Antimicrobial resistance at the human–animal interface in the Pastoralist Communities of Kasese District, South Western Uganda

Intensive usage of antimicrobials in the management of animal diseases leads to selection for resistance among microorganisms. This study aimed to assess antimicrobial use and to describe factors associated with the transmission of antimicrobial resistance between humans and animals in pastoralist communities of Kasese district. A mixed-methods approach was employed in this study. Rectal swabs were collected from the participants and cattle and transported in Carry–Blaire transport medium to the laboratory within 24 h of collection for culture and sensitivity to confirm carriage of multi-drug resistant bacteria. In-depth interviews were conducted among veterinary officers, veterinary drug vendors, human health facility in-charges in both public and private health facilities, and operators of human pharmacies and drug shops. Carriage of multi-drug resistant bacteria among humans was 88 (93%) and 76(80%) among cattle. Consumption of lakeshore water and carriage of multi-drug resistant bacteria in cattle were associated with carriage of multi-drug resistant bacteria in the human population. The prevalence of multi-drug resistance among organisms Isolated from both humans and animals was high. There is a high likelihood of transmission of multi-drug resistance between humans and animals.

the study design. A mixed-methods approach was employed in this study between January 2018 and March 2019. A total of 122 participants who previously tested positive for multi-drug resistant E. coli were followed up to the community (Fig. 1). In the community, questionnaires were administered to the traced patients/ parents/guardians who consented to take part in the study to assess antibiotic consumption both in animals and humans, self-medication, and other variables as per the conceptual framework. Presence or absence of multidrug-resistant isolates were our outcome variables while the predictor variables were cattle keeping, handling of cattle and their products, consumption of animal products, consumption of antimicrobial agents by both animals and humans, sanitation and social status, history of hospitalization in the household, education level, occupation, location of the household from the clinic and socio-demographic characteristics such as age and sex. Rectal swabs were then collected from two randomly selected members of the household, and two randomly selected animals (cattle) from each homestead using sterile cotton swabs for culture and sensitivity. The collected specimens were transported in the Carry-Blair transport medium to the Microbiology laboratory of Makerere University School of Biomedical Sciences for analysis. Speciation and antibiotic susceptibility of the isolates was done using the Phoenix automated microbiology system (Phoenix 100 ID/DST system) from Becton and Dickson (Franklin Lakes, NJ, USA) and the results interpreted using the CLSI guidelines.
In-depth interviews were conducted among the veterinary medicine practitioners including; Veterinary Officers and veterinary drug vendors, human health facility in-charges in both public and private health facilities and operators of human pharmacies and drug shops who were purposively selected due to their role in the health facility/pharmacy/drug shop and administrative positions occupied that directly affect the use of antimicrobials. The research assistants were trained to collect quality information at its depth. For convenience, appointments were scheduled with the prospective interviewees who met our selection criteria at the time which was most convenient to them and each interview lasted between 1 and 2 h. All the respondents were interviewed at their places of work to minimize the disruption of their normal routines. Before the commencement of the interview, the interviewer explained the aims and scope of the study and obtained written informed consent from every participant who accepted to take part in the study. Face-to-face interviews were conducted with the aid of an interview guide that was prepared by the investigator, reviewed by co-investigators, and refined as the final tool for data collection. All the 12 interviews were audio-recorded and later transcribed into manuscripts by the field research assistants for analysis.
Data management and analysis. Data were coded and entered into Epi-Data version 3.1 software and later exported to STATA V14 for analysis. Categorical data were presented as proportions and their associations Scientific RepoRtS | (2020) 10:14737 | https://doi.org/10.1038/s41598-020-70517-w www.nature.com/scientificreports/ determined by the Chi-square test. Bivariate analysis was performed to explore associations between the isolation of multi-drug resistant bacteria and associated factors thereof. Variables were included in the multivariate analysis basing on factors in the bivariate analyses that had either p ≤ 0.2 and/or other variables such as the history of chronic illness known from the literature to be associated with the risk of acquiring multidrug resistance. In all cases, P ≤ 0.05 was considered as evidence of significant statistical association. Data from the in-depth interviews were transcribed verbatim in the English language by the trained research assistants and thematic analysis as described before 11 was used to inform data analysis. For quality control, transcription was done by the first research assistant and the second research assistant listened to the audio recording while reading the transcripts to recheck on the quality of the transcripts. Data analysis was done using the NVivo V12 qualitative data analysis software (QSR International (Americas) Inc). Emergent themes were identified, coded, and organized into concepts that were later developed into explanations. Data analysis involved a process of familiarization with the data by listening to the audio recordings and reading the transcript several times while noting ideas. The principle investigator performed the thematic content analysis and the co-investigators verified the themes and contents. Significant statements relating to antimicrobial resistance were identified and coded. The codes were clustered and themes were generated. Emergent themes generated were then discussed with coinvestigators. The process of theme generation was reviewed and refined by going back and forth between the themes and codes, as well as between the themes and transcripts until the final themes were identified, defined and named. The final themes were discussed with some of the respondents in the process of member checking. All methods were performed according to the relevant guidelines and regulations. ethical considerations. The study was approved by the Makerere University School of Biomedical Sciences Higher Degrees and Ethics Committee (SBS-HDREC) and The Uganda National Council of Science and Technology (UNCST). Written informed consent was obtained from all respondents. The picture in Fig. 2 was taken from the backside after obtaining consent to ensure confidentiality.
Approval to use animals in this study was obtained from the Uganda Wildlife Authority (UWA) and written informed consent to collect, store and use samples from animals/cattle for research was obtained from the owners of the animals. All methods used in this study were following the relevant standard guidelines and regulations.

Results
Out of the 122 respondents who fulfilled our selection criteria for follow up to the community, 103 respondents were successfully traced, 19 were lost to follow, 95 consented to participate in the study, and 8 declined to participate with no specific reasons. The probable reason for a high number of the lost to follow was that at the time of follow up coincided with the cultivation season and many people had moved to the low lands to cultivate.
The mean age of the respondents was 35 years, the youngest participant was 18 years whereas the oldest was 66 years. Females constituted slightly over half, 55 (58%) of the total number of respondents.
The common livestock kept were cattle, goats, sheep, and pigs. The local breeds of livestock were generally preferred over the crossbreeds because they are resistant to diseases and can survive well under the local weather conditions. Crossbreeds are gaining acceptance among some pastoralists because they are high performers in terms of production. Cross-breeds grow very fast and they can be marketed before encountering a lot of expense of feeding because they are fast growers and the farmer can get quick returns. Pastoralists reported that that cross-breeds need a high level of management in terms of disease control yet drugs are very expensive.
Knowledge, opinions, and perceptions on antimicrobial resistance. Respondents were requested to respond with 'True' , 'False' , or 'I don't know' to a couple of questions to assess their knowledge, opinions, and perceptions on antimicrobial resistance. The majority of the respondents in this study 74 (78%) could not correctly define antimicrobial resistance. A high proportion of the respondents, 74 (78%) agreed with the statement that "antimicrobial resistance occurs when your body becomes resistant to antimicrobials and they no longer work well". Many respondents also agreed that "antimicrobial resistance is an issue that could affect them or their families" 68 (72%), "can be transmitted from person to person, and very challenging to treat" 75 (79%) ( Table 1).
The respondents agreed with several actions that would address the problem of antimicrobial resistance. Most of the respondents agreed that people should use antimicrobials only when they are prescribed by a doctor or nurse, 90(95%), parents should ensure all of their children's vaccinations are up-to-date, 87 (93%) and doctors should only prescribe antimicrobials when they are needed 92 (97%). All respondents agreed that regular hand washing was key in minimizing the problem of antimicrobial resistance.
There were also several misconceptions among the respondents about the use of antibiotics. Over 50% reported that antibiotics can be used to treat malaria, respiratory tract infections 76 (80%) sore throat 79 (83%), and common cold 87 (92%).
Different socio-demographic factors such as education level, having heard about antimicrobial resistance, sources of drinking water, a recent visit to a health facility, level of health care facility visited, History of medical procedure, chronic conditions, over the counter access to antimicrobials, having had a procedure involving the use of Internal devices, frequency of antimicrobial use, multi-drug resistance in bacteria isolated from Cattle www.nature.com/scientificreports/ were analyzed using bivariate and multivariate analysis to determine their association with multi-drug resistance carriage among the isolated bacteria from humans ( Table 2). Drinking lakeshore water and carriage of MDR bacteria in cattle were significantly associated with carriage of multi-drug resistant bacteria in humans (P < 0.05). None of the other socio-demographic characteristics was significantly associated with carriage of multi-drug resistant bacteria in humans. Two hundred and ninety-six (296) bacteria isolated from both humans and animals were tested for resistance to a panel of 15 antibiotics belonging to different classes. The bacteria isolated included; Escherichia coli, Klebsiella pneumoniae, Klebsiella oxitoca, Serratia fonticola, Kluyvera ascorbate, Enterobacter cloacae, and Enterobacter aerogenes. The antibiotics used were selected from the penicillins, β-lactamase inhibitors, 1st, 2nd, 3rd, and 4th generation cephalosporins, quinolones/fluoroquinolones, aminoglycosides, tetracyclines, nitrofurantoins, carbapenems, and folate pathway inhibitors. Generally, the prevalence of antibiotic resistance among the isolated organisms was high, with a number showing multi-drug resistance.
The E. coli isolated from cattle exhibited high resistance to cefazolin, ampicillin, and amoxicillin-clavulanic acid (96, 80, 67) %, respectively whereas low resistance to ciprofloxacin levofloxacin and imipenem was noted  Table 3). None of the E. coli isolated from cattle were resistant to gentamicin. Klebsiella pneumoniae of cattle origin was mostly resistant to cefazolin, amoxicillin-clavulanic acid, and cefuroxime (91, 82, 64) % respectively but showed low resistance to ertapenem, imipenem, ciprofloxacin, and levofloxacin. None of the Klebsiella pneumoniae isolates from cattle was resistant to gentamicin. All other organisms (Table 2) isolated from cattle were resistant to ampicillin and highly resistant to cefazolin, cefuroxime, and amoxicillin-clavulanic acid (92, 86, 79) % respectively. None of the other isolated bacteria was resistant to ciprofloxacin and levofloxacin and most

in-depth interviews
Demographic characteristics of the study respondents. A total of 12 in-depth interviews were conducted with key informants age between 24 and 50 years from different professional backgrounds handling antibiotics. All our respondents consented to be audio-recorded. The educational background of the respondents ranged from high school certificates to Master's degrees. The qualifications included a Masters in veterinary medicine, Masters of Science in Nursing, Diploma in Nursing, Certificate in nursing, Nursing Assistants who were high school certificate holders trained on the job to do nursing work, Bachelor of Business administration, Bachelors of Community development and Advanced certificate of education. We did not find any participant with a Bachelor of Medicine for our interview because most of the community rural pastoral community health infrastructure was left to the lower qualified staff. Even when one of the clinics in the community was owned by a qualified doctor, it was run by a nursing assistant who he trained on the job.

themes.
A total of six key themes relating to key informant views on antimicrobial resistance were identified using thematic analysis. In Table 3, we summarized all the six key themes before writing the theme in depth.

Key informant opinions and perceptions on Antimicrobial resistance.
Respondents reported several livestock diseases affecting humans and pastoralist communities in Kasese district for which antibiotics were commonly used. All the four key informants dealing directly in veterinary medicine (who we call veterinary key informants for the sake of this study) reported that the four most common livestock diseases were mainly tickborne diseases which include; East Coast Fever (ECF), anaplasmosis, heartwater, and babesia. Even when some of these diseases were viral, antibiotics were used as supportive treatment.
What the community has known is that these antibiotics are broad-spectrum. So when an animal shows any sign of disease, the mind goes back, runs very fast to give antibiotics without knowing whether the antibiotic will be effective or not and what disease it is. The first aid is an antibiotic; that is the practice they have developed over time (KII veterinary pharmacist, Kasese town).
One of the respondents also reported occasional outbreaks of other notifiable diseases that kill a lot of animals.   www.nature.com/scientificreports/ On the human medicine side, seven out of eight respondents mentioned malaria, bacterial infection, urinary tract infections, respiratory tract infections, and typhoid as the most common diseases treated with antibiotics.
"Mostly we have people with a cough. Everybody is coughing, everybody has stomach pain. These we sell to them. at least the antibiotics are helping us to manage the problem". "If you refuse to sell to them, they leave your place and buy from another place" (KII Nurse Katwe-Kabatooro). common drugs used in humans and animals. Respondents reported several livestock drugs commonly used among pastoralists in Kasese district. Oxytetracycline with different trade names in different concentrations was reported to be the main antibiotic used by the pastoralist followed by penicillin/streptomycin combination. Sulphadimidine was reported to be used usually when the animal had diarrhea. Respondents also reported common use of de-wormers like albendazole, and liver mysole in different trade names and a few pastoralists had started using ivermectin due to its double effect on worms and ticks though it was perceived to be a very expensive drug.
Acaricides were reported by all the veterinary key informants to be on very high demand. Four classes of acaricides commonly used by pastoralists include amidines, synthetic pyrethroids, the combination amidines, and pyrethroids, the organophosphorus acaricide, however, most pastoralists unknowingly use the same class of acaricide for quite a long time, changing just the brand names (Table 4).
Pastoralists don't know they are using the same drug so eventually, they begin complaining that the drugs are not working. They say that the drugs are fake (bicupuli) when the problem is lack of knowledge of whatever they are using (KII Veterinary officer, Kasese district).
On the human medicine side, many antibiotics were mentioned as commonly used in the pastoralist communities (Table 5). Though a long list of antibiotics was reported to be in common use in this community, the people selling or prescribing them were mostly of low medical cadre with a number of the key informants (custodians of antibiotics) practically struggling to pronounce the names of the antibiotics.

Awareness of antimicrobial resistance.
Pastoralists were reported to have limited awareness of antimicrobial resistance and they perceived that treatment failures in animals were due to rampant counterfeit medicines on the market. Most of the veterinary drug shop operators reported being aware of the problem though many could not correctly define antimicrobial resistance.
There was limited awareness about antimicrobial resistance among operators of human medicine clinics and drug shops. Generally, the community bought antibiotics and if they felt the antibiotic did not work they, would go and change to a stronger one, however, educated members of the community were reported to prefer obtaining a prescription from the hospital and then go and buy from the local drug shops and pharmacies since there were often stock outs of drugs in the hospitals.
Those people who are learned don't want antibiotics, when it comes to antibiotics, they go to the hospital for a prescription. The information I have received from my patients is that when they start on antibiotics and they take them for a long time, they become addicts. Next time they buy them and take, they will not be Table 4. Inter-host resistance comparison for E. coli isolated from humans and cattle. There was no significant difference in resistance to particular drugs between the human and cattle isolates except for Cotrimoxazole (P < 0.05).

Drugs
Human (n = 133), N (%) Cattle (n = 84), N (%) P value One of the respondents interviewed was a masters' degree nurse who stated clearly that awareness of antimicrobial resistance was not a concern for nursing as a profession.

You would get that information clearly from the clinician, mostly this would be for the clinicians but I am not a clinician. I am a nurse and not a nurse working in health facilities so you might not get the proper information (KII nursing officer Kasese district).
Drivers of antibiotic resistance. Several drivers of antimicrobial resistance were identified by the respondents (Table 5) most of who associated the problem to poverty, illiteracy, lack of veterinary personnel, and ignorance among the community and the low-qualified medical staff serving the community. One operator of a human medicine clinic reported frequently giving antibiotics as first aid because people had no money to buy full doses.
The drugs are expensive so when the patient says he has no money, you give him an injection for the little money he has, but it is not enough to complete the dose and you have nothing to do for him, you just give for first aid (KII nursing assistant, Nyakatonzi) . The first problem is the patients, who abandon their medication after feeling well, especially in the middle of a dosage. Secondly, the stock-outs of drugs from facilities which makes patients take underdose. Clinical treatment is also a cause of this problem but at a very small scale because chances are equally high that the right medication would be offered to the patients but the practice of sharing one dose of medication among family members is also common (KII nursing officer, Katwe-Kabatooro). www.nature.com/scientificreports/ Three out of eight human medicine practitioners reported that some families shared a drug dose among family members to treat the different infections which are not sensitive to these drugs, hence increasing the problem at hand. Also, patients reportedly did not follow treatment schedules.

The patients use these drugs badly. You tell the person to come in the evening for an injection, but he comes early in the morning. You tell him again to come in the evening he still comes in the morning. They do not keep time. That is the problem I have seen and it can cause resistance (KII Clinical officer, Hima sub-county).
Frequent stock out of medicines in the government facilities was reported to be a common problem which the community adapted to by sending children to the public health facilities with false complaints to get several doses of antibiotics and antimalarial to store at home for future emergencies.
Pastoralists were reported to prefer treating their animals because veterinary doctors were few and very expensive.
one of the reasons why this problem is coming up, pastoralists don't seek prescription. They do it themselves. So whatever they land on in the drug shops is what they go and use. And even the choice depends on the money they have. So the availability of money also has played a role in this. Then, the treatment done by the pastoralists depends on what the farmer feels. If the farmer feels the animal is fine, he stops medication without completing the dose and stores the rest of the medicine for future use. Drugs were mostly stored inappropriately in the ceilings, metallic boxes, and others place the pastoralists deem fit (KII Veterinary officer Kasese district).
Polypharmacy was commonly practiced in community pharmacies, clinics, and drug shops due to lack of laboratories in the community health facilities, and antimicrobial resistance was only suspected when the patients returned to the health facility without improvement.
One of the respondents cited political interference as a driver for antimicrobial resistance. Surveillance and monitoring. All the 12 respondents reported that there were no interventions in place to minimize the problem of antimicrobial resistance both in humans and in animals. Drug inspectors often visited drug shops and clinics but their main concern was cleanliness and smart display/arrangement of drugs. Most operators of drug shops admitted to selling antibiotics and the drug inspectors were aware but only advised them not to display antibiotics on the shelves.

I do not see any intervention because things are being done the way they were being done. Pastoralists still go and do whatever they want to do. I have not seen any intervention in livestock to fight antimicrobial resistance (KII veterinary shop owner, Kasese town).
Like for Septrin and metro; last time they told us that when we sell septrin and Flagyl, we shouldn't display them on the shelves (KII drug shop attendant, Kinyamaseke ward).
proposed intervention. Some interventions were proposed by the respondents. All the respondents proposed the sensitization of the community as a key intervention and all people who handle antibiotics and the general public should be sensitized. We have Class C drug shops; they don't bother telling us about antibiotic resistance. The inspectors mind about cleanliness and the arrangement of drugs. They told us we can use antibiotics but hide them and not display them on the shelves (KII drug shop attendant, Nyakotonzi).
Some proposed closing down clinics and drug shops operated by unqualified personnel while others proposed that the government should initiate and strengthen surveillance and monitoring of antibiotics at all levels.
They don't give them the right dose, even you know how Uganda operates, those drug shops you see, the people there are not real medical people. Those who are selling the drugs may be trained on the job. They just give antibiotics even when they give it in the wrong doses. If somebody was to take it for five days, they will give for two days (KII Clinical officer Hima).
All the respondents cited awareness campaigns among community health workers and the general population as an important component of the prevention and surveillance strategy for antimicrobial resistance coupled with equipping health facilities with the necessary diagnostic equipment.
Scientific RepoRtS | (2020) 10:14737 | https://doi.org/10.1038/s41598-020-70517-w www.nature.com/scientificreports/ I think they should improve the method of telling us, the drug dealers and sellers in drug shops, that we should be referring these people to hospitals because we are the first people these patients come to before they go to health centers. So we are the ones to tell them. So they should tell the drug sellers to be referring people to hospital (KII drug shop attendant, Kinyamaseke ward). I think after sensitization, facilities should be well equipped such that when people visit these health facilities, they will find the equipment available. But if the sensitization is not followed by the adequate equipment of these facilities, it will all be for nothing however much this information would be very well disseminated or even printed on posters. Besides, if this same information disseminated in the sensitization sessions is extended to radio stations, it would help to ascertain the seriousness of this information (KII nurse Katwe-Kabatooro).

Discussion
There was generally low knowledge and awareness on antimicrobial resistance in the pastoralist community of Kasese district and this may be one of the most important contributors to inappropriate use of antibiotics in this community leading to the emergence of multi-drug resistant bacteria. A majority of the community members 74 (78) did not understand the meaning of antimicrobial resistance. A misconception that antimicrobial resistance occurs when your body becomes resistant to antimicrobials and they no longer work as well, was common and this could be as a result of low literacy levels in this community. A majority of our respondents 59 (62%) either had no formal education or ended in primary school and only 6% of the people interviewed had attained a tertiary level of education. Though there was no statistically significant relationship between education level and knowledge/awareness of antimicrobial resistance in our study, in other studies, knowledge, and awareness of antimicrobial resistance was attributed to lack of formal education 12,13 . Awareness campaigns on antimicrobial resistance through effective education and communication coupled with the inclusion of the wider community are important and they constitute the major objective of the Global Action Plan on Antimicrobial Resistance 14 .
Many key informant respondents, 9 out of 12 agreed that antimicrobial resistance was an important issue affecting the pastoralist communities. All the key informant respondents reported that the practice of over the counter purchase of antibiotics was common and attributed this to the challenge in accessing qualified prescribers both for veterinary drugs and human medicine. Whereas some of the clinics in the community were owned by qualified medical doctors, they preferred to employ low cadre medical personnel to run them. The reason for this is that employing a highly trained person required a sound financial position. Previous literature has reported the crucial role of sound financial position to maintain a high quality of service 15 . Most of the respondents obtained drugs from the community drug shops and private clinics without consultation and mostly in sub-optimal doses. Respondents mostly depended on their experience, and advice from friends, family, or neighbors similar to the findings of another study that found that people preferred obtaining antibiotics and taking them basing on their knowledge or guidance from friends and family 16 . Improving the availability of qualified prescribers through deploying more qualified staff in the community health facilities can increase the number of appropriate prescriptions of antimicrobials obtained by the community members. In Kasese district, like other districts in Uganda 17 , only one veterinary officer who was mostly taken up with administrative roles and hardly reached by pastoralists due to his busy schedule was employed. Employing more veterinary officers in Kasese district can improve their availability to the service of the pastoralist communities and hence improve on judicious use of antimicrobials directly by treating animals and indirectly through training and sensitization of the pastoralists.
There were no guidelines found in the dug shops and clinics visited and most prescribers were using their knowledge. This was also observed by a similar study that found that prescribers in community-based practices were less supported than their hospital counterparts and they end up operating in isolation 18 . Because of the assumption that the risk of generating resistance is greater in the hospital environments that are heavily exposed to antimicrobials, most guidelines are tailored to hospitals 18 . Antimicrobial resistance in the community has been reported in several studies around the world 19-21 yet interventions and policies continue to focus on antimicrobial resistance in the hospital environments. If these efforts are not redirected to the community, the consumption of non-prescribed antibiotics will continue rising. Whereas the adverse effects of non-prescribed antibiotics are rarely reported, they are likely to be as common as the adverse effects of the prescribed antibiotics. Patient safety issues have rarely been of concern even when inappropriate dosing may result in severe allergic reactions, masking the diagnosis of infectious diseases or death 19 .
Though most of the respondents agreed that people should use antimicrobials only when they are prescribed by a doctor or nurse, there were mostly less trained cadres of nurses and nursing assistants operating drug shops and clinics in this community and one needs to get to the hospital and queue in long lines to get a physician's prescription. A similar challenge of a high prescriber to patient ratio was observed in a study in Nepal 22 leaving the less trained cadre as the only alternative source of antibiotic prescription. This presents a dilemma of choosing between judicious use and universal access to effective antibiotics. Other studies have noted that universal access to effective antibiotics is essential for tackling the problem of antimicrobial resistance 23,24 and millions of people die due to lack of access to effective antimicrobials. The reduction in mortality afforded by effective antibiotics in modern medical practice ranging from cancer therapy, neonatology, intensive care medicine to aggressive surgeries, and organ transplantation procedures are quite enormous 25 but may soon be short-lived. The policies and guidelines to tackle antimicrobial resistance need to put the underserved communities high. The non-prescribed use of antimicrobials in the community has continued to attract less attention leaving the burden of antimicrobial resistance in the hands of those taking care of their critically ill loved ones in the emergency rooms. The resistance developed due to the abuse of particular antibiotics today will affect the treatment of another individual tomorrow 26 .
Vaccination was generally considered an effective means of tackling antimicrobial resistance by the community and all respondents agreed that regular hand washing was key in minimizing the problem of antimicrobial Scientific RepoRtS | (2020) 10:14737 | https://doi.org/10.1038/s41598-020-70517-w www.nature.com/scientificreports/ resistance. Our findings agree with previous studies that report that simple infection prevention strategies including hand hygiene are effective in breaking the chain of infection and controlling the horizontal spread of antimicrobial-resistant organisms [27][28][29] . Other studies have also supported the fact that vaccination is an effective means of reducing the consumption of antimicrobials and controlling antimicrobial resistance [30][31][32] . Vaccination against bacterial and viral infections reduces morbidity due to the related infections thereby reducing the prescribing of antibiotics and controlling the development of resistance of bacteria to antibiotics. Drinking lakeshore water and carriage of MDR bacteria in cattle were significantly associated with carriage of multi-drug resistant bacteria in humans (P < 0.05). Lake George and the Kazinga channel play a central role in the daily routine of the pastoralist communities in Kasese district. Apart from being a source of drinking water for the cattle, it is a source of domestic drinking water and other uses including recreational purposes for children and adults (Fig. 2). Water is, therefore, a shared resource among the wild, domestic animals, and humans which play a key role in the transmission of antimicrobial resistance at a human-animal interface. Other studies have also similarly observed that shared environments such as water and soil play a vital role in the transmission of antimicrobial resistance 33,34 . Though the causal pathway is difficult to understand, several studies have noted that antimicrobial resistance can be transmitted from animals to humans 26,35 . Resistant bacteria are constantly shed by the cattle into the environment and surface water, directly contaminating caretaker who in Kasese district are often children. Resistant bacteria can then be transferred through several routes to different environments, humans, and animals 26 . Increased use of antibiotics on the farm is reported to drive the evolution of antimicrobial resistance in animals and bacteria confer similar mechanisms of resistance to the commonly used antibiotics in humans 34,36 . Control of antimicrobial resistance, therefore, requires joint efforts from all stakeholders from human health, animal health, agriculture, aquaculture, finance, environment, and well-informed consumers.
Multi-drug resistance in gut bacteria is a growing global public health concern. In this study, antibiotic resistance testing revealed high resistance patterns against cefazolin, ampicillin, cotrimoxazole, and amoxicillin-clavulanic acid among the human isolates of E. coli (98, 85, 72, 60) %, respectively. Similar findings were documented by another study 37 in Korea which also associated resistance of E. coli strains to ciprofloxacin with resistance to ampicillin, cotrimoxazole, and cefazolin. However, gentamicin, ciprofloxacin, levofloxacin, and imipenem showed low resistance patterns (2, 5, 5, 5) %, respectively similar to other studies 38,39 .
Escherichia coli isolated from cattle exhibited high resistance to cefazolin, ampicillin, and amoxicillin-clavulanic acid (96, 80, 67) %, respectively. Klebsiella pneumoniae from cattle origin was mostly resistant to cefazolin, amoxicillin-clavulanic acid, and cefuroxime (91, 82, 64) %, respectively. High resistance to cephalosporins among the E. coli and K. pneumoniae isolated from cattle has been reported in several studies in Africa and elsewhere [43][44][45][46] . Ampicillin and amoxicillin resistance has also been commonly observed in several studies 43,[47][48][49] . Low resistance to ciprofloxacin, levofloxacin, and imipenem among the E. coli, K. pneumoniae, and other isolates was observed in this study. Fluoroquinolones and carbapenems are antibiotics of critical importance that are still showing low or no resistance in some studies in Africa and elsewhere 50-52 but rising resistance is becoming a big concern. The low resistance of bovine gram-negative bacteria against imipenem could be because the carbapenems have not been approved for use in animals so use in food animals is likely to be low and therefore minimal direct selection pressure 51 .
None of the E. coli and K. pneumoniae isolated from cattle was resistant to gentamicin. Similarly, low rates of resistance (0.89%) of E. coli isolated from dairy cattle to gentamicin were reported by a study in Zambia 53 and less than 10% of the isolates from the dairy cattle from farms in Jordan were resistant to gentamicin 54 . On the contrary, higher resistance to gentamicin was reported from the E. coli isolated from milk in Bengal-India 43 . The varied prevalence of resistance against gentamicin from different geographical regions could be explained by different antibiotic use patterns in these regions.
Other gram-negative bacteria isolated from cattle were all sensitive to ciprofloxacin and levofloxacin indicating that these antibiotics are still useful in the management of bacterial infections in cattle. We detected a high prevalence of multi-drug resistance among cattle isolates 97 (81%) in this study; higher than the findings of another study in China that reported, 40% prevalence of multi-drug resistance from the cattle isolates and 54% in Germany 55,56 . This difference could be attributed to the difference in antibiotic use patterns in the different geographical locations and the recruitment procedure. In the China study, calves entering the fattening sage that might have not had a substantial exposure to antibiotic therapy were sampled whereas, in our study, we mostly sampled already fattened cattle. Overall, 65 (40%) of the bacteria isolated from humans were ESBL producers and 22 (12.5%) were carbapenemase producers and 44 (40%) of the cattle isolates were ESBL producers and 18 (15%) were Carbapenemase producers. This study also noted that there was no statistically significant difference in resistance to a particular drug except for cotrimoxazole in human and animal isolates of E. coli indicating a possibility of cross transmissibility (Table 4). ESBL and Carbapenemase production seems to be the primary cause of MDR in gut bacteria. Similarly, other studies have observed a high likelihood of ESBL carriage to contribute to MDR among gut bacteria 57,58 . Surveillance of ESBL and carbapenemase production could be one major and useful step in the control of the spread of MDR bacteria in both humans and animals.  The prevalence of multi-drug resistance among organisms isolated from both humans and animals was high. Consumption of lakeshore water and carriage of multi-drug resistant bacteria in cattle were associated with carriage of multi-drug resistant bacteria in the human population which suggests a possibility of transmission of multi-drug resistant bacteria between humans and animals.

Recommendations.
To attain optimal health for humans, animals, and the environment, the Global and the National Action Plans need to re-direct the efforts to reduce the burden of antimicrobial resistance from the hospital environment to the community. Investment in improving health care delivery systems in the pastoralist communities will improve the health in this population, reduce hospital admissions, and the number of infections that require antibiotics. Public information campaigns are important to improve the knowledge of pastoralist communities on the dangers of non-prescribed use of antimicrobials and the burden of antimicrobial resistance.

Data availability
All data on which the conclusions of this manuscript are drawn is available on request from the corresponding author.