How can we assess the burden of muscle, bone and joint conditions in rural Botswana: context and methods for the MuBoJo focused ethnography

Study design

The MuBoJo project is a focused ethnography. Researchers have long used ethnography
as a comparative method to investigate patterns of human behaviour and cognition through
observations and interactions in natural settings 30]. In short, ethnography is “the work of describing culture” 31],32]. Historically, ethnographers engaged in long-term field studies, most lasting a year
or longer, with the goal of comprehensively depicting a cultural group.

Focused ethnography is a methodology that emphasises anthropological theory and methods
while limiting the scope and duration of fieldwork to a specific program-relevant
research problem and is “often used to determine ways to improve care and care processes”
32]. One of the virtues of ethnographic research is “that it remains flexible and responsive
to local circumstances” 33]. Because we wanted to explore the meaning of a set of related concepts about the
burden and care of MuBoJo problems, alongside the temporal relationships of establishing
spine care centres in rural Botswana and conducting research during a three-year doctoral
programme, we designed a focused ethnography.

Although a priori research hypotheses are not appropriate for this methodology, research questions
adapted from Kleinman 25] guided development of the project aim, methods and analytic concepts. These included:

What are the meaning contexts of burden and muscle, bone and joint conditions in
Shoshong?

What are the paths people travel when they have muscle, bone or joint troubles?

What are the culture-specific and universal characteristics of caring for MuBoJo
conditions in Shoshong?

Setting

Botswana

Botswana is a land-locked country in the southern part of the African continent, bordered
by South Africa to the south and southeast, Namibia to the west and north, and Zimbabwe
to the northeast [see Additional file 1 for a map of Botswana]. The Kalahari Desert covers almost 70% of the land such that
with a population just over two million people 34], Botswana is one of the most sparsely populated countries in the world. The Okavango
Delta lies in the north and is the world’s largest inland delta. Most of the population
lives along the eastern border of the country with approximately 40% living in rural
areas. Botswana is comparable in size to France and slightly smaller than Texas in
the USA.

English is the official language of the government, business and education sectors,
however at least 80% of the people speak Setswana, with smaller proportions who speak
Kalanga (8%) and Sekgalagadi (3%) 34]. Fifty-one percent of Batswana are female 34]. The average life expectancy is 53 years for both men and women 35], although with the reduction in infant mortality rate and the government’s increased
access to antiretroviral drugs, life spans are expected to increase in the near future.

The Botswana healthcare system includes public (governmental), private for-profit,
private non-profit and traditional medicine practice, with 98% of health facilities
operating in the public sector 36]. Healthcare services are decentralized to the district level; delivery is based on
the primary health care model; and until 2011, oversight was provided by Village Development
Committees (personal communication, participant 36). The MoH now oversees health services
in the country. There are 14 district hospitals in country, more than 200 health clinics,
at least 330 health posts, and almost 850 mobile posts. Most of the population lives
within 8 km (about 5 miles) of a health facility. At present, health systems principally
provide care for acute episodic conditions and do not tend to chronic care needs,
particularly for the rapidly growing aged population.

Village Shoshong

Shoshong is a rural village of about 10,000 people in the Central District of Botswana,
just north of the Tropic of Capricorn and approximately 250 km (150 miles) north of
the country’s capital, Gaborone (more than 230,00 people) 34]. There are three wards in Shoshong: Phaleng, Bokaa and Kgamane. Wards are generally
areas where residences are clustered around the kgosi’s (chief’s) residence and ward kgotla (traditional village meeting place). Phaleng and Bokaa are the largest Shoshong wards,
each with roughly 35 sub-wards (personal communication, A. Plant). Shoshong is fairly
typical of a large village in Botswana where people maintain a ‘three-home system’
and travel from the village to the lands and to the cattle-posts 37]. There is one tarred road leading into Shoshong from the east; this road bifurcates
at the bus rank into two tarred roads that end up parallel to each other as one travels
west to the next major road (about 10 km) – known simply as ‘the road to Gaborone’
heading south and ‘the road to Serowe’ to the north. The remainder of the 700 square
miles (1813 square kilometers) that make up Shoshong include rugged terrain, corrugated
sandy roads, thick bush, and dry, rocky, red earth. During the rainy season, many
corrugated roads are impassable and a significant proportion of villagers are unable
to receive aid or emergency services.

Study access

Gaining access to Shoshong for this focused ethnography is epitomised by Hammersley
and Atkinson 33], where “(s)ometimes the setting itself comes first – an opportunity arises to investigate
an interesting setting or group of people; and foreshadowed problems spring from the
nature of that setting”. Certainly, it can be argued that, we harnessed investigator
interest in MSK burden with collaborative relationships amongst NGO and university
personnel to develop an ‘opportunistic research strategy’ 38] that would best serve villagers in rural Botswana.

Participant recruitment

People living in and around Shoshong who suffered from or cared for people suffering
from muscle, bone and joint conditions were eligible for this study, including villagers
who had and had not attended the NGO Clinic, health professionals and staff from two
village clinics and one health post, and traditional healers. We used purposive, opportunistic
and snowball sampling to invite participants regardless of age, gender, social status
or physical condition; in this way, we sought information-rich cases to illuminate
the research aims under study 39].

Villagers were approached while the researcher (MH) was ‘hanging out’ in the village
and during several scheduled kgotla gatherings. Other participants were self-selected during participant observations,
informal interviews, and by word of mouth from kgotla presentations, clinic staff meetings, and community events where permission was granted
to ask for volunteers. A color brochure [Additional file 2: Setswana and Additional file 3: English] was available at the three main kgotlas, village clinics, and library, and was offered to potential participants when MH engaged
villagers in conversations during her stay. The brochure aimed to position the researcher as researcher during fieldwork, rather
than as a NGO healthcare provider.

All healthcare providers and staff who worked at the NGO clinic during the time of
fieldwork were invited to participate. We also invited nurses, health education assistants
and health care auxiliaries working in the two Shoshong primary care clinics and one
health post to participate in the MuBoJo project.

Ethical considerations

The Botswana Ministry of Health, Health Research Development Committee granted ethics
approval (HRDC 00735) for this project, as well as approval for our continuing review
request for data analysis. The Botswana HRDC follows WHO ethical review standards
40] and in 2012 Botswana had “the highest proportion of Research Ethics Committees (five)
to population and of number of people trained (four) to population” in 25 African
countries evaluated, despite its ranking seventh and eighth on those parameters, respectively
41]. In addition, we obtained verbal permission to conduct the study in Shoshong from
the traditional authorities, the dikgosi, from the three main wards.

For transparency and to assure village leaders that we were treating people in a respectful
way, we provided the detailed informed consent document (ICD) in Setswana and English
to the dikgosi, the Shoshong Clinic head matron and senior health education assistant, and all villagers
who requested these details [Additional file 4: Setswana and Additional file 5: English]. The ICD outlined the study purpose, procedures, risks and benefits, methods
to ensure privacy and confidentiality, voluntary nature of participation and the right
to withdraw from the study at any time. These details follow Western bioethical principles
on which the Botswana MoH application guidelines are framed. For in-depth interviews,
the researcher and interpreter administered consent orally at the time of the interview
and participants signed the Statement of Consent [Additional file 6: Setswana and Additional file 7: English], which was a one-page summary about the voluntary nature of participation;
that they could stop the interview at any time and not need to give a reason; that
detailed ICDs were freely available for review; and they could ask questions at any
time. There was also a selection where the participant did or did not give permission
for audio recording the interview. For photographs, we obtained separate (recorded)
verbal and/or written permission [Additional file 8: Setswana and Additional file 9: English] for use in scientific presentations and publications related to the MuBoJo
project. We prepared child assent ICDs for potential participants under 18 years;
however, no children participated in the project.

Because the interpreters and transcriptionists may be acquaintances of study participants,
MH conducted individual training sessions about the importance of confidentiality,
the protection of human participants, and complying with secure data transfer protocols
for interview transcripts. MH required all Botswana-based project personnel to complete
the online certification course from the U.S. National Institutes of Health Office
of Extramural Research entitled ‘Protecting Human Research Participants’ 42]; MH has completed annual recertification from this program for the past 10 years.
At the end of the online course, personnel submitted their certificate of completion
for project files and met with MH to discuss course content.

Data collection

Earlier visits to Botswana in late 2011 and early 2013 for five and four weeks, respectively,
afforded opportunities to: establish relationships with government officials and health
care personnel; seek approval from traditional authorities to conduct research in
Shoshong; tour existing healthcare facilities and develop relationships with local
healthcare providers and WSC clinic personnel; build local capacity with Batswana interested in assisting with the project; prepare forward and backward translations
of text documents; and, pre-test the interview and transcription processes in two
languages.

Ethnographic material is drawn primarily from six months of fieldwork during October
2013 to March 2014. Routine efforts to generate daily fieldnotes, transcribe and translate
digital recordings, index and file material, write memos and reflexive notes were
demanding, time intensive activities 33]. Verbal fieldnotes were habitually captured in a digital voice recorder, particularly
during the 40 km commute between the study site and NGO accommodations or during the
two-hour commute to Gaborone to train and work with transcriptionists, transport NGO
volunteers to/from the airport, and build relations with university-based faculty.

The researcher engaged in non-participant observation in natural settings where people
interact with one another, including kgotlas, bus rank, shopping areas, post office, and clinic waiting rooms. Participant observations
included daily work routines in people’s compounds, at the lands and at cattle posts.
In this way, MH was able to interact and participate with people during their work
and leisure time to observe and talk with them about what they were doing, thinking
and saying 43] in relation to muscle, bone and joint health. These data enhanced the content of
interviews and some observations led to in-depth interviews or photographs.

In depth interviews were conducted in settings convenient and comfortable for participants
and, with two exceptions, were audio recorded. With few exceptions, villager interviews
were conducted in Setswana with an interpreter. Interviews with healthcare providers
were typically conducted in English.

Fifty-five interviews were conducted with 34 villager participants in April-May 2013
and between October 2013 and March 2014. Twenty-four villagers were interviewed once,
five were interviewed twice, two interviewed three times, and three villagers were
interviewed four, five and six times, respectively. Ten pre-test interviews ranged
from 10–35 minutes each with an average interview duration of 20 minutes. The remaining
45 interviews ranged from 30–90 minutes with an average duration of 60 minutes. Of
the 34 villager participants, 25 were women and nine were men. The median age of participants
was 57 (range?=?20-97).

Fifteen interviews were conducted with 14 healthcare providers; one provider was interviewed
twice, approximately ten months between interviews. Providers were licensed, certified
or apprenticed as a chiropractor, healthcare auxiliary, health education assistant,
nurse or nurse/midwife, osteopath or traditional healer. Of the 14 providers, 10 were
women and four were men. Ten providers were Motswana and one each were from France, South Africa, Switzerland and the United States. Age
was collected in 10 year increments such that four providers were in their 20s, four
in their 30s, five in their 50s, and one more than age 60. On average, provider interviews
lasted for one hour.

Language strategies

Given the opportunistic research setting with researchers and NGO volunteers who do
not speak Setswana and the aims of this research to reveal the everyday burdens of
MuBoJo conditions, we designed several strategies to work with translators (for written
language), interpreters (for spoken language), and transcriptionists (to transcribe
audio recorded Setswana and provide contextual translations in English).

Written language

During MH’s earlier visits to Botswana, along with relationship building between WSC
personnel and people in Gaborone and the Central District of Botswana, two women agreed
to provide forward translations (English to Setswana) for text documents. Both women
are Botswana nationals, fluent in both languages with Setswana their mother tongue.
One woman in her 50s has diplomas in general nursing and midwifery, a BS in health
sciences, and more than 20 years experience in the public health arena, including
local, national and regional community health efforts; she currently lives in Mahalapye
(about 40 km from Shoshong). The other woman is in her 40s with a BA in Humanities
and post-graduate certificates in AIDS management, monitoring and evaluation; she
has served as a field coordinator for over ten years with a consultant’s agency based
in Gaborone that provides project management services in Botswana. The translators
do not know one another and both prepared independent forward translations for project
text documents, including the informed consent document, photograph permission form,
information sheet, human placard, and semi-structured interview guides.

Next, two Shoshong villagers who were not involved with forward translations conducted
independent backward translations. These villagers also happened to be women, both
in their 20s, and both having completed senior secondary schooling (high school equivalent
in the US) in Botswana. At the time, these young woman were volunteers at the NGO
clinic, serving the capacity of healthcare auxiliary (clinic staff).

Through this process, one set of forward translations emerged as the least complex
translation and best suited for villager understanding. Back translations allowed
us to simplify content in both languages to adapt textual information to suit the
culture 44]. Through discussions with three of the four translators, we agreed that the information
sheet was cumbersome with far too much detail; this was completely revised and designed
as the color brochure [Additional file 2: Setswana and Additional file 3: English] to present simple and consistent messages about the research project.

Spoken language

The translator who lives in Mahalapye was enthusiastic to continue work with the MuBoJo
project and offered her services as an interpreter. During the pre-test phase in April
2013 and after text document translations were complete, MH and the interpreter spent
one week working together. We discussed issues of confidentiality, semi-structured
and open ended questions, and conducting interviews in two languages. We pre-tested
the interview process by conducting 11 patient and provider interviews together; interviews
were audio recorded and the average duration was 20 minutes. On-going dialogue between
MH and the interpreter focused on: the ideas expressed with interview questions, refining
our approach so that questions had relevance in the local setting 44], and techniques to maximise the richness and depth of the data obtained. We reconnected
in October 2013 and worked together for six months during the primary data collection
phase, although there were several days during each month of fieldwork when the interpreter
was unavailable. Thus, a second interpreter was hired during the early days of fieldwork.
This man is a Motswana in his 30s, whose first language is Kalanga and also speaks Setswana and English.
He has served as an interpreter for staff at the WSC clinics in Shoshong and Mahalapye
at various times over the past three years.

Transcribed language: data transcriptions and translations

Before departure in April 2013 MH worked with two independent persons (one female;
one male) who expressed interest to serve as transcriptionists for interviews conducted
in Setswana. The woman from Gaborone had prepared one set of forward translations
for text documents. A serendipitous conversation with a key informant led to a young
man in his 20s from Shoshong who also expressed interest in transcription work. This
man spent his formative years in Shoshong and his teenage years in Gaborone before
entering university in Malaysia; he has a BA in Mass Communication and Broadcasting.

One-on-one training with each potential transcriptionist included discussion about
issues of confidentiality, secure data transfer protocols, and pre-testing the transcription
process. We created a transcription guide [Additional file 10] with examples of notation preferences, particularly for inaudible sections of audio
files and conversations with emotional content 45]. MH transcribed English from two of the pre-test interview audio recordings and inserted
placeholders when Setswana was spoken. Transcriptionists agreed to provide verbatim
Setswana transcription and contextual English translation of dialogue. All parties
agreed to pre-test the transcription process during the next month and then revisit
negotiations for future work.

One transcriptionist, the Shoshong villager, completed two transcripts within two
weeks. Several attempts to reach the second potential transcriptionist were unsuccessful
by the time the MoH ethics application was submitted in June 2013. Hence, one transcriptionist
was listed on the ethics application and we acknowledged that we would need several
transcriptionists to keep pace with interviews during fieldwork. We were optimistic
that university graduate students would be interested in the work once MH was on the
ground for an extended stay.

During fieldwork that commenced in October 2013, MH or a professional transcriptionist
(Way With Words Ltd., London) prepared verbatim English transcriptions and inserted
placeholders for Setswana spoken during the interviews. MH relistened to entire audio
recordings to review and correct vendor-prepared English transcripts before passing
documents to Botswana-based personnel for verbatim Setswana transcription and contextual
English translation. As predicted, one Motswana transcriptionist could not keep pace with the volume of in-depth interview data collected
during fieldwork. Consequently, five Motswana were trained and hired over the course of six months. In the end, three of these
transcriptionists only transcribed seven interviews, four of which were unusable and
re-assigned. Of 56 audio recordings with two languages, two Motswana prepared 22 and 31 transcripts, respectively.

Researcher team profiles

Regardless of the analytic approaches we adopt for this research, personal and professional
identities and interests will inevitably shape how we describe and interpret the data
46],47]. Harnessing these resources was important for planning and conducting the MuBoJo
project; acknowledging the subjectivity of researchers is important for analysis.
All five authors are white and from upper middle class backgrounds.

MH is a second generation Greek-American born in the eastern United States. Her manual
therapy interests date back to the early 1970s when she practiced massage therapy.
She received a graduate degree in chiropractic from the National College of Chiropractic
(now the National University of Health Sciences) in 1989. She earned a master’s degree
in public health (epidemiology) in 1993 and maintained a limited home-office chiropractic
practice for 15 years. She spent 23 years teaching critical appraisal of the literature
and fundamentals of epidemiology in chiropractic curricula, graduate and post-graduate
programs in the US. Since 1990, she has been involved with the design and conduct
of clinical trials in manual therapy, particularly chiropractic. By 2009, her research
interests shifted to qualitative research. Since then, she served as co-facilitator
for several focus groups, in which she assisted with the analysis and publication
of the work 48], and is engaged in analysis of semi-structured interviews with participants from
several clinical trials. In 2012 she enrolled in a doctoral programme at the University
of Southern Denmark (SDU) Faculty of Health Sciences. She embarked on the MuBoJo project
with a passion to provide empirical work for the burden and care of MSK disorders
in underserved communities.

CM is a chiropractor who received his training at Technikon Natal (now Durban University
of Technology) and a doctorally-prepared social scientist (University of Stellenbosch)
who was born, raised and educated in South Africa. After one year in private practice,
he assumed several lecture and administrative roles in the chiropractic department
at Durban University of Technology. He joined SDU as a post-doctoral fellow in 2006
and is now an associate professor in the Clinical Biomechanics Research Unit. He remains
clinically active in private and public settings. His research interests are in the
politics and legislation of chiropractic, as well as in the notion of recovery from
musculoskeletal conditions.

JH is a Danish chiropractor who received his chiropractic degree from Palmer College
of Chiropractic (PCC) in Davenport, Iowa and an epidemiologist who trained at SDU.
He is professor and head of the SDU Clinical Biomechanics Research Unit; he is also
head of the SDU Graduate Programme for Physical Activity and Musculoskeletal Health.
His research focus has been on spinal and MSK pain in the population and he has been
active in Danish national and international task forces and health technology assessment
groups for spine pain, traumatic brain injury and evaluation of MSK research.

SH was born in Canada but spent his formative years in South Africa where he completed
his undergraduate and master’s degree. He completed professional training in North
America, obtaining his chiropractic degree from PCC, and his doctoral (neurophysiology)
and medical degrees from the University of British Columbia. He is a US board certified
neurologist, a fellow of the Royal College of Physicians of Canada, and currently
in clinical practice. He also is the World Spine Care president. His current clinical
and research foci are aimed at providing evidence-based, culturally integrated prevention,
assessment and treatment of spinal disorders in the developing world. He has long
supported interprofessional and multidisciplinary approaches to care for people with
spinal disorders.

HJ is a professor and anthropologist having received her doctoral training at the
University of Copenhagen. She has 30 years experience in academia at the intersection
between anthropology and health sciences. In addition to medical anthropology, her
research interests include qualitative methodology in health research, medical pluralism,
and embodiment of political structures. As a native Dane, her professional trajectory
has included multiple international collaborations with complementary and alternative
medicine investigators.

Data analysis

On-going data analysis is a prominent feature of ethnographic methods, whereby dimensions
important to participants unfold during fieldwork and create iterative loops through
various phases of data collection. Our iterative analytic process is guided by the
‘grounded theorizing’ approach described by Hammersley and Atkinson 33]. With this approach, there is no formula or recipe for ethnographic data analysis,
but it is important to recognise that data management and manipulation are not enough;
rather “(D)ata are materials to think with 33]”.

While analysis continues at the time of this writing, initial stages included reading
and re-reading interview transcripts and fieldnotes, and generating memos for concepts
that made sense of the data. Following close reading of textual data and initial categorical
coding, MH creates substantive codes for each topic and groups these into major themes.
She uses constant comparative methods to identify similarities and differences within
and across cases, returning to the raw data iteratively to review categories and themes,
clarify meaning, and reflect on patterns and connections emerging from the data 33]. When in the field, she consulted with her interpreters and key informants to ensure
that her understanding of the cultural context was the villager’s understanding of
cultural context. Although technological challenges in rural Botswana thwarted communication
with international colleagues during fieldwork, upon returning from the field, MH
regularly reviews and discusses data analysis and interpretation with co-authors.

This structured approach or template organizing style 49] of analysis aims to facilitate pattern finding to reveal insights and meaning about
the burden of musculoskeletal conditions. We are using NVivo 10 (QSR International,
Victoria, Australia) computer software for data storage and retrieval from a large
volume of fieldnotes, transcripts, and memos, to assemble these data in one place
for the interpretive process. The software allows us to organise codes derived from
the data and merge codes into larger categories and themes as the analysis progresses.

Notwithstanding the value of coding these data to identify similarities and differences
between cases for the research questions related to the paths traveled to attend to
MuBoJo conditions and potential comparisons of culture-specific and universal characteristics
of MuBoJo conditions, we are sensitised to the notion that the meaning contexts may
be non-itemizable 50]. Drawing on Biernacki’s argument “that coding frustrates retrieval of the nuanced
meanings that explain action,” we also believe the immersion/crystallization (I/C)
analytic technique will be valuable to unpack MuBoJo burden from these data. I/C was
originally coined by Crabtree and Miller 51] and further elucidated by Borkan 52] and demands ‘personal immersion in the data, by repeated reading of texts, until
insights become apparent’ 53]. Findings and interpretations will be presented and discussed among research team
members to reach consensus about explanatory models for the burden of living with
and caring for others with MuBoJo conditions in Shoshong.

Data quality

Given that the researcher does not speak Setswana and given that most elders in Shoshong
speak little English, we used rigorous and transparent methods in our approach to
gather data. We did not impose preconceived themes on the data collected. Reflexive
reading will locate the researcher as part of the data generated and will seek to
explore her role and perspective in the process of generating and interpreting the
data.

We aimed for credibility and confirmability in several ways. First, we developed an
early familiarity with village culture 54] during visits in 2011 and 2013 before designing the study and entering the field
for primary data collection. Second, we provide in this report, a detailed description
of the context and methods used with this focused ethnography. Next, we are in the
midst of prolonged engagement with the texts and coding process, including debriefings
with experienced research mentors. For fieldnotes and transcripts generated during
fieldwork, the researcher consulted key informants fluent in both languages to reduce
misinterpretations of the data. Further on, we sought a wide range of informants to
confirm or disconfirm patterns that emerged during data collection. Finally, on-going
reflections about the nature of this journey with villagers adds to the contextual
framework for analysis and for future World Spine Care work in Botswana. We have no
reason to believe that Shoshong villagers, who suffer from or care for others living
with muscle, bone and joint conditions, differ from villagers in other rural Botswana
communities. Readers will need to determine the transferability of findings to other
settings.