). Another FGDs participant reported that "There is much trust in their services and it is likely that working closely with them will reduce diseases burden through improved animal management practices".
Laws and regulations regarding the handling, use and dispensing veterinary drugs Almost all participants were not aware of the laws and regulations restricting veterinary drug handling, distribution, and sale. They reported that they obtain antimicrobials over the counter and have never experienced di culties in obtaining drugs for their animals.
On the other hand, veterinary drug sellers and animal feed manufacturers and sellers were aware of the presence of laws and regulations but acknowledged weak implementation. "There are laws and regulations but the implementation is weak" (KI number 7). However, none of them could name the speci c laws and regulations.

Waste Management With AMR Development And Spread
The majority of the respondents admitted that the wastes generated from different sources might contribute to AMR development and spread. They reported that wastes from domestic, human, and veterinary drug selling points were not sorted before collection and disposal. Additionally, wastes from the collection centre were not disposed of on time or daily basis so they are most often scattered all over the place, thus they may become a potential source of AMR. "Waste from different households, e uents from the hospital including the nearby hospitals, industries like breweries and pharmaceuticals and others are lodged in this area during ooding. All these waste might contain residues of antimicrobials that contribute to the development of AMR in the environment" (KI number 5).

Withdraw Period Of Antimicrobials
The use of human drugs was mentioned to be common in pig and poultry farming, mostly amoxicillin, doxycycline, ampicloxacillin, tetracycline, chloramphenicol eye drops, cipro oxacin, metronidazole, and chloramphenicol. Almost all key informants and FGDs participants admitted to being aware of the withdrawal period, however, they found it di cult to implement for the fear of economic loss. One of the FGDs participants reported, "Farmers are aware of the withdrawal period but it is close to impossible to adhere with the time stipulated". In another interview it was reported, "The capital invested is small, and it is a loan from the community serving cooperatives. The pro t obtained from selling the poultry, domestic pig, and eggs are the ones used to run the family, feed the producing ock (for layers) and the remaining herd, and repay the loan. Waiting for the withdrawal period or discarding the eggs means extra costs which are di cult to overcome" (The KI number 8).

Discussion
This study was conducted in Tanzanian's most densely populated area with highest concentration of both pharmaceutical and commercial industries, hospitals, pharmacies and drug sellers, intense use of antimicrobials including insecticides in crop and animal farming, and intense environmental contamination (24,25,27,32,33) and is therefore certainly the hottest hotspots for the spread of AMR organisms between human, animals and environment in the country. Additionally, the Msimbazi River drains to the Indian Ocean, which is international water bodies therefore, the practice of AMU and AMR re ects wider implications.
We found a very high usage (87.6%) of veterinary antimicrobials in poultry and pig farming mostly for prophylaxis rather than treatment. The high usage might be attributed to the prolonged exposure in animal farming coupled with negligence in adhering to good hygienic practices and other biosecurity measures. The higher proportion in this study corresponds to the one reported in Sudan, Ghana, Nigeria, and Cameroon where routine antimicrobials use are the normal practices even in absence of disease outbreak (31,(34)(35)(36). Predominantly, farmers were making treatment decision based on presentation of clinical signs, which carries a high possibility of misdiagnosis, administration of inappropriate drugs and improper dosage. Similar studies in Sudan and Nigeria reported that antimicrobial use without involving veterinary practitioners and lack of laboratory ndings in diagnosis of animal diseases leads to improper use of antimicrobial (31,37).
Disturbingly, many farmers were using human medicines including amoxicillin, doxycycline, ampicloxacillin, tetracycline, chloramphenicol eye drops, cipro oxacin, metronidazole and chloramphenicol to treat animals. Due to limited extension services and poor animal health delivery systems, farmers tend to buy veterinary and non-veterinary drugs from private drugs shops and treat their livestock themselves. There is high chance that farmers seek to maintain animal health welfare by using drugs that are cheaper, readily available, easy to use and more effective. The use of human medicine in animals proves the weakness in implementation of laws and regulation governing handling, sale and use of antimicrobials both in human and animal and limited extension services. This weakness spurs the problem of AMR organisms. This nding is consistency with (35,38) that antimicrobial use in food animals is accompanied by lack and or weak regulation, limited veterinary services and higher costs of private veterinary consultants.
Apparently many farmers were stocking veterinary drugs at home, with a likelihood of poor handling and storage. The stocking of drugs might interfere with the active ingredient leading to reduced e cacy of the antimicrobials as reported by (39,40). Moreover, metaphylaxis was very common, implying that even the healthy animals in a herd or ock were also treated thus escalating the chance of AMR development Similar ndings was reported in Tanzania (10) that stocking of veterinary antimicrobials was associated by poor improper usage leading to drug residues in animal products.
Frequently, farmers were using a combination of veterinary drugs on advice of nancially motivated veterinary drug sellers who do not have enough knowledge on drug use, side effects, proper dosage and withdrawal period taking advantage of inadequate veterinary services and extension o cers. Similar ndings were reported in Cameroon and Ethiopia that private veterinary services are expensive (36,41) and the veterinary professionals are responsible for the abusive use of antimicrobials (42). We found tetracycline and quinolones being the most used antibiotics due to being cheap, readily available and accessed easily, without restrictions (12,43), and are often sold by informal vendors at informal markets and even along the road (10,44). Use of tetracycline in animal has been documented in Cameroon and Ghana probably since they are cheaper and has broad-spectrum activity against variety of diseases (34,45,46). The wide spread use of quinolone observed in this study corresponds to the one reported in the previous studies (37,47,48) that quinolones are widely used in animals for diarrhoea treatments and prophylactic despite the fact that they are expensive.
Antibiotics are commonly sold during cattle auction days by informal vendors, such as petty traders and livestock keepers. Antibiotics found in markets like these are often unregistered, and therefore sold at very cheap prices (Abdu Hayghaimo-Former director of Veterinary services in Tanzania, personal communication, April 1, 2013). The quality of these medicines is undetermined. economic loss ii) regulatory bodies in Tanzania have not yet set withdrawal periods for veterinary drugs and farmers rely, primarily, only on veterinary drug sellers. As a result, farmers rarely comply with the recommendations and the responsible regulatory authorities do also not monitor usage. Consequently, veterinary drug residues are likely to be present in food of animal origin, which poses a potential hazard to human health (49). Failure to observe withdrawal periods in animal has been reported in Tanzania, Nigeria, Ethiopia, Sudan and Malaysia that it was among the reason for antimicrobial residues in food of animal origins and propagates development of antimicrobial resistance (10,12,39,(50)(51)(52).
Additionally, our survey found different types and brands of antimicrobial used were from different sources that are not monitored and controlled. In Tanzania, the quality and quantity of veterinary antimicrobials are di cult to assess. This nding is comparable with the one reported by (31,53) that there is variation in quality of veterinary antimicrobials which is in tandem with insu cient system for monitoring of antimicrobial use in animal production. While the Tanzania Food and Drug Authority (TFDA) perform quality assessments, on imported drugs at the port of entry, there is very weak post-market surveillance on veterinary medicines. We found the government does not control antibiotics included in animal feed and that informal feed manufacturers that are found in small kiosks do not regulate the quantity of antibiotics included, leading to unnecessary exposure to antibiotics. During focus group discussions it became apparent that frequently drug importers, distributors and wholesalers supply drugs direct to consumers.
With regard to IPC, we observed poor housing with unhygienic conditions and limited air circulation in poultry houses. This most likely facilitated occurrence of most of the reported diseases by the farmers such as respiratory infections and typhoid in poultry and intestinal worms and skin diseases in domestic pigs, prompting increased use of antimicrobial agents. Poor housing conditions are in contrary to the Animal welfare act 2008 (54) that requires animals to be kept in approved structures that conforms to the quality hygienic and management practices. Collectively our ndings suggest the acute need for the availability of livestock extension services (government or regulated private) at local administrative level for the farmers to seek advice. We envisage that livestock extension personnel will be much trusted and their consultation service will be cost-effective and public health relevant, as it will reduce the magnitude of AMU in animal farming and improve the safety of short-cycle stock derived food commodity. Furthermore, we recommend strengthening of the awareness and understanding of antimicrobial resistance by farmers and community level service providers (traders, extension personnel, and community based animal health workers) through effective communication, education and training through collaboration between government and Civil Society organizations (CSO).
With regard to environmental contamination, respondents identi ed a number of activities taking place in the Msimbazi basin as potential drivers of AMR. These include e uents from households, hospitals, abattoirs and pharmaceutical and commercial industries, use of pesticides, ooding, and emptying of sewage into the environment during rainy seasons. Despite this, most of these activities are still on-going; posing health risks to humans and livestock by causing infections that are di cult to treat (25). This is contrary to the Environmental Management Act, 2004 (55), which provide for legal and institutional framework for sustainable management of the environment, including impact and risk assessments, prevention and control of pollution, waste management, environmental quality standards, public participation and compliance.
We found out that the existing Veterinary Acts  (63) that aims at developing a sustainable, competitive, vibrant and more e cient commercialized sheries and aquaculture industry has no mention of AMU and AMR. We recommend that these acts and regulatory documents should be updated and work logistically to embraces a One Health approach, which is cost-effective strategy for curbing AMR. The need for a cross agency and cross-disciplinary collaborations has been suggested in some studies in Tanzania (14,64) for the purpose of optimizing AMU, control quality, distribution, handling and awareness creation in human and animals and other related sectors.
On a positive note the government of Tanzania has developed a National Action Plan for Health Security 2017-2021 that aim to create and maintain active collaboration between the sectors for addressing health security using "One health approach concept" so as to ensure that there is timely preparedness, and a consistent and coordinated response in the event of occurrence of an event of public health concern. The plan is implemented under the guidance of the Prime Minister's O ce in order to achieve an Inter-ministerial Committee to administer the plan, and monitor and evaluate its implementation from all relevant line ministries. At the same time the government has also developed a National action plan on antimicrobial resistance (2017-2022) (65), which adopted the One Health Approach, and has strategies that are related to monitoring and surveillance of AMR and antimicrobial consumption in human and animal, improving antibiotic stewardship and control the spread of AMR in both clinical and farm settings, and increase knowledge and public awareness on AMR and establish national governance for inter-sectoral actions.
Fortunately, there are a number of research and OH AMU and AMR projects that are currently being implemented in Tanzania including i) Supporting the National Action Plan on AMR in Tanzania (SNAP-AMR) that assesses prescribing practices in different health care settings as well as investigating community access and attitudes to antibiotics among householders, community 'drug' shops and unregulated sources such as roadside traders and examine use in livestock by individuals with different levels of knowledge about AMR, e.g. district vets, community livestock o cers and livestock holders ii) Fleming Fund Country Grant that aims to strengthen Tanzania's national Antimicrobial Resistance (AMR) surveillance strategy by addressing the gaps in AMR data and strengthening antimicrobial stewardship. Ensuring that veterinary laboratories have access to a cadre of expert trainers tools, methods for analysis and interpretation and propose minimum data sets. Existing tools such as AfyaData can be adapted to collect AMR data and standardized protocols will allow for integration of data and comparative analysis among countries. These projects provide a platform for implementation of the national action plan on addressing AMU and AMR.

Conclusion
This study found a high usage of veterinary antimicrobials primarily for prophylactic purposes among poultry and domestic pig farming communities. Most farmers have inadequate knowledge on IPC and antimicrobial use and have limited access to veterinary and extension services, paving way self-treatment and opportunism by pro t driven non-professional veterinary drug sellers. The existing veterinary legal framework is weak and is hardly implemented due to a number of reasons including inadequate veterinary and extension services. Our study proposes implementation OH interventions that focus on optimizing antimicrobial use in animals and humans and measures to minimize environmental contamination to minimize occurrence of infections and use of antibiotics and promote health and productivity to realise the sustainable development goals (SDGs). Additionally, the study calls for more research to evaluate the active ingredients of antimicrobials consumed in speci c animal species in the study area and across the country in order to get the true re ection of the magnitude on AMU in food animals.

Declarations Ethics approval and consent to participate
The permit to perform this study was approved by the Muhimbili University of Health and Allied Sciences, Local Government Authority along the Msimbazi river basin and the National Institute for Medical Research Ethical Clearance board. In addition, participants signed written informed consent for participation in the study.

Consent for publication
Not applicable.

Availability of data and material
Questionnaire, FGDs and In-depth interview guiding questions used to generate information for the study is attached (supplementary le).

Competing interests
Not applicable as there have no competing interests.

Funding
The study was supported by the grants from the Government of Tanzania in collaboration with the World Bank to the SACIDS foundation for One Health, Sokoine University of Agriculture, Tanzania.
Authors' contributions Zuhura I. Kimera: Designed the study, Collected data and performed the analysis, generated the manuscript.