Introduction

Over the last twenty-five years, anthropologists have provided important critical insights into experimental HIV-related research. They have explicated hidden cultural transcripts and unintentional ‘social side effects’ of HIV interventions (Taylor 2007); raised ethical concerns about recruitment, consent, and confidentiality in cross-cultural contexts (Molyneux and Geissler 2008; Reynolds et al. 2013); questioned the growing commercialization of experimental science; and highlighted the power relations embedded in HIV interventions that are carried out among economically and politically marginalized people (Biehl 2007; Epstein 1996; Farmer 2001; Nguyen 2010).

Reflecting on anthropologists’ different modes of engagement with biomedicine during the AIDS epidemic, Hardon and Moyer (2014) differentiate between subaltern alignment, reflexive dialogue, and collaborative design (note these modes are not limited to HIV research, see also Hardon 2010). ‘Subaltern alignment’, in which anthropologists reveal what is at stake for people who have little power to influence global health efforts, is arguably the most common and comfortable approach for anthropologists. Two more active forms of engagement are ‘reflexive dialogue’, in which the anthropologist does not align solely with the subaltern but engages with different actors, and ‘collaborative design’, in which the anthropologist does not only enter into dialogue but also participates in developing interventions with, for example, people affected by the HIV epidemic and/or biomedical practitioners (Hardon and Moyer 2014).

Anthropologists’ collaboration in experimental HIV research can take different forms, such as: executing preliminary qualitative explorations to inform interventions, instruments or designs of trials; conducting field studies and observational studies in parallel with trials in order to increase external validity, among other goals; and undertaking critical ethnographies of ‘trial communities’ and investigating practices of knowledge production (Geissler 2011; Le Marcis 2015; Pool et al. 2010; Reynolds et al. 2016). Some have argued though that with the increased biomedicalisation of HIV, much anthropological research has shifted toward applied qualitative health research and away from ethnographic fieldwork characterised by ‘deep hanging out’ (Moyer 2015, 261). So far there are few accounts from anthropologists (and other social scientists) working within intervention-oriented HIV research that critically reflect on what is going on in collaborative design and how it is done (Montgomery and Pool 2011), and whether it can be mutually constructive for anthropologists and their collaborators (see Elliott and Thomas, this issue, for a discussion on collaborative partnerships between anthropologists and epidemiologists). In this article, I draw upon my own experience of a collaborative research project in Swaziland to reflect on my attempts to simultaneously conduct ‘insider ethnography’ as a doctoral student and applied qualitative research as the social science coordinator of the project I was studying.

Insider ethnography

‘Insider ethnography’ is often multipositioned as well as multisited (Marcus 1995). This means that the anthropologist is not a detached observer but rather participates as an active (and possibly paid) member of the institutional alliances and interventions under study. This may include being involved in the coproduction of research designs, tools, and interventions, and it necessitates that the anthropologist explores rather than conceals the personal connections and contributions one has as an actor in the social processes that are being analysed (Mosse 2005). Feminist scholars have long advocated for analysing the positionality of the researcher and taking account of the situatedness of knowledge (Haraway 1988). Such analysis posits that the power dynamic between the researcher and researched shapes how knowledge is created and encourages explicit documentation of the ‘micropolitics of research projects’ (Conti and O’Neil 2007).

Accounting for one’s own subject position has received attention in various fields of ethnographic inquiry, most notably in science and technology studies and development studies, in which methodological challenges of ‘studying up’ are discussed when studying (global) elite or powerful epistemic communities (Marcus 1983; Nader 1972; Ortner 2010). In a personal reflection, Forsythe (2001) discusses the ethical and political dilemmas she experienced as an insider studying the world of artificial intelligence in the Unites States. Despite problems of data ownership and powerful informants who were able to contest how they were represented, Forsythe was largely successful in her attempt to demonstrate that ethnographic approaches can be usefully applied to analyse informatics. This, ironically, placed her in direct competition with her informants who became (more) successful at appropriating funds for doing ‘ethnography light’, without any formal training or understanding of ethnographic theory (Forsythe 2001, 136).

From a methodological perspective, Erikson (2011, 27) remarks that when the anthropologist’s attention goes ‘macro’ ‒ when studying government agencies and corporate boardrooms ‒ people often go missing; scholars tend to employ fuzzy concepts, using ‘words like flows,circulation and processes to describe global phenomena’, instead of providing rich ethnographic detail of social situations and the actors involved. When writing from an insider perspective in global health, this rich ethnographic detail becomes difficult to include because it tends to destabilise the interpretation of events, something that organisations are keen to control (Mosse 2005). Besides directly censoring critical insider accounts by objecting to publications (see for example Mosse 2006), insider ethnography can be influenced by more subtle forms of disciplinary power (Foucault 1979) that structure behaviour during fieldwork (and writing), such as when researchers internalize when to speak or when to remain quiet, without the coercion of others. Moreover, multiple and fluid alliances across different settings characterise complex global health projects and therefore, when a project’s boundaries are porous and ambiguous, insider ethnography is more contested. It may therefore be necessary to occupy different subject positions to be accepted into different field situations (for example health facilities, study investigator meetings, and community events).

In this article I reflect on my multipositionality in the MaxART [note 1] project. Having worked in the project over the last five years I have familiarised myself with the public health discourse of CD4 counts, retention-in-care, and stepped-wedge designs. In fact, I have become familiarised to such a degree that I sometimes struggle to write ‘anthropologically’ about the project and must either forgo discussing the technical details or frame them as discursive objects in the making. I also gained new managerial skills through my work as the social science coordinator of the overall project, such as producing standard operating procedures [note 2] and logistical frameworks. Additionally, I acquired a great deal of tacit knowledge about working in a public health project, having embodied part of its practices and language. As time passed, however, these gains made it difficult for me to distance myself because of the different loyalties that emerged. In the following sections, I describe my social interactions with study team members and health facility staff. I discuss how I balanced my role as an ‘insider’ who was on the study team with my role as a doctoral student researching the team and facility staff. I demonstrate that this balancing act was not only an internal process but a relational and intersubjective practice. I show this by discussing three ways of ‘making space’: 1) through writing a doctoral research protocol that provided the flexibility to study events as they were unfolding and to ask critical questions, 2) by variously aligning with and detaching from the study team to avoid being seen as one of the study’s ‘implementers’, and 3) by taking moral distance from certain decisions related to the study’s design and implementation. 

Changing protocols

In 2011 the MaxART project received 8.8 million euro from the Dutch Postcode Lottery Dream Fund to, as the project’s original proposal states: ‘Virtually Eliminate New HIV Infections in Southern Africa … by “breaking the back” of the epidemic in Southern Africa by using a tool the world already has at its fingertips: anti-retroviral therapy’ (SAN! and CHAI 2010, 3).

Swaziland, a small country of around 1.1 million people with a HIV prevalence estimated at 26 percent among fifteen- to forty-nine-year-olds (MOH 2012), was chosen as the test location for the project. Stop AIDS Now! (SAN!) and the Clinton Health Access Initiative (CHAI), who jointly wrote the proposal, argued that if they could ‘demonstrate that the epidemic can be halted in Swaziland, the impact would be felt and multiplied around the world’ (SAN! and CHAI 2010, 3). The project entailed two phases. In the first phase, the focus was on implementing several interventions to increase access to HIV testing and antiretroviral therapy (ART) throughout the four regions of the country under the current HIV-treatment guidelines.[note 3] The second phase was originally conceptualised as the HIV Treatment as Prevention ‘Proof of Concept’ Study, which would test the feasibility and estimate the impact of ‘early’ [note 4] ART initiation on reducing new infections (SAN! and CHAI 2010, 32).

As the Treatment as Prevention study was designed, I conducted qualitative research to inform the team about the current state of HIV testing, counselling, and treatment in the facilities that would participate in the study. I designed this research with a local advisor from the Swaziland Network of People Living with HIV. We shared similar ideas about participatory ways of conducting research and involving people living with HIV in the design of interventions. Together, we trained four Swazi researchers and conducted three months of observations, interviews, and focus group discussions in eight different facilities and surrounding communities. We assessed the understanding of ART as prevention amongst health workers and people living with HIV, inquired about the key requirements they felt should be in place for the implementation of immediate ART initiation, and investigated the current state of HIV testing and counselling and ART service delivery.

In July 2013, we presented the preliminary findings at the biannual ‘Face-to-Face’ meeting, when all eight organisations of the MaxART consortium convened in Swaziland, and we invited about twenty study participants (people living with HIV, health providers, and administrators) to attend the presentation. During the presentation, one of the hospital administrators asked the principal investigator why the planned study was called ‘Treatment as Prevention’, which, in her opinion, was an ambiguous term and could lead to ART being seen as a substitute for condom use. The principal investigator, a medical doctor from the Swaziland Ministry of Health, responded by indicating that the ministry was also unhappy with the terminology, and arguing that the study should be called ‘Treatment for All’ (for a more detailed discussion this terminology see Vernooij et al. 2016). One of the conclusions the MaxART consortium drew from our presentation was that there was need for developing tailor-made health messages to explain the benefits of early ART for different groups: health providers, people living with HIV, and the general public. I became interested in studying the social processes of knowledge production and translation in relation to the project’s focus and the evident concerns about correct ‘messaging’. I therefore wrote a separate PhD proposal to study the interactions between different actors involved in the design and implementation of messages as an insider in the project.

At the same time, a public health expert from CHAI was taking the lead in writing the protocol for the Treatment as Prevention study, later renamed Early Access to ART for All (EAAA), and between December 2012 and September 2014, twenty-nine versions of the protocol were developed. During this period, I received seven versions on which to provide feedback and to revise our social science sections, in line with the changing study design. My qualitative study with the Swaziland Network Of People Living with HIV/AIDS had placed me in a more powerful position in the consortium because I had acquired experiential knowledge of the study facilities and relationships with the staff there, which were of particular interest to CHAI, and I was able to use that knowledge to try to influence parts of the study design.

When the EAAA study finally started in September 2014 it was no longer a proof-of-concept study to test the impact of early HIV treatment on reducing new infections. Instead it was defined as an ‘implementation study designed to answer critical implementation questions (i.e., acceptance and retention among patients initiated on ART at higher CD4 thresholds) and determine the “real world” potential of this new prevention intervention’(MaxART EAAA 2015, 14). In practice, this meant that the study team was tasked to implement the intervention in the fourteen participating health facilities, not as a study but rather as a new approach to HIV care. At the same time, the study had to produce usable and reliable data to inform the ‘scale-up’ of early ART in the rest of the country and several standardisation methods, such as including randomisation and blinding to make the study more ‘robust’.

Performing alignment and detachment

Vincanne Adams (2016, 23) has recently described ‘research as intervention’ as the key characteristic of global public health today, writing that interventions ‘must produce not just health but also statistically robust data that can be powered, subjected to chi-square or linear regressions, and then used for scaling up’. The discrepancies between doing research and doing implementation created ethical dilemmas for members of the study team and shaped my interactions with the rest of the team, as the following field note excerpt illustrates.

22 August 2014

It is a Friday afternoon in a government health clinic in northern Swaziland. Nine facility staff members, three study team trainers and I sit in the cramped tea room of the clinic for the training of the Standard Operating Procedures of the EAAA study which will commence next week. The room gets filled with the smell of Kentucky Fried Chicken, the customary lunch brought to health facilities whenever there are on-site trainings. When the trainer is explaining to the health care staff which steps to follow when enrolling people living with HIV into the study, another study team member, a representative from the Ministry of Health, corrects her and tells the staff they should not explain the project as a study to patients. ‘People don’t like to be studied’, she says. Rather she argues – as it is also written in the patient information guide – they should mention that the project is a new approach to HIV care. The trainer’s supervisor, who oversees the clinical procedures of the study in the 14 participating facilities, questions whether this is ethical. She looks at me. Nobody responds.

At this point, I had been doing observations of HIV counselling and care practices in this clinic for four months, before the study started. I was interested to learn how the study would affect HIV-prevention messages given to people living with HIV and how the staff would incorporate the study into their daily routines. I divided my time, location, and loyalties between my rural field site, where I shared a house with one of the staff members from the facility, and the CHAI office, a two-hour drive from my house, where I participated in weekly study team meetings. When the clinical study team came from Mbabane to the clinic for trainings I tried to participate as an observer in the training to avoid being seen as one of the study’s ‘implementers’, because I feared my participation would affect the staff’s perception of me and thereby their willingness to share their views. However, just a few weeks previously, during a lunch meeting with the clinical supervisor and a study team manager, I had tried to convince them to talk to the principal investigator of the Ministry of Health to change the procedures to inform people they were enrolling in a study about starting ART early, and to obtain their consent.

I remained quiet when the question about ethics came up during the training at the clinic, as I wanted to see how the health workers would react. However, I did not learn this by observing the training sessions, but rather by spending time with the clinic staff and hanging out in the clinic’s dispensary. Because the dispensary is the only room in the clinic that has air-conditioning the staff usually sit there on a hot day when there are no patients. It is also the place where I could make myself useful – helping to count pills and placing them into small plastic bags of pre-packed medications – whilst doing ‘participant observation’. In the following two excerpts from my field notes, I describe how I negotiated my different subject positions in the field(s).

29 August 2014

It is the day before the EAAA study starts. I am sitting in the dispensary with the nurse-in-charge, the lay counsellor, one of the expert clients, and one orderly. A study team member walks in to drop off the study enrolment stickers and forms, and the blood collection tubes for the additional viral load and resistance test to be done for study participants to generate evidence for one of the study’s primary outcomes: viral suppression. She quickly explains the study enrolment procedures to the nurse-in-charge and leaves.

After the study team member left the staff starts discussing in siSwati how they should explain ‘viral load test’ to patients, since the test is not routinely available in government facilities in Swaziland. Noticing that I am trying to follow their conversation, the nurse-in-charge turns to me and says in English, ‘we are talking about the consent now, I’m worried everyone will be consenting into the study’. She explains that it is uncommon in these rural areas for a patient to refuse an offer from the health provider. Patients fear that this might damage the relationship with the provider. She asks me whether she understood correctly that they should not ask for consent for starting ART early, but just for the extra lab tests performed. I confirm this. Two other staff members walk in and we have a long discussion about their concerns about the study, which are mainly related to being understaffed and the fear that the study increases their workload without adding extra personnel. I ask why they did not mention any of these concerns to the study team during the training last week. One of the nurses explains that in such trainings there is not much room for discussion. The presence of the Ministry of Health during the training carries weight and since this is a government clinic they have to respect their authority. Furthermore, the study was going to start in a week, how much room for changes was there? I encourage them to share their concerns about the study with the clinical study team. One staff member gets enthused and makes a plea that if they want to change something, this is the time to act, not after the study has started.

2 September 2014

It is the first day of the study. I am in the CHAI boardroom for a study meeting. The clinical team is celebrating that thirteen people enrolled into the study today. The clinic in which I do fieldwork is a topic of concern, they have not enrolled anybody yet, and they have called today to ask for a meeting to discuss their concerns. I don’t say anything but after the meeting I tell the clinical supervisor that I encouraged the staff to share their concerns. ‘Ah, so you are the culprit!’, she says. A few days later the supervisor goes with a representative from the Ministry of Health to the clinic to discuss their concerns.

My attempts to disassociate myself from the study team at times, trying not to influence the situation at one moment but nonetheless stirring up debate in the next, illustrate my navigation between trying to be accepted in both field sites, the study team and clinic. I tried to be part- study-team-insider during investigator meetings and part-study-team-outsider when they visited the facility to train and oversee the study implementation. One reason for this was methodological: I did not want the clinic staff to perceive me as a representative of the study team, which could have influenced my rapport with the staff. Another reason was moral: I found it difficult to support certain aspects of the study, such as blinding the study team and participants to the timing of the intervention, which I had tried to argue against in previous meetings during the study design process. It was repeatedly argued during study team meetings that randomization and blinding would make the study more ‘robust’. Since the fourteen facilities were not randomly selected, but instead were selected to avoid overlap with another large implementation study going on, I questioned whether sequence randomization affected the robustness or power of the study. A final reason was epistemological, as I was concerned with the manner in which knowledge was being obtained. Instead of supporting efforts to control behaviour by blinding, I argued in favour of a participatory approach that would involve health care providers and ‘community mobilisers’ in discussions of why it was important that the intervention not be offered during the control phase of the study.

Moral entanglements

During the time the study was being (re)designed I emailed my PhD supervisor (who was a coinvestigator in the MaxART study) and described the proposal to blind participants (people living with HIV), health care providers and the study team about the intervention start date and which facilities would deliver it as ‘ridiculous’. She advised that I take the proposition as seriously as I did traditional healers’ stories about being trained in underwater worlds by creatures who were half-man and half-horse, which I had written about without normative undertones. Like many anthropologists, I was finding it more complicated to employ (cultural) relativism with ‘bureaucrats, politicians and corporate types’ (Erikson 2011, 28). It was more challenging to suspend judgment when seeking to unravel the specific social context that produced particular scientific norms and values (such as blinding and standardization) than when trying to understand the life-worlds of traditional healers.

Part of this was linked to my stake in the project. Because of limited funding, we could only conduct our ‘applied’ mixed-methods research in the facilities where I had previously conducted qualitative research with the Swaziland Network of People Living with HIV. Randomizing and blinding would make it impossible to select the facilities where we had worked before and plan data collection. Furthermore, several months before the discussions of randomization and blinding unfolded, I had started my fieldwork at one of the study sites. It was therefore decided that my field site could be taken out of the randomization so that I could continue my research there, the site was selected to offer the intervention first, and I was unblinded in order to be able to plan our data collection in advance. The rest of the facilities were randomized and the facility staff, community leaders, and (part of the) study team were blinded from knowing when the intervention started until six weeks before.

Although I was just as much part of the study team as, for example, the clinical supervisor was, I tried to align myself more with the study participants (health care providers and people living with HIV) than with the study team, as I had done previously with the Swaziland Network of People Living with HIV. However, this was not the type of alignment that was expected of an ‘insider’; in study team meetings, a collective sentiment was expressed that ‘we are all in this together’ and that it was important that we all speak in one voice about the study to the outside world. From the start of the study, my two research assistants and I were expected to share information from our observations in order to improve the study’s implementation. Because we were dependent on the study team for access to conduct our observations and interviews, as time passed I felt less able to openly critique several practices, for example, the informed consent procedure, because of these conflicting loyalties.

By the end of my fieldwork, in February 2015, it had become confusing for some that I was both the social science coordinator and simultaneously studying the study’s implementation. When I asked to be part of a training in the clinic where I had previously done fieldwork, the trainer first preferred not to have me present. She asked me: ‘Did you not say yourself that you wanted to be seen as separate from the clinical team?’ It was important to her, and to the consortium at large, that we present a collective voice during the training sessions. During the training sessions, if I asked – instead of answering – questions this would come across as if we were not on the same page. I explained that I wanted to participate as an observer and was allowed to do so. The conflicts between my roles followed me beyond the field sites, as this excerpt shows:

19 February 2015

It’s about 10 o’clock on a Thursday morning. My housemate, who works as an HIV counsellor in the clinic, asks me for help. Can I tell her how she should explain the ‘Early Access to ART for All’ in the daily morning health education sessions? By this time the facility had ‘transitioned’ to offering early ART as the new standard of care in the facility. I am hesitant to answer and ask her if the study’s clinical mentoring team had not trained her how to explain this to patients. My housemate looks puzzled and says: ‘That training was only one day, you work in the project, why won’t you tell me, what is the problem?’

I resisted telling my housemate how to explain the study (or new standard of care, as it was to be described to patients) for the same reason I did not want to speak out during the training: I did not agree with how the study’s benefits were supposed to be explained to patients. At the same I did not want (nor did I feel allowed) to assert my opinion about this to the health care providers or other ‘outsiders’. My proximity to the study team thus created certain expectations when interacting with study team members and clinic staff. As the field note excerpts show, when I decided to distance myself, by not participating or not speaking in one collective voice, this sometimes produced frustration amongst the study team and clinic staff, and a sense that I was being disloyal.

Conclusion

Throughout my involvement in MaxART, I have been balancing active participation in a collaborative project with trying to maintain critical distance as an ethnographer. During my fieldwork, my position changed from that of a relative outsider, aligned with the Swaziland Network of People Living with HIV, to that of a study insider, concerned with maintaining access to multiple sites and occupying different subject positions in order to access them. By participating in study meetings, contributing to the design of consent forms, and attending messaging workshops, I became a coproducer of the study, which created expectations on the part of the study team and clinic staff. This involvement turned maintaining a degree of distance into a matter of disloyalty. During my time in MaxART, I shifted between the three modes of engagement (subaltern alignment, reflexive dialogue, and collaborative design) (Moyer and Hardon 2014). Adopting these different forms of engagement within the same project, however, caused others to view me with ambivalence. At the end of my fieldwork, the dominant subject position allotted to me was that of social science coordinator working for MaxART. The tensions described in this article illustrate the distinctly performative mode of knowing implied in ethnographic fieldwork that links understanding to participation (Hastrup 2004, 464).

I am still negotiating this multipositionality, albeit from a distance as I have returned home to the Netherlands. Although removed from the everyday intricacies of study implementation, I remain involved in weekly phone calls with research assistants, as well as email exchanges and conference calls with study managers to talk about data collection, preliminary findings, and report writing. Indeed, when conducting such insider research, it is often the social exit rather than the entry that produces a significant shift in relationships (Mosse 2006). Moreover, it is often the act of writing and transforming one’s experience into public knowledge that can disturb and break fieldwork relations (see Hastrup cited in Mosse 2006, 951). This article is an effort to gain (more) distance from my concern with the study’s implementation through the process of writing. However, my experience of insider research shows that, rather than being achieved post hoc through writing, the processes of alignment and detachment are a continual practice, navigated throughout the stages of ethnographic fieldwork as one’s position shifts.

Balancing proximity and distance is a key effort in every anthropological encounter. But conducting ethnographic research in a highly political field such as global health highlights particular methodological and ethical challenges regarding how anthropologists relate to power (Marcus 2008). Being part of a collaborative project requires the anthropologist to contribute during the project; ‘deep hanging out’ is therefore unlikely to meet other participants’ expectations and assumptions about the anthropologist’s role in their interactions. In response, one can take up different roles in the project and contribute through more ‘applied’ research that serves the agenda of the project (which may not necessarily be the agenda of the critical ethnographer). But, to grasp the social and political processes through which global health projects are made and transformed in practice, anthropologists must make space to produce critical ethnographies of global health.

Doing ethnography ‘of’ global health, Pigg (2013, 132) argues, requires that the ethnographer ‘linger on after the consultant’s site visit’ and ‘pay attention to what falls out of view or falls between the cracks’. For me, this ‘lingering on’ – for example in the clinic after the study’s trainings were finished – became increasingly difficult with time as the distinction between being the study team coordinator and a doctoral student started to fade. The three ways of ‘making space’ were strategies to deal with the challenges of doing ethnography ‘of’ and ‘in’ global health simultaneously. Besides the methodological aspects of making space by having a separate PhD protocol and by aligning and detaching myself from the study team, I struggled with morally supporting certain aspects of the ‘research as intervention’ strategy. My position as insider presented a challenge for ethnographic knowledge production based on principles of openness to and empathy for the ‘other’ (Armbruster and Laerke 2008), as I was creating distance rather than seeking closeness.

Doing insider ethnography may yield a specific kind of ethnographic knowledge. The results may be less theoretical and critical because of self-censoring, collaborators’ objections, or insufficient time. Yet insider accounts that are more attuned to what is at stake for powerful collaborators may be powerful in their own terms too. Anthropologists have the ability to influence research projects in the making, but this depends on our presence in and the level of influence on the networks and institutions that create and implement health policy. The actual effect that the anthropologist might have on the project outcome is also related to the level of flexibility of the project and other people involved. Today, many anthropologists are involved in global health research, and more reflexive accounts of insider ethnography are needed to guide students and budding researchers on positioning oneself and dealing with expectations in collaborative research. Further exploration is needed about intersubjectivity in practice: How does the anthropologist engage in collaborations with differently positioned actors without (being perceived as) ‘taking sides’? Are the ethical principles of commitment and openness in anthropology being applied when ‘studying up’ or doing a critical ethnography of global health? How to combine moral engagement with (scientific) detachment? 

Acknowledgments

I am grateful to the MaxART consortium for their acceptance and support of my PhD research. I wish to thank Eileen Moyer, Ria Reis, Catherine Montgomery, and Tanja Ahlin for commenting on earlier drafts of this paper and Christopher Pell and Erin Martineau for their editorial assistance. Financial support was received from the Amsterdam Institute for Social Science Research.

About the author

Eva Vernooij is a doctoral candidate in the anthropology department at the University of Amsterdam. She is writing her dissertation on the production and transformation of knowledge practices about HIV treatment as prevention in Swaziland. She is a lecturer in the Department of Theory and History of Psychology at the University of Groningen.

References

Adams, Vincanne, ed. 2016. Metrics:What Counts in Global Health. Durham, NC: Duke University Press.

Armbruster, Heidi, and Anna Laerke, eds. 2008. Taking Sides: Ethics, Politics and Fieldwork in Anthropology. New York: Berghahn Books.

Biehl, João. 2007. Will to Live: AIDS Therapies and the Politics of Survival. Princeton, NJ: Princeton University Press.

Conti, Joseph A., and Moira O’Neil. 2007. ‘Studying Power: Qualitative Methods and the Global Elite’. Qualitative Research 7, no. 1: 63–82. https://doi.org/10.1177/1468794107071421.

Epstein, Steven. 1996. Impure Science: AIDS, Activism and the Politics of Knowledge. Berkeley: University of California Press.

Erikson, Susan L. 2011. ‘Global Ethnography: Problems of Theory and Method’. In Reproduction, Globalization and the State, edited by Carole H. Browner and Carolyn F. Sargent, 23–37. 1st ed. Durham, NC: Duke University Press.

Farmer, Paul. 2001. Infections and Inequalities: The Modern Plagues. Berkeley: University of California Press.

Forsythe, Diana E. 2001. ‘Ethics and Politics of Studying up in Technoscience’. In Studying Those Who Study Us: An Anthropologist in the World of Artificial Intelligence, edited by David J. Hess, 119–131. 1st ed. Stanford, CA: Stanford University Press.

Foucault, Michel. 1979. Discipline and Punish: The Birth of the Prison. New York: Vintage Books.

Geissler, Paul W. 2011. ‘Studying Trial Communities: Anthropological and Historical Inquiries into Ethos, Politics and Economy of Medical Research in Africa’. In Evidence, Ethos and Experiment: The Anthropology and History of Medical Research in Africa, edited by Paul W. Geissler, and Catherine Molyneux, 1–28. 1st ed. New York: Berghahn Books.

Haraway, Donna. 1988. ‘Situated Knowledges: The Science Question in Feminism and the Privilege of Partial Perspective’. Feminist Studies 14, no. 3: 575–99. https://doi.org/10.2307/3178066.

Hardon, Anita. 2010. ‘From Subaltern Alignment to Constructive Mediation: Modes of Feminist Engagement in the Design of Reproductive Technologies’. In Feminist Technology, edited by Linda L. Layne, Sharra L. Vostral, and Kate Boyer, 154–178. Urbana: University of Illinois Press.

Hardon, Anita, and Eileen Moyer. 2014. ‘Anthropology of AIDS: Modes of Engagement’. Medical Anthropology 33, no. 4: 255–62. https://doi.org/10.1080/01459740.2014.889132.

Hastrup, Kirsten. 2004. ‘Getting It Right: Knowledge and Evidence in Anthropology’. Anthropological Theory 4, no. 4: 455–72. https://doi.org/10.1177/1463499604047921.

Le Marcis, Frederic. 2015. ‘Life Promises and “Failed” Family Ties: Expectations and Disappointment within a Clinical Trial (Ivory Coast)’. Anthropology & Medicine 22, no. 3: 295–308. https://doi.org/10.1080/13648470.2015.1081671.

Marcus, George E., ed. 1983. Elites: Ethnographic Issues. Albuquerque: University of New Mexico Press.

Marcus, George E. 1995. ‘Ethnography in/of the World-System: The Emergence of Multi-Sited Ethnography’. Annual Review of Anthropology 24: 95–117. https://doi.org/10.1146/annurev.an.24.100195.000523.

Marcus, George E. 2008. ‘The End(s) of Ethnography: Social/Cultural Anthropology’s Signature Form of Producing Knowledge in Transition’. Cultural Anthropology 23, no. 1: 1–14. https://doi.org/10.1111/j.1548-1360.2008.00001.x.

MOH (Ministry of Health). 2010. National Comprehensive HIV Package of Care for Adults and Adolescents in Swaziland. Mbabane: Ministry of Health.

MOH. 2012. Swaziland HIV Incidence Measurement Survey (SHIMS). Mbabane: Ministry of Health.

Molyneux, Sassy, and Paul W. Geissler. 2008. ‘Ethics and the Ethnography of Medical Research in Africa’. Social Science & Medicine 67, no. 5: 685–95. https://doi.org/10.1016/j.socscimed.2008.02.023.

Montgomery, Catherine M., and Robert Pool. 2011. ‘Critically Engaging: Integrating the Social and the Biomedical in International Microbicides Research’. Journal of the International AIDS Society 14, Suppl 2: S4. https://doi.org/10.1186/1758-2652-14-S2-S4

Mosse, David. 2005. Cultivating Development: An Ethnography of Aid Policy and Practice. New York: Pluto Press.

Mosse, David. 2006. ‘Anti-Social Anthropology? Objectivity, Objection, and the Ethnography of Public Policy and Professional Communities’. Journal of the Royal Anthropological Institute 12, no. 4: 935–56. https://doi.org/10.1111/j.1467-9655.2006.00371.x.

Moyer, Eileen. 2015. ‘The Anthropology of Life After AIDS: Epistemological Continuities in the Age of Antiretroviral Treatment’. Annual Review of Anthropology 44: 259–75. https://doi.org/10.1146/annurev-anthro-102214-014235.

Nader, Laura. 1972. ‘Up the Anthropologist: Perspectives Gained from Studying Up’. In Reinventing Anthropology, edited by Dell H. Hymes, 284–311. 1st ed. New York: Pantheon Books.

Nguyen, Vinh-Kim. 2010. The Republic of Therapy: Triage and Sovereignty in West Africa’s Time of AIDS. Durham, NC: Duke University Press.

Ortner, Sherry B. 2010. ‘Access: Reflections on Studying Up in Hollywood’. Ethnography 11, no. 2: 211–33. https://doi.org/10.1177/1466138110362006.

Pigg, Stacy L. 2013. ‘On Sitting and Doing: Ethnography as Action in Global Health’. Social Science & Medicine 99: 127–34. https://doi.org/10.1016/j.socscimed.2013.07.018.

Pool, Robert, Catherine M. Montgomery, Neetha S. Morar, Oliver Mweemba, Agnes Ssali, Mitzy Gafos, Shelley Lees, et al. 2010. ‘A Mixed Methods and Triangulation Model for Increasing the Accuracy of Adherence and Sexual Behaviour Data: The Microbicides Development Programme’. PloS One 5, no. 7: e11600. https://doi.org/10.1371/journal.pone.0011600.

Reynolds, Lindsey J., Carol S. Camlin, Norma C. Ware, and Janet Seeley. 2016. ‘Exploring Critical Questions for the Implementation of “Universal Test and Treat” Approaches to HIV Prevention and Care’. AIDS Care 28, Suppl 3: 1–6. https://doi.org/10.1080/09540121.2016.1178960.

Reynolds, Lindsey, Thomas Cousins, Marie-Louise Newell, and John Imrie. 2013. ‘The Social Dynamics of Consent and Refusal in HIV Surveillance in Rural South Africa’. Social Science & Medicine 77: 118–25. https://doi.org/10.1016/j.socscimed.2012.11.015. 

SAN! and CHAI (STOP AIDS NOW! and Clinton Health Access Initiative). 2010. Ending New HIV Infections in Swaziland: A Catalytic Model for Southern Africa. Amsterdam: Stop AIDS NOW!

Taylor, Julie J. 2007. ‘Assisting or Compromising Intervention? The Concept of “Culture” in Biomedical and Social Research on HIV/AIDS’. Social Science & Medicine 64, no. 4: 965–75. https://doi.org/http://dx.doi.org/10.1016/j.socscimed.2006.10.030.

Vernooij, Eva, Mandhla Mehlo, Anita Hardon, and Ria Reis. 2016. ‘Access for All: Contextualising HIV Treatment as Prevention in Swaziland’. AIDS Care 28, Suppl 3: 7–13. https://doi.org/10.1080/09540121.2016.1178954.