Am. J. Trop. Med. Hyg., 100(6), 2019, pp.

1378–1390
doi:10.4269/ajtmh.18-0171
Copyright © 2019 by The American Society of Tropical Medicine and Hygiene

Rural South African Community Perceptions of Antibiotic Access and Use: Qualitative Evidence
from a Health and Demographic Surveillance System Site
Jocelyn Anstey Watkins,1* Fezile Wagner,2 Francesc Xavier Gómez-Olivé,2 Heiman Wertheim,3,4,5 Osman Sankoh,6,7,8
and John Kinsman9,10
1
Division of Health Sciences, Warwick Medical School, The University of Warwick, Coventry, United Kingdom; 2Medical Research Council, Wits
Rural Public Health and Health Transitions Research Unit (Agincourt), School of Public Health, Faculty of Health Sciences, University of the
Witwatersrand, Johannesburg, South Africa; 3Wellcome Trust Major Overseas Programme, Oxford University Clinical Research Unit, Hanoi,
Vietnam; 4Nuffield Department of Medicine, Centre for Tropical Medicine, University of Oxford, Oxford, United Kingdom; 5Department of Medical
Microbiology, Radboudumc Center for Infectious Diseases, Nijmegen, The Netherlands; 6Faculty of Health Sciences, School of Public Health,
University of the Witwatersrand, Johannesburg, South Africa; 7International Network for the Demographic Evaluation of Populations and their
Health (INDEPTH) Network, Accra, Ghana; 8Statistics Sierra Leone, Freetown, Sierra Leone; 9Faculty of Medicine, Department of Public Health and
Clinical Medicine, Epidemiology and Global Health (Umeå Centre for Global Health Research), Umeå University, Umeå, Sweden; 10Department of
Public Health Sciences, Global Health (Division of International Health - IHCAR), Karolinska Institutet, Stockholm, Sweden

Abstract. Knowledge and practices of rural South African populations with regard to antibiotic access and use
(ABACUS) remain understudied. By using the case of four villages in the north east of the country, our aim was to
investigate popular notions and social practices related to antibiotics to inform patient-level social interventions for
appropriate antibiotic use. To achieve this, we investigated where community members (village residents) were accessing
and sourcing medication, and what they understood antibiotics and antibiotic resistance (ABR) to be. Embedded within
the multicountry ABACUS project, this qualitative study uses interviews and focus group discussions. A sample of 60
community members was recruited from the Agincourt Health and Demographic Surveillance System, situated in
Mpumalanga Province, from April to August, 2017. We used the five abilities of seek, reach, pay, perceive, and engage in
access to healthcare as proposed by Levesque’s “Access to Healthcare” framework. Respondents reported accessing
antibiotics prescribed from legal sources: by nurses at the government primary healthcare clinics or by private doctors
dispensed by private pharmacists. No account of the illegal purchasing of antibiotics was described. There was a mix of
people who finished their prescription according to the instructions and those who did not. Some people kept antibiotics
for future episodes of infection. The concept of “ABR” was understood by some community members when translated
into related Xitsonga words because of knowledge tuberculosis and HIV/AIDS treatment regimens. Our findings indicate
that regulation around the sale of antibiotics is enforced. Safer use of antibiotics and why resistance is necessary to
understand need to be instilled. Therefore, context-specific educational campaigns, drawing on people’s understandings
of antibiotics and informed by the experiences of other diseases, may be an important and deployable means of pro-
moting the safe use of antibiotics.

INTRODUCTION developments in the rise of AMR pathogens in Africa were

found for commonly prescribed antibiotics in a systematic
The problem of antibiotic resistance (ABR). Antibiotic review by Tadesse et al.2; nonetheless, some country-level
effectiveness has decreased in recent years because of data were unavailable. The authors called for researchers to
overuse and misuse, resulting in increasing bacterial re- investigate the status of AMR and to identify knowledge gaps
sistance and threatening modern medicine.1 Antibiotic use to design suitable local and global interventions. Research
relates to “ABR” and is a subset of “antimicrobial resistance” into access and use of antibiotics are presently limited in most
(AMR). Both are widely acknowledged as a public health SAA countries. South Africa has a relatively well-functioning
concern and global problem.2 Drug-resistant infectious health system with national surveillance systems routinely
“superbugs” are increasing,3–5 and global efforts to curtail the generating representative and robust data on antimicrobial
emergence of resistance are gaining priority.6 In low- and use in tertiary care12 but not in primary care, where most citi-
middle-income countries (LMICs), the burden of infectious zens access antibiotics from.
diseases is exacerbated by limited access to, and availability Social sciences literature on AMR and ABR. The impor-
and affordability of antibiotics to treat infections.7 Little is tance of social science research for tackling AMR and ABR is
known about the prevalence of ABR in sub-Saharan Africa gaining momentum, with multidisciplinary projects being
(SAA), despite its known risks.8 There have been calls for funded. Chandler et al.13 take an anthropological lens, con-
continent-wide surveillance to inform empirically based treat- sidering ABR as “a discourse, social practice and natural fact”
ment guidelines.9 The WHO has designed a “Policy Package to by describing varying works on the use of medicines by con-
Combat AMR”10 for governments to commit to a national plan; sumers and prescribers, and meanings of medicines for pa-
yet to date, few African countries have initiated it. tients and providers. Likewise, Wood14 draws on sociological
The Center for Disease Dynamics, Economics, and Policy perspectives, arguing that the responsibility of using antibi-
offers an online “Resistance Map”11 with search tool by otics responsibly is a societal problem, not only the “re-
country for ABR by either antibiotic or pathogen. General sponsibility of government health services.”
Several authors summarize key social factors involved in
antibiotic use by “agents”: providers, dispensers, and the
* Address correspondence to Jocelyn Anstey Watkins, Social Science
and Systems in Health, Division of Health Sciences, Warwick Medical
general population, especially within primary research exam-
School, Gibbet Hill, The University of Warwick, Coventry, CV4 7AL ples in developing countries.15 For example, Broom’s et al.16
United Kingdom. E-mail: [email protected] study argues that antibiotic misuse by doctors is “better

1378
 RURAL SOUTH AFRICAN COMMUNITY PERCEPTIONS OF ABACUS 1379

understood in terms of social relations.” Professional etiquette • provide a context for developing social interventions for
as a logic of practice can be resistant to change. “Antibiotics safe antibiotic use.
are often perceived as strong medicine, capable of curing al-
most any kind of disease,”15 with color, taste, and size as In this article, the definition describing the “appropriate or
factors in determining perceived efficacy by the consumer.17 inappropriate” use of antibiotics is based on the biomedical
van der Geest18,19 discussed factors relating to the illegal view: inappropriate use is likely to make the medicine less
distribution of biomedical medicine, particularly in Africa. They effective in terms of treating the disease in question, whereas
describe why lay people self-medicate and buy from drug appropriate use will likely make it more effective and safe.
peddlers due to long queues at overcrowded hospitals, not What is “appropriate” from a biomedical perspective need not
wanting others to know of their illness and weak drug control. be appropriate from a sociocultural or from the patient’s
In another article, they20 refer to biomedical pharmaceuticals perspective.
as the “charm of medicines,” “meaningful,”21 and tangible, Access to healthcare framework. Sen’s Capability
and as “time-saving commodities”22 with implications for Approach35,36 was used for the design of the ABACUS proj-
social relations. Over the decades, knowledge gained from ect. Yet, for this study, we decided to use another framework
these studies around social factors of medicine prescribing that directly focused on access to healthcare than on capa-
and consumption lay the backdrop for our study and situate bility and freedom.
this work’s qualitative contribution. Access to healthcare has been described as “the opportu-
South African health system context and ABR policy. nity to reach and obtain appropriate healthcare services
South Africa’s pluralistic healthcare system is based on both in situations of perceived need for care.”37 Many academics
traditional and biomedical healthcare.23 The public health system have defined access to healthcare38–40 with most conceptual
offers free primary healthcare to all citizens and hospital ex- frameworks including the three dimensions of acceptability,
emptions for those who qualify,24 resulting in high use availability, and affordability. There are varying access to
of government facilities.25 In Gauteng Province, 96 percent of healthcare models, and in this article, we use an analytical
people use these.26 Research has documented patient reports of framework by Levesque et al.37 called “Access to Healthcare”
clinical neglect, and verbal and physical abuse by nurses, which where five abilities are discussed under the five related di-
is often based on organizational issues and professional inse- mensions of accessibility. We justify this choice, given that
curities.27 The government is investing in its “Community Health their interpretation of access uses a patient-centered ap-
Worker” program to support household healthcare.28 proach by conceptualizing access at the interface of health
The National Medicines Policy and Essential Medicines systems and populations. Also, this framework includes two
list29,30 underpin healthcare services. The Global Antibiotic further dimensions of approachability and appropriateness.
Resistance Partnership7,31 aims to contribute to investigating Levesque et al.37 link each of the five dimensions to five abil-
strategies and solutions to curb AMR. In 2016, the South Af- ities of perceive, seek, reach, pay, and engage, as illustrated in
rican National Department of Health32 released “AMR National their diagram (see Figure 1). We used this analytical framework
Strategy Framework” to facilitate its implementation at all to interpret and analyze our findings, to describe the phe-
levels of the healthcare system. The policy documents the nomena of interest, and to structure the results. We focus all
country’s national response plan to move beyond tuberculo- five abilities of people in the process of seeking healthcare
sis (TB) resistance, HIV/AIDS, and malaria by also focusing on which goes beyond other available frameworks that only de-
acute bacterial infections.33 scribe the dimensions. The various dimensions of access are
The study and field site context. This article focuses on an not completely independent constructs and can influence
area in rural South Africa. We report on qualitative research each other during an episode of illness and care.
centered around access to and use of antibiotics from a
sample of people residing in the Agincourt Health and De- METHODS
mographic Surveillance System (HDSS) site in Mpumalanga
Province. This study is embedded within the antibiotic access The research design is cross-sectional and exploratory and
and use (ABACUS)34 project conducted in six countries in is used to assess and compare community-based antibiotic
Africa and Asia. The overall project’s main purpose is to access and consumption, as well as the factors underpinning
compare antibiotic access and consumption practices across them. This study, is a subset of the ABACUS34 project under
communities (defined here as the population living within the the International Network for the Demographic Evaluation of
catchment area of the Agincourt HDSS). This is to inform the Populations and their Health Network,41 conducted in seven
design of and identify targets for intervention strategies that HDSS within six LMICs: South Africa, Mozambique, and
may be used to promote safe and appropriate antibiotic use. Ghana in Africa; Vietnam, Bangladesh, and Thailand in Asia,
Our study’s aim was to explore factors and practices around further explained in Wertheim et al.34
access and use of antibiotics and comprehension of “ABR.” Study setting. Our article is based on data from the Rural
By asking community members who reside in villages within Public Health and Health Transitions Research Unit of the
the Agincourt HDSS about their experiences and perspec- South African Medical Research Council (MRC) and the Uni-
tives, we can understand the cultural and social dimensions of versity of the Witwatersrand (the MRC/Wits Agincourt Unit,
their medicine use. Our specific objectives were to referred to here as the “Agincourt HDSS”). It is located in the
Agincourt subdistrict, Ehlanzeni District of Mpumalanga
• investigate where community members are accessing and Province, 500 km from the city of Johannesburg (Figure 2)42,43
sourcing healthcare treatment and antibiotics in north east South Africa.
• explore their understandings and experiences of antibi- The site was established in 1992 to support district health
otics and ABR system development and now investigations into causes and
1380 ANSTEY WATKINS AND OTHERS

FIGURE 1. A conceptual framework of access to healthcare by Levesque et al.37 indicating the five dimensions of accessibility of services and five
related abilities (permission to use figure granted by the publisher). This figure appears in color at www.ajtmh.org.

consequences of complex health, population, and social HDSS database,”47 which we used as a sampling frame. The
transitions are at its core, such as observational studies and relational database represents the life histories of the local
trials around preventing HIV transmission and reducing met- population (individuals and households) and takes account of
abolic disease risk.43 The site covers 420 km2, encompassing in- and out-migrations.43 The Agincourt HDSS’s “Public En-
32 villages with approximately 16,000 households which have gagement” has established a long-standing relationship with
been under an annual population census monitoring births, the HDSS population and their leaders, based on mutual trust
deaths, and migrations to update resident status and vital and respect,46 and it serves as a platform for information
events since 1992.44–46 These data are held in the “Agincourt sharing between them and the community members.

FIGURE 2. Map of the Agincourt Health and Demographic Surveillance Systems field site and its geographic location within Mpumalanga, South
Africa. This figure appears in color at www.ajtmh.org.
 RURAL SOUTH AFRICAN COMMUNITY PERCEPTIONS OF ABACUS 1381

The Agincourt HDSS is a typical marginalized rural com- TABLE 2

munity in South Africa. Unemployment rates remain high with Respondent categories for the antibiotic access and use project
common labor-related out-migration.43 Many villages are Agincourt Health and Demographic Surveillance Systems focus
group discussions (Focus groups) and geographical area
densely settled, remote, and underserved by government
Respondent categories Female respondents Male respondents Total
service providers. The district healthcare system offers citi-
zens free primary healthcare at the point of care accessed at 18–29 years old 6 (village A) 7 (village B) 13
clinics, community health centers, and district hospitals.48 30 years and older 7 (village C) 7 (village D) 14
Community group 8 (home-based 8 (community 16
Accessing the community: recruiting for community carers) elders)
member in-depth interviews. Participant recruitment fol- Total Focus groups 23 24 43
lowed a standardized approach in all ABACUS project country
sites. A sample of 17 community members was randomly
selected from the Agincourt HDSS database for participation Engagement Office, so as not to overburden the community.
in a semi-structured in-depth interview. The sample was These respondents were sampled randomly, using the data-
stratified by gender and age (Table 1), with mothers with base, from each respective village. A sample of 43 respon-
children aged five years and younger as a key population dents were involved in six focus groups.
group. The remaining participants consisted of men and Field-workers. Two female Xitsonga-speaking public health
women who cared for children older than 5 years. Children field-workers, experienced in qualitative data collection were
younger than 5 years and adults older than 60 years may have employed to assist with recruitment and data collection. They
a higher frequency of healthcare seeking. Household female were trained by F. W. and a pharmacist regarding 1) antibiotics
members, particularly mothers, play a crucial role in managing and use, 2) interview and focus group content, and 3) relevant
childhood and family illness. Our sample was standardized ethical considerations. When recruiting and interacting with
and stratified to make the qualitative findings comparable from respondents, the field-workers identified themselves as
the other seven ABACUS sites in six countries and so there Agincourt HDSS employees conducting research on antibi-
would be no systematic sampling bias in any given site. otics and antibiotics use. They also introduced themselves by
All respondents were visited at their homes and if not explaining which village they resided in (which was not any of
available, an appointment was made to return at a suitable the four villages in this research study).
time. When respondents refused or were unavailable to take Data collection. Data collection was conducted from April
part in the study, another person in the same category was to August 2017. Field-workers took turns to lead the focus
randomly sampled from the database. We over sampled to groups, whilst the other took handwritten field notes to com-
account for refusals and deaths. plement the audio recordings. They conducted half of the total
Recruiting for community member focus group dis- number of interviews each. Field-workers went through the
cussions (Focus groups). In total, six Focus groups were study information sheet (Supplemental Appendix 1) in detail
conducted. Respondents for two Focus groups were recruited with the respondents during the informed consent process,
via convenience sampling, as they belonged to existing giving them time to ask any questions and seek further clarity
community groups (female home-based carers and male where required. Respondents then made the decision whether
community elders) and were stratified by age and gender or not to participate in the study. None of the respondents
(Table 2). The recruitment of these community groups was were known to the field-workers. Light refreshments were
assisted by the voluntary Community Advisory Group (CAG) provided after each focus group.
with permission from the Induna (Village Chief). The concept of Interviews lasted 40–60 minutes and the focus groups las-
the CAG was designed by the Public Engagement Office to ted around 90 minutes. They were conducted in Xitsonga, the
ensure that community concerns and ideas are taken fully into local language. All verbal data were audio-recorded and
account in any research activities that take place in the HDSS translated into English while transcribing onto a computer by
site and to provide a forum within the community where re- the same two field-workers who collected the data, within a
search projects are discussed.46 This would, we hoped, en- week after data collection. The field-workers were not involved
courage favorable group dynamics as the respondents knew in interpreting the data for analysis but were approached if
each other, and it was convenient for us to organize as they anything was unclear in the transcript.
met regularly anyway. Recruitment of the four remaining Fo- A general interview guide was used across all country sites
cus groups was restricted to four villages who had the least (site-specific questions will be asked in a further qualitative
amount of HDSS research activity, as advised by the Public phase). This was developed by J. K., based on a literature
review which identified key issues that could be investigated
qualitatively. The questions were then divided up into sections
TABLE 1 such as 1) accessing care, 2) the suppliers, and 3) the medi-
Respondent categories for the antibiotic access and use project cines. This process was inductive, but with the concept of
Agincourt Health and Demographic Surveillance Systems in-depth access taken as a theoretical construct at the core. The in-
interviews terview guide (Supplemental Appendix 2) focused on the fol-
Female Male
Respondent categories respondents respondents Total lowing: any medications taken, where people purchase
treatment, and how they pay for it. Several questions were
Mothers who care for children 8 – 8
aged 5 years or younger specifically on antibiotic use, where respondents were asked
18–59 years with children older 3 2 5 to describe instances when they remembered being pre-
than 5 years scribed antibiotics and for what condition, and what ABR was.
60 years or older 2 2 4 We used the Xitsonga word, xitsongwatsongwana, which
Total interviews 13 4 17
means “microorganisms,” to refer to bacteria. Our team also
1382 ANSTEY WATKINS AND OTHERS

coined the term ku ala ku tira ka tiantibiotic for “ABR,” which given before data collection by signature or inked thumb-
directly translates to “the antibiotic no longer works in my print, if illiterate. The respondents were given a study in-
body.” formation leaflet with the research team’s telephone number
A study identification number was assigned to each re- in case of withdrawal from the study or to report any ethical
spondent and all identifiers were removed from the data. We concerns.
have ensured that respondents cannot be identified from any Ethical approval was granted by the University of the Wit-
text in quotations. We used different types of qualitative data, watersrand Medical Human Research Ethics Committee:
each for a specific purpose: the one-to-one interviews allowed M160753, the University of Oxford Tropical Research Ethics
for more private settings with people to elicit specific experi- Committee, OxTREC: 31-15 and the South African De-
ences and individual opinions. Whereas the focus groups partment of Health, Mpumalanga Provincial Health Research
enabled us to gain a variety of opinions from several different Committee: MP_2017RP48_440.
people on the same topic, as they interacted,49 also offering
insights into community norms. Both sets of data were ana- RESULTS
lyzed in the same manner.
Data analysis. Thematic analysis was used to analyze Description of respondents. A total of 60 respondents (34
the primary qualitative data to search for themes to organize female and 26 male) were involved in this study (excluding
data in rich detail.50 We developed a coding framework directed three refusals from the interviews and three dropouts from the
by categories from the interview guide and dimensions in the focus groups). There were 21 community members unavail-
Access to Healthcare Framework37 that specifies five demand- able for an interview and 78 community members unavailable
side abilities associated with healthcare access. Data were for focus groups. Most respondents were unemployed or
coded line by line by J. A. W., with 10 percent recoded by F. W. pensioners, and one community member was a traditional
to check for similar interpretation. When minor discrepancies healer. Most people who participated defined themselves as
were found, these were discussed and, if necessary, referred on literate and half held the final high school examination of
to J. K. for further input. In reporting this study, we have applied matriculation. As we used the data together during analysis,
the 32-item Consolidated Criteria for Reporting Qualitative there was no need to stratify the sample having found no
Studies (COREQ) checklist for interviews and focus groups.50,51 substantive differences in the different groups during our initial
As per COREQ, we describe the disciplinary backgrounds and readings.
research paradigms that our research team collectively share Themes. Data are presented under the five themes that
(including experiences in using social theory): health scientist, comprise the theoretical framework on healthcare access,
public health researcher, anthropologist, biologist, demogra- pertaining to how community members accessed healthcare
pher, clinician, and statistician. These complementary skills, in and treatment within this HDSS. This includes a subsection on
combination with our varying degrees of being “insiders” and community understandings of ABR. There was no substantial
“outsiders” within the HDSS (as described under “reflexivity”), difference in the findings from the community members in the
provided a strong basis for a reflexive and comprehensive set of interviews and focus groups, and we, therefore, present the
perspectives during the analysis. data together.
We used NVivo version 10 qualitative data analysis computer Ability to seek (personal and social values and autonomy).
software package (QSR International, Doncaster, Australia) to The community members have several choices regarding
manage our data and look for positive, negative, ambivalent, where to seek healthcare. Most of the respondents used the
and nuanced data under each topic. Codes were categorized government public health system to access healthcare and
into themes.52 J. A. W., F. W., and J. K. met four times to refine treatment. This was for acute illnesses such as bacterial infec-
themes until consensus was reached to ensure themes repre- tions requiring antibiotic treatment and for long-term, chronic
sented the breadth and depth of the dataset. diseases such as hypertension (medications collected monthly).
Reflexivity. We were aware throughout the research pro- Although the option to consult a private doctor was available in
cess of the importance of being reflexive when collecting and the locality, low socioeconomic status of the population and
interpreting the data. From the onset, we considered where we social factors resulted in their seeking care predominately from
stood on issues as a means of ensuring that we did not impose government healthcare providers.
our values on the data, thus not taking our own prejudices and
values for granted. Given we are a multidisciplinary team, with “I think the clinic is the best institution where you can get
people from different backgrounds, including some from proper help.” (Interview 1, female, 31 years old, volunteer)
Agincourt (insiders) and others from elsewhere (outside), this
gave us a greater awareness of our respective analytical and Some people described using traditional healers and then
disciplinary differences which helped to ensure trustworthi- combining biomedical treatments with traditional medicine.
ness during the analysis and interpretation stages. The re- Other potential sources of antibiotics included door-to-door
spondents only met the two field-workers and were fully aware sellers, church pastors, or market stall sellers at “pension
of who they were as employees of Agincourt HDSS, by de- points” (pop-up monthly village markets on the day pen-
scribing who they were and about their role in the study. This sioners draw social grant state pensions). People have
was intended to make sure their introductions were trans- knowledge about the different options of where to seek
parent as possible as to who they were as well as the intention healthcare from, but patient choice is often guided by financial
of the research and data collection process. means, cultural norms, or habit.
Ethical considerations. Each field-worker read the con- Ability to reach (availability of transport and social
sent form out loud to each respondent. Written informed support). Related closely to the ability to seek healthcare is
consent and permission to audio-record was voluntarily the person’s ability to get themselves or their children
 RURAL SOUTH AFRICAN COMMUNITY PERCEPTIONS OF ABACUS 1383

physically to the health facility. The cost of public transport, I am not sure if they fall under antibiotics, but I have once
which involves using local minibus taxis from the village to the used the Penicillin I was given by the nurse at the clinic.
clinic, was a concern for many respondents. Transport costs (Interview 3, female, 44 years old, domestic worker)
were described as an inhibiting factor when a household
member required healthcare. Issues around safety were A variety of bacterial infections, such as tonsillitis, urinary
raised by women with young children who had traveled by foot and respiratory tract infections, and wounds, and what anti-
to the clinic. They described how vulnerable and frightened biotics were used to treat them were described.
they felt of being attacked and raped.
Mmm. . . . I think it [antibiotics] should be taken when you
I was walking back from the clinic with a baby on my back have bladder infection. (Interview 5, female, 47 years old,
and I was with this other friend of mine. When we were unemployed)
about to cross over on the stream, two men appeared
from the bushes and wanted to rape us. We had to run A few people explained how they needed antibiotics when
back from where we were coming from with the babies on their “body soldiers” were down, referring to their immune
our backs. And the worst part of it was that we did not have system, a concept that many people have learned in relation to
even a single cent [money]. So, we waited for another cluster of differentiation 4 count (measurement of white blood
group of people who were also at the clinic and walked cells). Most respondents recognized how to take antibiotics as
home with them as a group. So, it is a challenge for us typically recommended: by finishing the prescribed course, as
when we have to go to the clinic. Those men had knives directed by the nurse or pharmacist.
with them. (Interview 1, female, 31 years old, volunteer)
I must not stop using them just because I am feeling
This illustrates the challenging social setting of this area, as
better, I have to continue until they are finished. (Interview
well as some of the concerns that people have to face when
3, female, 44 years old, domestic worker)
traveling to access care.
Ability to pay (income and social capital). Most people
Nevertheless, some said they stop taking these tablets,
relied on treatment from primary healthcare facilities, including
which they knew was ill-advised:
antibiotics, free at the point of access. However, many de-
scribed buying supplementary over-the-counter medicines
such as painkillers, anthelminthics for deworming, and cough WhenIfeelbetter,Istop.(Focusgroup3, female,below30years).
syrups from “spaza” shops (informal convenience shops). We stop taking the treatment before we even finish the
They perceived these as affordable and meant they did not course. We cheat the healthcare workers because we
have to queue at the clinic. Where possible, people also made
know that they don’t see us and no one will tell them that I
home remedies for coughs and diarrhea, using traditional
did not finish the course of my treatment. (Focus group 4,
recipes from local plants and trees. Only two of the community
home-based carer, female)
members reported consulting at a private doctor and paying
for the antibiotic prescriptions at the pharmacy. They did value
People who admitted to not finishing the prescribed course
pharmaceutical medicines as higher quality because they had
of antibiotics said they disposed of unused antibiotics into
paid for it, as compared with medication prescribed at the
their outdoor pit latrines. Others said they kept them for the
clinic, which was free and so considered lesser quality.
next episode of household illness or to be shared with other
family members or neighbors.
The treatment is for free but people are doubting to use it.
Instead they want something that they will pay for. (In-
We keep them so that we can use them again when we get
terview 2, male, 64 years old, pensioner)
ill. (Focus group 4, home-based carer, female)

Ability to engage (information, adherence, and care-

giver support). There was a very mixed understanding of the Accessing antibiotics from private healthcare. Community
term “antibiotics.” For the most part, the community had not members knew it was not permitted to buy antibiotics directly
heard of this word before. However, when described using from the pharmacy without a prescription because of
relevant Xitsonga words (as described in the Methods sec- regulation.
tion), most could then relate to the term. One retired teacher
explained what he thought the role of antibiotics was: Here in South Africa, you cannot get antibiotics without
the referral from the doctor. He will prescribe that you have
According to my understanding, as I was a teacher. . . to go and buy in the chemist but besides that, they cannot
when they talk about ‘biotic’, I think they are talking about give them to you. (Focus group 1, female, older than 30 years)
bacteria. When they talk about ‘anti’, it means they are
talking about things that are fighting against bacteria. One woman believed that pharmacies would sell antibiotics
(Interview 2, male, 64 years old, pensioner) to her on the basis they are money-making businesses.

Some people recalled taking legitimately prescribed and Chemists, they don’t care. You can buy without those
dispensed antibiotics by primary healthcare clinic or by the letters [prescription]. (Interview 6, female, 40 years old,
district hospital. traditional healer)
1384 ANSTEY WATKINS AND OTHERS

However, another respondent described how a pharmacy they are consulting to the hospital when their illness be-
refused to sell antibiotics to her without prescription and told come worse. Their traditional healers by then will say they
her to go back to the doctor. are not seeing anything wrong with the patient as they will
tell you that they tried everything according to their
I went to the chemist. I didn’t know that I was supposed to medicine. (Focus group 1, female, older than 30 years old)
come with a letter from the doctor. It’s because those pills
were strong. . . They strictly sent me back and say they The traditional healers are finding that people are turning to
want referral letter from the doctor. (Focus group 1, fe- biomedical medicine over traditional practices:
male, older than 30 years old)
As traditional healers, we are not able to work as people are
Our data provide no account of informal or unqualified going to buy those medicines straight from the chemist.
vendors selling antibiotics to the community illegally. (Interview 6, female, 40 years old, traditional healer)
Those who had consulted at a private surgery in the past
said the doctor had only prescribed antibiotics when necessary. Other health beliefs also included prayer.

They [the doctors] know the danger of abusing them. I understand that demons are operating nowadays. If you
(Interview 6, female, 40 years old, traditional healer) don’t pray for your child and think that it’s just a minor
headache, you might lose her/him due to death. (Interview
Ability to perceive (trust and expectations, health be- 1, male, 23 years old, unemployed)
liefs, and health literacy). Trust and expectations. Older
people, in particular, who were able to afford to buy drugs Health literacy. In general, when primary healthcare nurses
privately, had purchased tablets from market stall sellers using dispense antibiotics, the patient receives them in a packet with
their pension. Respondents were unsure whether the medi- the drug name printed on the packaging. This does not mean
cation bought was antibiotics as they were sold under the that patients have necessarily understood they are being
guise of nutritional supplements or analgesics. prescribed drugs classified as “antibiotics.” All respondents
reported receiving verbal and written instructions from the
nurses on how to take the medication, but they also described
They are selling mixed pills which are seven different
how they do not always read or look at the pictorial labels.
colors in one sachet. There were saying they are for flu. . . ,
others were vitamins. But in one sachet! Others were for
pains. People used to buy those pills. That is why they We don’t bother to know the names of the medicines and
don’t get well as they don’t want to go to the clinic. (Focus pills that we get from the clinics. And sometimes we don’t
group 1, female, older than 30 years) even read the name on the containers. (Interview 10, male,
23 years old, unemployed)
The young men were conscious that their grandmothers are
exploited or “hood winked” at the pension markets. Older Community members do not typically ask for a particular
women trusted market stall sellers over the clinics and phar- antibiotic at the clinic or pharmacy. With the exception, one
macies and also so that they could avoid queuing. person said she specifically requested “amoxicillin” again
because it had previously worked. We do not know if she
Our grannies are buying medicines in the pensions. The wanted it for the same condition or for another illness.
problem is that they are trusting the medicines they meet Prescribed antibiotics were referred to as “so strong” and
on the way than the ones from the chemists or from the therefore important to take because from experience, they
clinic ones. (Focus group 2, male, younger than 30 years) were effective.

One woman who had bought medicine from the market said Another thing is that many people are ill nowadays. They
it did not cure her ailments: don’t know where to go and what to do. That is why they
are trusting every information. If you can come with those
They are selling in the pension. We are buying as they are antibiotics you are talking with, people can buy them very
advertising them. But after completing the bottles you won’t fast. (Focus group 1, female, older than 30 years old)
see any change. (Focus group 1, female, older than 30 years)
Community members generally had a limited understanding
This is causing lack of trust among who these sellers are, of the purpose of antibiotics, with some associating them with
where they come from, and what the medication is. viral treatment. One woman did not understand why the an-
Health beliefs. Discussion around health beliefs was not tibiotic treatment was not reaching her “nerves.”
specifically raised by respondents. No one was hesitant to
take biomedical medicine owing to their traditional beliefs. They will stop the virus to multiply. . . Antibiotics we are
Many people described when they seek traditional healers for taking are not working. They are saying this virus is hiding in
certain ailments. our nerves. (Interview 7, female, 44 years old, housewife)

As Africans, we have too many different beliefs. There are However, many people had some understanding of antibi-
still people who are consulting to traditional healers. Eh. . . otics through their direct or indirect experience of pulmonary
 RURAL SOUTH AFRICAN COMMUNITY PERCEPTIONS OF ABACUS 1385

TB and cotrimoxazol for HIV/AIDS infection. Treatment of TB, will not respond because the first course was not finished
and the prevention and treatment of opportunistic HIV-related and the infections have gotten worse in your system. So
infections, which involves taking a long-term course of antibi- the course needs to be finished. (Interview 4, female, 31
otics, have both become very much a part of the community’s years old, unemployed)
lives, significantly enhancing people’s general health literacy.
We found that peoples’ frame of reference around antibi- A few people gave reference to how the body fights bacteria
otics usually came from their experience of the ongoing TB and how resistance to antibiotics is likely to occur.
epidemic, prevalent in this locality. Many respondents knew
someone who had been on a 6-month course of combined The reason why they resist may be because the bacteria have
antibiotics for active pulmonary TB. This direct or indirect built up and multiplied in a way that the pill cannot be able to
experience created a knowledge base that provided people treat it. (Interview 5, female, 47 years old, unemployed)
with some understanding of antibiotics. This was relevant to
the idea of needing to finish the course and the consequences
of not doing so. By applying this knowledge of adhering to DISCUSSION
antibiotic treatment, many people were able to describe what
they thought antibiotics were. This article provides insights into the ways that people
living in a rural South African HDSS site access and use anti-
I have heard that once you start using the treatment, you biotics, and also their understandings of the concept of ABR.
have to finish the course. For instance, when you are ini- We considered each of the five abilities in the Access to
tiated on TB treatment, there is a period that they give to Healthcare framework with respect to where people are
take those pills for; and after that period, it is then that they accessing and sourcing their healthcare to obtain antibiotic
tell you to stop using them. (Focus group 4, home-based treatment within this rural area. These comprised the abilities
carer, female) to seek, reach, pay, engage, and perceive.
With regard to peoples’ ability to seek healthcare, we found
Those with personal experience of being diagnosed with TB that, as in other South African studies, popular over-the-
explained what they had learnt from previously taking counter low-cost nonantibiotic medication for fever or
antibiotics. coughing in children was commonly purchased,53 whereas
antibiotics were prescribed primarily by nurses at clinics. To
some extent, people in the Agincourt HDSS have autonomy to
By the time I was diagnosed with TB, I was weak and I was
choose where to get treated (e.g., traditional healer, church
confused. But after taking antibiotics I gained energy and I
pastor, or at the clinic), but this personal choice is limited when
became strong. I was able to do my household activities. I
it comes to antibiotics, which, by law, are available exclusively
was told to take them for 6-months (Laughing) and I was
from government clinics and from (a few) private doctors. We
not to stop. If I stopped without finishing the course, this
found that antibiotics are a normalized and an acceptable line
can lead the TB not to get cured. I was told that they are
of treatment sought, as biomedicine has been part of daily life
working to help me and also to help those who are at
for some time alongside traditional medicine.
home, for the TB not to spread to them. (Interview 7, fe-
People’s ability to reach healthcare services and treatment
male, 44 years, housewife)
was subject to the availability of transport to reach the re-
People had been taught to take their treatment “well.” This spective facility, and the associated costs, which in some
was another example of understanding antibiotic knowledge cases meant that antibiotic access was limited. The threat of
stemming from TB experience. rape, although only described by one female community
member, could mean access can also be compounded by
serious safety concerns for females, in particular. This type of
With those who are on TB treatment, they are feeling
social disorganization is not uncommon in South Africa.54 We
good as long as they take their medication in a good way.
did not find any evidence for people self-medicating antibi-
That’s why they have to complete 6 months while on
otics, largely because the purchasing of antibiotics without
treatment. This will prevent the spread of TB to other
prescription was not seen as feasible for the aforementioned
people. (Interview 8, female, 27 years old, unemployed)
reasons.
The actual source and location of where to seek and obtain
Antibiotic resistance. The term “ABR” was unfamiliar to all
healthcare corresponded with community members’ ability to
respondents.
pay for healthcare, based on their income. Antibiotic medi-
cation is free from primary healthcare services, making it the
I have never heard of antibiotic resistance. (Interview 9, most common derived source. Many people were willing to
female, 31 years old, volunteer) purchase affordable and commonly available medication,
such as paracetamol, used for pain and fever, as found in other
However, as described earlier, people did understand the South African studies.54 For a minority, this “choice” extended
concept, in relation to TB and HIV/AIDS treatment, and several to more expensive prescribed medication from a private
respondents could give accurate descriptions of what it is. doctor and then purchased from a local, legitimate private
pharmacy. Other studies have found rural patients may be-
. . . In my understanding, there is resistance because come dissatisfied with the clinic and choose to visit a private
people don’t finish the course. And when you use them doctor expecting to get the (right) treatment.55,56 Some people
[antibiotics] again, maybe the illness is getting worse; they in our study also chose to buy medications using their pension
1386 ANSTEY WATKINS AND OTHERS

money, from alternative unlicensed providers, usually to from other medicines, and that, therefore, the introduction of
supplement medication from the clinic. Older people were new concepts alters behaviors for other medicines as well—
found to be more susceptible to purchasing from market stall with potentially unforeseen consequences. Much research
sellers. Little is known about what medicine they are selling, has been conducted on social and psychological factors
as most published studies are on traditional medication influencing health outcomes,64,65 and this has had a direct
practices.57,58 In our data, the illegal selling or buying of anti- impact on the ways in which both TB and HIV/AIDS have been
biotics in this area was not reported. This indicates that at addressed in South Africa. Media campaigns targeting rural
worst, this is only a minor problem in this rural setting, given and urban communities have successfully sought to convey
the regulation of sales of antibiotics. This is unusual in an LMIC information about the two diseases,66,67 which suggest that a
setting, as it is well documented that nonprescription sales similar method could be used in the fight against ABR. Com-
and the dispensing of antibiotics in more urban areas of munity groups could also be established to discuss antibiotic
Zambia and Tanzania are a widespread problem due to stewardship and the impact of resistance, following principles
weaker regulatory enforcement.59,60 developed in an urban township in Cape Town, in which pa-
Medication from the clinic was sometimes perceived of tient adherence groups were found to be effective as a model
lesser quality than the pharmacy’s supplies. Free medication of care.68
from the healthcare facilities was valued less by several peo- These results suggest that people are aware of antibiotics in
ple. The reasoning was that if they were given something for some capacity and have specific local interpretations of this
free, it must be of poorer quality than something they paid for. type of medicine. Also, people can relate to the concept and its
In line with other literature, a study in China has also demon- extension to ABR because of their TB/HIV knowledge. We do
strated that a “concern with the quality of medicines led to not know the extent to which these new conceptions were
distrust in the public sector” with people preferring medicines specific to antibiotics rather than just gaining new knowledge
purchased from private sector pharmacies.61 Yet, we found about medicine more generally. Further research would have
that people mostly trusted the free antibiotics from the clinic to discover whether these new conceptions mapped perfectly
nurses. Factors for this include (potential) trust in the public or imperfectly onto clinical definitions as these existing data
healthcare system and not always having the financial means do not give light to this.
to pay for private healthcare. Study strengths and limitations. As random sampling for
The community members described their engagement with the interviews were used, this study included respondents
information around antibiotic instructions given by the nurses from four different HDSS villages. This was considered a
as satisfactory. However, they did describe instances of not strength because it meant their individual healthcare experi-
finishing the antibiotic course and disposing of them in- ences differed and so were the clinics that healthcare was
appropriately. Some people could not determine the dif- accessed from. By using community leaders to assist with
ference between antibiotics for bacterial infections and focus group recruitment, we were able to recruit preexisting
analgesics for pain relief, perceiving both as “just tablets to get groups of people, already known to one another. Although this
better.” The respondents displayed some understanding of may have encouraged a flowing discussion, however, we
ABR by describing its causes. People were also aware that recognize that the community leaders may have selected
antibiotics are strong, with the capacity to cure a range of people for their own particular reasons, and this could have
ailments.62 Even though people knew they should finish a introduced some bias into the dataset.
course of antibiotics, some conceded that they do not always A further limitation of this study was the difficulty in in-
follow the verbal or written instructions. This was also found in vestigating a topic that the HDSS population were unfamiliar
a study by Friend-du Preez et al.54 on health seeking behavior with when asked about medical terms that cannot be
for childhood illness in urban South Africa, whereby antibiotics directly translated into the local language. To counteract this
were not always used as intended or according to the rec- problem, we used Xitsonga words xitsongwatsongwana
ommended instructions.63 (microorganisms/bacteria) and ku ala ku tira ka tiantibiotic
The community’s ability to understand antibiotics and ABR (ABR). In the end, we found issues around terminology did not
was derived from their experiences of other well-known ill- in fact matter because although people did not know the
nesses, such as TB, and opportunistic infections, such as “antibiotic,” they could mostly understand the concept.
pneumonia, often associated with HIV/AIDS. The relationship For three of the five themes presented in the conceptual
between new concepts and existing notions of medicine and Access to Healthcare Framework, significantly more data
care in the context of TB and HIV/AIDS is of interest. These were found (for the abilities to pay, to engage, and to perceive).
diseases have played a prominent role in the community over The abilities to seek and to reach antibiotics could therefore be
a long period of time and have been covered extensively by investigated further in future research into this topic. Also, we
public health campaigns. Therefore, we found that people’s used a different theoretical framework to help analyze and
health literacy, their beliefs, trust and expectations (their ability interpret our qualitative data compared with the overall mixed
to perceive), and their level of health information, adherence, methods design of the larger, six-country study (Sen’s Ca-
and empowerment (their ability to engage) were often based pability Approach36). Sen’s theory did not fit with the specific
on illnesses such as TB. Through this knowledge, people research questions that we were trying to answer here. Fur-
could grasp the concepts of “antibiotics” and “resistance” thermore, the data collection methods that we used allowed
even if they were not familiar with the words themselves. It is us to collect a standardized set of qualitative data from all six
not yet clear whether by “grasping” new concepts of “antibi- participating countries, which in turn will permit comparisons
otics” and “resistance,” the respondents’ concepts over- between all of the sites. Although there was the opportunity for
lapped with clinical concepts. For example, it could be probing specific topics that arose during the semi-structured
possible that people are still not able to distinguish antibiotics interviews and focus groups, and we sought to encourage an
 RURAL SOUTH AFRICAN COMMUNITY PERCEPTIONS OF ABACUS 1387

open atmosphere during the discussions, this need for stan- there, which is usually a target for intervention (as is the case
dardization required us to follow a broad set of predefined for the other ABACUS countries). In that sense, awareness
topics. It was not feasible, given the nature of the larger raising is one of a very limited number of options open to the
ABACUS project, to engage in more ethnographic research Department of Health in this setting (and we are aware of the
which could have opened up more unexpected lines of en- inherent limitations even these have).
quiry, concerning, for example, different conceptions of anti- Many policy debates have been focused on the right of
biotics and the meanings behind existing behaviors. access to antibiotics,79 rather than on patient education and
The field-workers who collected, translated, and tran- safety. This study gives evidence for the need for more edu-
scribed the interviews and focus groups were not involved in cation and training around taking antibiotics appropriately for
the analysis and interpretation of the data. Conversations be- the community and the healthcare workers. However, by im-
tween the two field-workers and researchers post-interviews proving patient education around antibiotics and resistance
allowed them to give feedback about the provisional findings. will not solve unsafe antibiotic use among patients as this is
Not specific to our study, further limitations include the posi- unrealistic. Also, a focus on education and knowledge can
tionality and reflexivity of us, as the authors, and how we acted ignore structural and contextual facets of behavior, for ex-
on our pre-assumptions and consideration into how we ample, cultural meanings of good care, economic constraints,
influenced the research process,69 and also the relationship discrimination, patients’ despair and experiences of un-
between the local field-workers and respondents (both of certainty, and social relationships between patients and
whom are from the HDSS population). healthcare workers. If substantial changes to a country’s
Implications for policy, practice, and future research. structures are not addressed, then bringing about safer anti-
Many of our respondents expressed a desire to learn more biotic use may not happen. We are not in a position to make
about antibiotics as a result of participating in the research. substantial changes to the structures and the context, thereby
This finding is not uncommon: a study by Norris70 found that bringing about appropriate antibiotic use, although we may be
a Samoan population in New Zealand also lacked under- in a position to bring about an improvement in person’s un-
standing around the antibiotics and wanted to know more derstanding of the issue through education and awareness-
about biomedical medicine. The study concluded that de- raising programs.
veloping appropriate messages around preventing and man- Another source of public health education can be using the
aging infections and “building on culturally based practices is community pharmacy model whereby pharmacy staff cham-
a safe strategy.” pion AMR and ABR and are empowered to initiate conversa-
Two sources of information already available to the com- tions with patients around antibiotics.80 The impact and
munity are the nurses’ health talks at clinics and communi- practicality of trying to “empower” local pharmacists to be
cations from community healthcare workers or home-based AMR/ABR stewards may be problematic. It may possibly en-
carers, whose role includes teaching and advising the com- tail unforeseen social consequences (e.g., changing the rela-
munity about how to take medications. These may be valuable tionships between patients and the health system) and
avenues for informing people about their role in ABR, both pharmacists may comply strategically, to pursue non-
when taking antibiotics themselves or when administering altruistic goals that could entail worsened patterns of phar-
them to their children. By making the nurses and community maceutical use.
healthcare workers’ pledge to be “Antibiotic Guardians,”71,72 At national or provincial level, an additional intervention
this could be part of the South African Department of Health could involve the regulation of the drug sales at pension points
wider antibiotic stewardship programme.6,73 Both groups of and to check their drugs are safe, legal, and not counterfeit.
healthcare workers can be trained to include relevant patient Another study to interview the suppliers of medication in the
information about safe antibiotic use, tailored to the local Agincourt HDSS site is presently underway and we may get a
communities’ needs and existing knowledge levels and par- better sense through this of the medications they are selling,
adigms. This would speak to “Education and Communication as well as whether these may include any antibiotic and/or
and Public Awareness,” one of the pillars (objectives) of the counterfeit drugs.
“South African Antimicrobial Resistance Strategy Framework”74,75 There has been a modest number of studies on social in-
(which also include enhancing ABR surveillance, governance, terventions in first-line public health primary healthcare
stewardship, and prevention). settings.73,81 All of these educational and awareness-raising
Antibiotic awareness education could also be supported by interventions outlined would need to be developed, imple-
national public health campaigns via social media and stan- mented, and then evaluated rigorously, before being
dard communication techniques. This approach is being used scaled up.
in the United Kingdom, for example, where the National Health
Service has launched a campaign called “Treating your in- CONCLUSION
fection without antibiotics,”76 using principles similar to those
used in the “World Antibiotics Awareness Week” campaign.77 From the perspective of the Access to Healthcare frame-
There are potential limitations of knowledge and awareness work, our study has found that despite free provision at pri-
campaigns that need to be evaluated,78 and other comple- mary healthcare clinics, “seeking” and “reaching” antibiotics
mentary activities—within and outside of the public health could be problematic because of difficulties reaching the fa-
sector that might be necessary to assist in altering behavior. cilities. Moreover, “paying” for the transport to the clinics
Broader structural interventions are necessary; yet in this area, constitutes a significant barrier to many people. The respon-
it has not yet become immediately clear what these could be dents’ ability to “engage” with and “perceive” what antibiotics
as the antibiotics are not available illegally outside the health and ABR are was based largely on their prior health literacy,
system, so there are no laws that need to be enforced better which, for many people, has been developed in relation to TB
1388 ANSTEY WATKINS AND OTHERS

and HIV/AIDS treatment. We recognize education is only one Francesc Xavier Gómez-Olivé, MRC/Wits Rural Public Health and
facet in addressing the misuse of antibiotics yet factors such Health Transitions Research Unit (Agincourt), School of Public Health,
Faculty of Health Sciences, University of the Witwatersrand, Johan-
as poverty, insecure income, and lack of access to healthcare nesburg, South Africa, E-mails: [email protected] and f.gomez-
may impact of the benefit of this education.82 If locally con- [email protected]. Heiman Wertheim, Wellcome Trust Major
textualized and used in coordination with suitable training for Overseas Programme, Oxford University Clinical Research Unit,
healthcare workers, the safe use of antibiotics could be pro- Hanoi, Vietnam, Nuffield Department of Medicine, Centre for Tropical
moted through the range of preexisting materials on com- Medicine, University of Oxford, Oxford, United Kingdom, and De-
partment of Medical Microbiology, Radboudumc Center for Infectious
munity antibiotic education, from organizations such as the Diseases, Nijmegen, The Netherlands, E-mail: heiman.wertheim@
WHO.78 gmail.com. Osman Sankoh, College of Medicine and Allied Health
The ABACUS project aims to provide an empirical basis for Sciences, University of Sierra Leone, Freetown, Sierra Leone, and
informing future, patient-level social interventions for appro- Faculty of Health Sciences, School of Public Health, University of the
Witwatersrand, Johannesburg, South Africa, E-mail: osman.sankoh@
priate and safe ABACUS across six LMICs. This study’s focus statistics.sl. John Kinsman, Faculty of Medicine, Department of Public
on knowledge and social practices related to antibiotic use Health and Clinical Medicine, Epidemiology and Global Health (Umeå
has implications related to educational and awareness-raising Centre for Global Health Research), Umeå University, Umeå, Sweden,
interventions relevant to rural South Africa, rather than social and Department of Public Health Sciences, Global Health (IHCAR),
intervention recommendations that focus on the social and Karolinska Institutet, Stockholm, Sweden, E-mail: john.kinsman@
umu.se.
cultural environment in which people consider and seek care.
Educational campaigns for this specific context could be de- This is an open-access article distributed under the terms of the
Creative Commons Attribution License, which permits unrestricted
veloped by drawing on people’s experience with TB and HIV/ use, distribution, and reproduction in any medium, provided the
AIDS treatment. By developing effective patient education original author and source are credited.
and health promotion materials to reduce unsafe antibiotic
use, we need to understand how people talk about and think
REFERENCES
about antibiotics and infection.14,83 Also, the absence of
existing notions of antibiotics and resistance among margin- 1. Brink AJ, van den Bergh D, Mendelson M, Richards GA, 2016.
alized groups requires other forms of interventions that involve Passing the baton to pharmacists and nurses: new models of
different approaches than standardized material on commu- antibiotic stewardship for South Africa? S Afr Med J 106:
947–948.
nity antibiotic education. Rather, a contextually relevant long- 2. Tadesse BT, Ashley EA, Ongarello S, Havumaki J,
term antibiotic-based “curriculum” that is driven by the South Wijegoonewardena M, González IJ, Dittrich S, 2017. Antimi-
African health system, for members of the public to un- crobial resistance in Africa: a systematic review. BMC Infect Dis
derstand the concepts of illness and associated antibiotic 17: 616.
treatments, would be welcomed. 3. O’Neill J, 2016. Tackling Drug-Resistant Infections Globally: Final
Report and Recommendations. The Review on Antimicrobial
Resistance. London, United Kingdom: HM Government and
Received February 27, 2018. Accepted for publication January 26, Welcome Trust.
2019. 4. Laxminarayan R, Duse A, Wattal C, Zaidi AK, Wertheim HF,
Published online April 15, 2019. Sumpradit N, Vlieghe E, Hara GL, Gould IM, Goossens H, 2013.
Antibiotic resistance - the need for global solutions. Lancet
Note: Supplemental appendices appear at www.ajtmh.org. Infect Dis 13: 1057–1098.
Acknowledgments: At the Agincourt HDSS, we thank the community 5. NICE, 2015. Antimicrobial Stewardship: Systems and Processes
members and the Indunas for their contribution to this study, to Floidy for Effective Antimicrobial Medicine Use. London, United
Wafawanaka (pharmacist) and Sizzy Ngobeni (project coordinator) Kingdom: The National Institute for Health and Care Excellence.
and Meriam Maritze and Princess Makhubele (qualitative field- 6. Papoutsi C, Mattick K, Pearson M, Brennan N, Briscoe S, Wong G,
workers). We also appreciate Chandrika Nair and Matt Keeling (War- 2017. Social and professional influences on antimicrobial pre-
wick) and Déirdre Hollingsworth (Oxford) for their assistance during scribing for doctors-in-training: a realist review. J Antimicrob
J. A. W.’s Early Career Fellowship; the ABACUS Kanchanaburi HDSS Chemother 72: 2418–2430.
team (Thailand) for their input; the coauthors; journal reviewers; and 7. Apalata T et al., 2011. Global antibiotic resistance partnership -
editor for their patience in the delay in finalizing this paper. This is for situation analysis: antibiotic use and resistance in South Africa.
my Jago. S Afr Med J 101: 549–596.
8. Agyepong IA et al., 2017. The path to longer and healthier lives for
Financial support: The ABACUS project is funded by The Wellcome
all Africans by 2030: the Lancet Commission on the future of
Trust (grant 109595/Z/15/Z), Major Overseas Programme, United
Kingdom, through INDEPTH Network; and Volkswagen Stiftung, led health in sub-Saharan Africa. Lancet 390: 2803–2859.
by Project Principle Investigator: H. W., University of Oxford: http:// 9. Leopold SJ, van Leth F, Tarekegn H, Schultsz C, 2014. Antimi-
www.indepth-network.org/projects/abacus. The funders had no role crobial drug resistance among clinically relevant bacterial iso-
in the project or this study’s design, data collections and analysis, the lates in sub-Saharan Africa: a systematic review. J Antimicrob
decision to publish, or preparation of this manuscript. The lead author Chemother 69: 2337–2353.
was sponsored by the INTEGRATE AMR—Integrating antimicrobial 10. Leung E, Weil DE, Raviglione M, Nakatani K, 2011. The WHO
resistance research expertise across disciplines at the Warwick Anti- Policy Package to Combat Antimicrobial Resistance. Available
microbial Interdisciplinary Centre (WAMIC), School of Life Sciences, at: http://www.who.int/bulletin/volumes/89/5/11-088435/en/.
University of Warwick, United Kingdom, on an Early Career Fellow- Accessed February 20, 2018.
ship, funded by the Engineering and Physical Sciences Research 11. CDDEP, 2018. Resistance Map. Available at: https://resistancemap.
Council (Grant number: EP/M027503/1). cddep.org/AntibioticResistance.php. Accessed December 10,
2018.
Data availability statement: The supporting data of the results or an-
12. Essack S, Desta A, Abotsi R, Agoba E, 2017. Antimicrobial re-
alyzes can be made available on request.
sistance in the WHO African region: current status and roadmap
Authors’ addresses: Jocelyn Anstey Watkins, Social Science and for action. J Public Health 39: 8–13.
Systems in Health, Division of Health Sciences, Warwick Medical 13. Chandler C, Hutchinson E, Hutchison C, 2016. Addressing Anti-
School, Gibbet Hill, The University of Warwick, Coventry, United microbial Resistance through Social Theory: An Anthropologi-
Kingdom, E-mail: [email protected]. Fezile Wagner and cally Oriented Report. Technical Report. London, United
 RURAL SOUTH AFRICAN COMMUNITY PERCEPTIONS OF ABACUS 1389

Kingdom: London School of Hygiene & Tropical Medicine, 35. Sen AK, 1990. Development as Capability Expansion. Oxford,
1–42. United Kingdom: Oxford University Press.
14. Wood F, 2016. Antimicrobial Resistance and Medical Sociology: 36. Sen A, 1999. Development as Freedom. Oxford, United Kingdom:
Research Brief. Cardiff, Wales: ESRC AMR Research Champion/ Oxford University Press.
University of Bristol, Cardiff University, 1–15. 37. Levesque J-F, Harris MF, Russell G, 2013. Patient-centred ac-
15. Haak H, Radyowijati A, 2010. Determinants of antimicrobial use: cess to health care: conceptualising access at the interface of
poorly understood, poorly researched. Sosa ADJ, Byarugaba health systems and populations. Int J Equity Health 12: 18.
DK, Amabile-Cuevas CF, Hsueh PR, Kariuki S, Okeke IN, eds. 38. Shengelia B, Murray CJ, Adams OB, 2003. Beyond access and
Antimicrobial Resistance in Developing Countries. New York, utilization: defining and measuring health system coverage.
NY: Springer, 283–300. Murray CJL, Evans DB, eds. Health Systems Performance As-
16. Broom A, Kirby E, Gibson AF, Post JJ, Broom J, 2017. Myth, sessment: Debates, Methods and Empiricism. Geneva, Swit-
manners, and medical ritual: defensive medicine and the fetish zerland: World Health Organization, 221–234.
of antibiotics. Qual Health Res 27: 1994–2005. 39. Peters DH, Garg A, Bloom G, Walker DG, Brieger WR, Hafizur
17. Bledsoe CH, Goubaud MF, 1985. The reinterpretation of western Rahman M, 2008. Poverty and access to health care in de-
pharmaceuticals among the Mende of Sierra Leone. Soc Sci veloping countries. Ann N Y Acad Sci 1136: 161–171.
Med 21: 275–282. 40. Aday LA, Andersen R, 1974. A framework for the study of access
18. van der Geest S, 1987. Self-care and the informal sale of drugs in to medical care. Health Serv Res 9: 208–220.
South Cameroon. Soc Sci Med 25: 293–305. 41. Sankoh O, Byass P, 2012. The INDEPTH Network: Filling Vital
19. van der Geest S, 1982. Part 1: the illegal distribution of western Gaps in Global Epidemiology. Oxford, United Kingdom: Oxford
medicines in developing countries: pharmacists, drug pedlars, University Press.
injection doctors and others. A bibliographic exploration. Med 42. Kinsman J, Norris SA, Kahn K, Twine R, Riggle K, Edin K,
Anthropol 6: 197–219. Mathebula J, Ngobeni S, Monareng N, Micklesfield LK, 2015. A
20. van der Geest S, Whyte SR, 1989. The charm of medicines: model for promoting physical activity among rural South Afri-
metaphors and metonyms. Med Anthropol Q 3: 345–367. can adolescent girls. Glob Health Action 8: 28790.
21. Whyte SR, van der Geest S, Hardon A, 2002. Social Lives of 43. Kahn K et al., 2012. Profile: agincourt health and socio-
Medicines. Cambridge, United Kingdom: Cambridge University demographic surveillance system. Int J Epidemiol 41: 988–1001.
Press. 44. Sartorius B, 2013. Modelling determinants, impact, and space-
22. Vuckovic N, 1999. Fast relief: buying time with medications. Med time risk of age-specific mortality in rural South Africa: in-
Anthropol Q 13: 51–68. tegrating methods to enhance policy relevance. Glob Health
23. Goudge J, Gilson L, Russell S, Gumede T, Mills A, 2009. The Action 6: 19239.
household costs of illness in rural South Africa with free public 45. Agincourt, 2014. Agincourt Village Fact Sheet. Acornhoek, South
primary care and hospital exemptions for the poor. J Trop Med Africa: MRC/Wits Agincourt HDSS.
Int Health 14: 458–467. 46. Twine R, Hundt GL, Kahn K, 2017. The ‘experimental public’ in
24. Cunnama L, Honda A, 2016. A mother’s choice: a qualitative longitudinal health research: views of local leaders and service
study of mothers’ health seeking behaviour for their children providers in rural South Africa. Glob Health Res Policy 2: 26.
with acute diarrhoea. BMC Health Serv Res 16: 669. 47. Kahn K, Tollman SM, Collinson MA, Clark SJ, Twine R, Clark BD,
25. Goudge J, Russell S, Gilson L, Gumede T, Tollman S, Mills A, Shabangu M, Gómez-Olivé FX, Mokoena O, Garenne ML, 2007.
2009. Illness related impoverishment in rural South Africa: why Research into health, population and social transitions in rural
does social protection work for some households but not oth- South Africa: data and methods of the Agincourt Health and
ers? J Int Dev 21: 231–251. Demographic Surveillance System. Scand J Public Health 35:
26. Goudge J, Gilson L, Russell S, Gumede T, Mills A, 2009. Afford- 8–20.
ability, availability and acceptability barriers to health care for 48. Anstey Watkins JOT, 2016. The Vutivi Study: Understanding the
the chronically ill: longitudinal case studies from South Africa. Potential Role for Appropriate Digital Technological Solutions in
BMC Health Serv Res 9: 75. the Innovation of Health System Design, Implementation and
27. Jewkes R, Abrahams N, Mvo Z, 1998. Why do nurses abuse pa- Normalisation in Rural South Africa for both Patients and
tients? Reflections from South African obstetric services. Soc Health-Workers: A Critical Exploratory Analysis. PhD Thesis,
Sci Med 47: 1781–1795. Warwick Medical School, The University of Warwick, Coventry,
28. van Heerden A, Harris DM, van Rooyen H, Barnabas RV, United Kingdom, 54–139.
Ramanathan N, Ngcobo N, Mpiyakhe Z, Comulada WS, 2017. 49. Kitzinger J, 1994. The methodology of focus groups: the impor-
Perceived mHealth barriers and benefits for home-based HIV tance of interaction between research participants. Soc Health
testing and counseling and other care: qualitative findings from Illness 16: 103–121.
health officials, community health workers, and persons living 50. Braun V, Clarke V, 2006. Using thematic analysis in psychology.
with HIV in South Africa. Soc Sci Med 183: 97–105. Qual Res Psychol 3: 77–101.
29. Ameh S, Gomez-Olive FX, Kahn K, Tollman SM, Klipstein- 51. Crowe S, Brugha R, Clarke N, 2017. ‘You do not cross them’:
Grobusch K, 2014. Predictors of health care use by adults 50 hierarchy and emotion in doctors’ narratives of power relations
years and over in a rural South African setting. Glob Health in specialist training. Soc Sci Med 186: 70–77.
Action 7: 24771. 52. Kabudula CW, Houle B, Collinson MA, Kahn K, Gómez-Olivé FX,
30. NDOH, 2017. Essential Drugs Programme (EDP). Available at: Clark SJ, Tollman S, 2017. Progression of the epidemiological
http://www.health.gov.za/index.php/essential-drugs- transition in a rural South African setting: findings from pop-
programme-edp. Accessed October 30, 2017. ulation surveillance in Agincourt, 1993–2013. BMC Public
31. GARP, 2010. The Global Antibiotic Resistance Partnership Re- Health 17: 424.
port: South Africa. Available at: https://cddep.org/partners/ 53. Tong A, Sainsbury P, Craig J, 2007. Consolidated criteria for
global-antibiotic-resistance-partnership/garp-network/garp-south- reporting qualitative research (COREQ): a 32-item checklist for
africa/. Accessed Febraury 9, 2018. interviews and focus groups. Int J Qual Health Care 19:
32. NDOH, 2015. Antimicrobial Resistance - Background Document. 349–357.
Pretoria, South Africa: National Department of Health. 54. Friend-du Preez N, Cameron N, Griffiths P, 2013. “So they believe
33. Abera Abaerei A, Ncayiyana J, Levin J, 2017. Health-care utili- that if the baby is sick you must give drugs...” the importance of
zation and associated factors in Gauteng province, South medicines in health-seeking behaviour for childhood illnesses
Africa. Glob Health Action 10: 1305765. in urban South Africa. Soc Sci Med 92: 43–52.
34. Wertheim HFL et al., 2017. Community-level antibiotic access 55. Bogart LM, Chetty S, Giddy J, Sypek A, Sticklor L, Walensky RP,
and use (ABACUS) in low- and middle-income countries: finding Losina E, Katz JN, Bassett IV, 2013. Barriers to care among
targets for social interventions to improve appropriate antimi- people living with HIV in South Africa: contrasts between pa-
crobial use—an observational multi-centre study. Wellcome tient and healthcare provider perspectives. AIDS Care 25:
Open Res 2: 58. 843–853.
1390 ANSTEY WATKINS AND OTHERS

56. van der Hoeven M, Kruger A, Greeff M, 2012. Differences in health 69. Gilgun J, 2010. Reflexivity and qualitative research. Curr Issues
care seeking behaviour between rural and urban communities Qual Res 1: 1–8.
in South Africa. Int J Equity Health 11: 31. 70. Norris P, Churchward M, Fa’alau F, Va’ai C, 2009. Understand-
57. Semenya SS, Potgieter MJ, 2014. Bapedi traditional healers in the ing and use of antibiotics amongst Samoan people in New
Limpopo Province, South Africa: their socio-cultural profile and Zealand. J Prim Health Care 1: 30–35.
traditional healing practice. J Ethnobiol Ethnomed 10: 4. 71. Kesten JM, Bhattacharya A, Ashiru-Oredope D, Gobin M, Audrey
58. Maimela E, Geertruyden J-P, Alberts M, Modjadji SE, Meulemans S, 2018. The Antibiotic Guardian campaign: a qualitative eval-
H, Fraeyman J, Bastiaens H, 2015. The perceptions and per- uation of an online pledge-based system focused on making
spectives of patients and health care providers on chronic better use of antibiotics. BMC Public Health 18: 5.
diseases management in rural South Africa: a qualitative study. 72. DOH, 2018. Become an Antibiotic Guardian. Available at: http://
BMC Health Serv Res 15: 143. antibioticguardian.com/south-africa/. Accessed December 7,
59. Dillip A, Embrey M, Shekalaghe E, Ross-Degnan D, Vialle-Valentin 2018.
C, Kimatta S, Liana J, Rutta E, Valimba R, Chalker J, 2015. What 73. Cox J, Vlieghe E, Mendelson M, Wertheim H, Ndegwa L, Villegas
motivates antibiotic dispensing in accredited drug dispensing M, Gould I, Hara GL, 2017. Antibiotic stewardship in low- and
outlets in Tanzania? A qualitative study. Antimicrob Resist In- middle-income countries: the same but different? Clin Micro-
fect Control 4: 30. biol Infect 23: 812–818.
60. Kalungia AC, Burger J, Godman B, Costa JO, Simuwelu C, 2016. 74. Mendelson M, Matsoso M, 2015. The South African Antimicro-
Non-prescription sale and dispensing of antibiotics in com- bial Resistance Strategy Framework. Monitoring, Surveillance
munity pharmacies in Zambia. Expert Rev Anti Infect Ther 14: and National Plans. Cape Town, South Africa: AMR Control,
1215–1223. 54–61.
61. Yu M, Zhao G, Lundborg CS, Zhu Y, Zhao Q, Xu B, 2014. 75. NDOH, 2017. Guidelines on Implementation of the Antimicrobial
Knowledge, attitudes, and practices of parents in rural China on Strategy in South Africa: One Health Approach & Governance.
the use of antibiotics in children: a cross-sectional study. BMC Pretoria, South Africa: Ministerial Advisory Committee on An-
Infect Dis 14: 112. timicrobial Resistance, South African National Department of
62. Luque-Fernandez MA, Van Cutsem G, Goemaere E, Hilderbrand Health.
K, Schomaker M, Mantangana N, Mathee S, Dubula V, Ford 76. Gelband H, Molly Miller P, Pant S, Gandra S, Levinson J, Barter D,
N, Hernán M, 2013. Effectiveness of patient adherence White A, Laxminarayan R, 2015. The state of the world’s anti-
groups as a model of care for stable patients on anti- biotics 2015. Wound Healing South Afr 8: 30–34.
retroviral therapy in Khayelitsha, Cape Town, South Africa. 77. NHS, 2017. Treating Your Infection without Antibiotics. Available
PLoS One 8: e56088. at: https://www.nhs.uk/video/Pages/treating-your-infection-
63. Llewelyn MJ, Fitzpatrick JM, Darwin E, Gorton C, Paul J, Peto TE, without-antibiotics.aspx. Accessed October 31, 2017.
Yardley L, Hopkins S, Walker AS, 2017. The antibiotic course 78. Charoenboon N, Haenssgen MJ, Warapikuptanun P, Xayavong T,
has had its day. BMJ 358: j3418. Zaw YK, 2019. Translating antimicrobial resistance: a case
64. Peltzer K, Naidoo P, Matseke G, Louw J, Mchunu G, Tutshana B, study of context and consequences of antibiotic-related com-
2012. Prevalence of psychological distress and associated munication in three northern Thai villages. Palgrave Commun
factors in tuberculosis patients in public primary care clinics in 5: 23.
South Africa. BMC Psychiatry 12: 89. 79. WHO, 2017. World Antibiotic Awareness Week. Available at:
65. Sherr L, Cluver LD, Betancourt TS, Kellerman SE, Richter LM, http://www.who.int/campaigns/world-antibiotic-awareness-
Desmond C, 2014. Evidence of impact: health, psychological week/en/. Accessed November 20, 2017.
and social effects of adult HIV on children. AIDS 28: 80. Cooper RJ, Tsoneva J, 2017. Benefits and tensions in delivering
S251–S259. public health in community pharmacies—a qualitative study of
66. Tariq S, Hoffman S, Ramjee G, Mantell JE, Phillip JL, Blanchard K, healthy living pharmacy staff champions. Int J Pharm Pract 25:
Lince-Deroche N, Exner TM, 2018. “I did not see a need to get 351–357.
tested before, everything was going well with my health”: a 81. Sumpradit N et al., 2012. Antibiotics smart use: a workable model
qualitative study of HIV-testing decision-making in Kwazulu- for promoting the rational use of medicines in Thailand. Bull
Natal, South Africa. AIDS Care 30: 32–39. World Health Organ 90: 905–913.
67. Karim SSA, Churchyard GJ, Karim QA, Lawn SD, 2009. HIV in- 82. Haenssgen MJ, Xayavong T, Charoenboon N, Warapikuptanun P,
fection and tuberculosis in South Africa: an urgent need to Khine Zaw Y, 2018. The consequences of AMR education and
escalate the public health response. Lancet 374: 921–933. awareness raising: outputs, outcomes, and behavioural im-
68. Naidoo P, Simbayi L, Labadarios D, Ntsepe Y, Bikitsha N, Khan G, pacts of an antibiotic-related educational activity in Lao PDR.
Sewpaul R, Moyo S, Rehle T, 2016. Predictors of knowledge Antibiotics 7: 95.
about tuberculosis: results from SANHANES I, a national, 83. Edgar T, Boyd SD, Palamé MJ, 2008. Sustainability for behaviour
cross-sectional household survey in South Africa. BMC Public change in the fight against antibiotic resistance: a social mar-
Health 16: 276. keting framework. J Antimicrob Chemother 63: 230–237.