HIV testing attitudes and practices amongst 'wealthy men': qualitative evidence from Tanzania

Abstract Evidence from nationally representative surveys conducted in sub-Saharan Africa shows that significant proportions of men in the wealthiest quintile report never having tested for HIV. Despite high prevalence rates in this quintile, no research has been conducted on the HIV testing attitudes and practices of wealthier men. This article reports findings from qualitative research conducted with 23 wealthy men in Tanzania. Whilst wealthy men reported barriers to and enablers of HIV testing previously reported by the general population, concerns around loss of social status and community standing were amplified for members of this demographic. Furthermore, HIV stigma among members of this group remains high. However, enhanced access to HIV testing through private clinics, regular healthcare appointments, health insurance schemes and the means to travel to other countries enables wealthy men to avoid stigma. In settings such as the workplace, wealthy men were able to test in public in their roles as ‘leaders’ to encourage others to test. Future interventions to increase testing amongst men should target settings in which these leadership roles can be taken advantage of. HIV services also need integrating into the health system to remove the need for testing and treatment to be accessed at separate clinics.


Introduction
HIV transmission in sub-Saharan Africa has to-date been primarily viewed through a poverty lens in both policymaking and academic spheres (Long and Deane 2015). Early narratives concerning transmission focused on lack of information, engagement in sex work and lack of access to and use of condoms (Prual, Chacko, and Koch-Weser 1991), a theme that has persisted in HIV research and policy-making. This is, in part, due to initial assumptions that those who are wealthy would change their behaviour first, due to higher levels of education, incentives related to enjoying their living standards, and an overall belief they were more likely to respond to the epidemic in a 'rational' way (Bujra 2006). However, the poverty narrative is at odds with evidence from nationally representative Demographic Health Surveys (DHS) from numerous sub-Saharan countries that show the wealthiest often have higher rates of HIV, and across the continent HIV does not disproportionately affect the poor (Mishra et al. 2007;Parkhurst 2010). For a range of reasons, this evidence has not impacted the policy and research agenda, with few, if any, interventions targeted at those of higher economic status. The wealthy are, with some exceptions (see e.g. Jangu, Tam, and Maticka-Tyndale [2017] and Aggleton, Bell, and Kelly-Hanku [2014]), also absent from social science HIV research.
This poverty narrative has also been implicitly applied to the issue of HIV testing. For example, behavioural experiments designed to calculate the amount of money needed to incentivise HIV testing make it clear who the target population is when the sums involved equate to as little as one or two US dollars (see, e.g. Osterman et al. [2015,1] and their article entitled "Would you take a test for 5,000 shillings?"). Unlike the work on HIV transmission, this is more reflective of the epidemiological evidence, with the DHS data showing a positive correlation between HIV testing and household wealth, a trend that is consistent across sub-Saharan Africa (Staveteig et al. 2013). In Tanzania, the most recent data shows that the proportion of men reporting having taken an HIV test increases across wealth quintiles; in the lowest wealth quintile, 49.5% report having taken a HIV test in comparison to 69.3% in the highest wealth quintile (TACAIDS 2018). Whilst this data may be interpreted to suggest that testing interventions and research should focus on the poor, almost one third of men in the wealthiest quintile (30.7%) report they have not taken a test. Significant proportions of men in the wealthiest quintile in other sub-Saharan countries also report never having taken a HIV test, despite prevalence rates comparable to (and often higher than) other wealth quintiles (TACAIDS and ZAC 2013;Mishra et al. 2007). With these higher prevalence rates in mind, more needs to be done to understand testing behaviours amongst wealthy men. Furthermore, DHS data may potentially overstate HIV testing in wealthier quintiles, as participants in this quintile may be more socially aware and thus give socially desirable answers to sensitive questions (such as concerning HIV testing) to survey enumerators, introducing social desirability bias into the data (Kelly et al. 2013;Lowndes et al. 2012). Despite these concerns, as with research on HIV transmission, the wealthier quintiles are also absent from the literature on HIV testing.
The poverty narrative is also reinforced by the limited way in which HIV testing is understood theoretically and in conceptual frameworks. One theoretical strand has applied psychosocial models, such as the Health Belief Model or the Theory of Planned behaviour, to the issue of HIV testing (Ayodele 2017;Kabiru et al. 2011;Mirkuzie et al. 2011). In general, these approaches focus on the role of perceptions in the decision-making process, such as the perceived threat of being infected with HIV, the perceived seriousness of being infected, and cues to action influence testing behaviour. Another strand of thinking draws from economic approaches, where health seen as an 'investment' that may involve short-term trade-offs for longer term benefits, thus creating a clear set of incentives for individuals, and particularly wealthier individuals, to know their status (Oster 2012). Opportunity cost also forms the starting point for experiments that draw on behavioural economics that aim to quantify non-economic barriers to testing by estimating the subsidies that would change individuals' incentive structures so that they would take an HIV test (Linnemayr et al. 2018;Ostermann et al. 2015). Conceptual frameworks, on the other hand, emphasise accessibility to testing services, ancillary services available at the testing site, confidentiality of testing and test results, and the quality of counselling and testing (Njau et al. 2014). However, as with the literature on sexual behaviour and risk, the role of wealth is often unaccounted for, with models rooted in an individualistic paradigm. Whilst HIV risk has been conceptualised at the 'structural' level (Seeley et al. 2012;Sumartojo et al. 2000), beyond a number of concerns raised that reflect limited contextual issues such as stigma, and empirical work that documents the determinants of HIV testing (Nnko et al. 2019;Obermeyer et al. 2013;Staveteig et al. 2013) there have been few attempts to theorise HIV testing in a structural manner. Furthermore, consideration of the health-seeking behaviour of those who are wealthier is not always incorporated in behavioural models with the result that they tend to provide theoretical support for the notion that the wealthy are not a concern.
Epidemiological evidence also stresses the need to target men, as men consistently report lower levels of HIV testing compared to women in sub-Saharan Africa (Shand et al. 2014). In Tanzania, the most recent Population-based HIV Impact Assessment (PHIA) found that 29.2% of women compared to 40.2% of men had never been tested (TACAIDS 2018). Although this 'testing gap' is well documented, it continues to persist, leading to increasing efforts to understand and address the barriers that men face to taking a HIV test (Conserve et al. 2018;Quinn et al. 2019). Prominent amongst explanations for this gap, setting aside issues such as the enhanced access that women have to HIV testing at antenatal clinics, are concerns about HIV stigma (Young et al. 2010) and cultures of masculinity (DiCarlo et al. 2014;Siu, Wight, and Seeley 2014). However, it is unclear to what extent men from different socio-economic backgrounds experience similar barriers to HIV testing, and whether there are unique barriers for men from the highest wealth quintile.
This article engages with this missing demographic in HIV research. It reports findings from a qualitative research project conducted in Tanzania that explored attitudes towards, and practices of, HIV testing amongst wealthy men. This research, given the general absence of interventions that take seriously the issue that many wealthier men have not taken a HIV test, is particularly pertinent given the focus that HIV testing has in the UNAIDS 95-95-95 targets that aim to end the AIDS epidemic by 2030 (UNAIDS. 2015), and the well-documented benefits of HIV testing and subsequent take-up of treatment. With the research and policy agenda now focused primarily on treatment, HIV testing as the gateway to treatment will remain central to future efforts. Whilst some progress has been made, there remain significant barriers, especially in the sub-Saharan African context, to increasing the take-up of HIV testing services for men. Further progress is needed if the goal of universal coverage of treatment is to be achieved.

Methods
The data reported below derive from fieldwork conducted in Mwanza city, Tanzania, in February and March 2017, which involved 23 semi-structured interviews with the owners or regional managers of key public sector institutions and prominent private sector businesses within the city. This purposive sample was chosen so that the research team, in an exploratory qualitative research project, would not have to measure or quantify the wealth of potential participants given that men in these positions would comfortably be in the wealthiest quintile (if not in most cases in the top 1% in the Tanzanian context).
Mwanza is the second largest city in Tanzania with a HIV prevalence rate of 6.5% (4.3% men, 8.9% women) which is above the national average of 4.7% (3.1% for men, 6.2% for women) (TACAIDS 2018). Whilst three previous DHS have shown high prevalence rates amongst the wealthiest quintile in Tanzania (Long and Deane 2015), this has changed in the most recent survey, though prevalence rates amongst the wealthiest men (1.8%) are not significantly higher than for the poorest (2.8%). Antiretroviral Treatment is provided free of charge and is in general (especially in urban areas) widely available 1 . As in many sub-Saharan settings, public HIV services are primarily provided in separate HIV clinics.
The research team comprised one researcher (TD) from the UK, and three researchers from Tanzania (JW, SM and JC). All interviews were conducted by SM with the other authors not present. This approach (see Deane and Stevano 2016) for a full discussion of the strengths and limitations of this approach) reflected the potentially sensitive topic under discussion, and in the light of the linguistic limitations of one of the authors (TD), ensured that interviews could flow without the interruptions necessary if this had been done using the research assistant or other authors as interpreters (Pitchforth and Van Teijlingen 2005). Most usually, the interviews were conducted in the participant's workplace according to their convenience, an environment in which they were comfortable and which afforded significant privacy. Mgunga, who conducted the interviews, was younger than all the participants (and less wealthy), and so this power-relation was used to confer 'expert status' on participants as a means to get them to open up about potentially sensitive topics (Harvey 2011).
The project adopted a two-step approach to sampling. Firstly, potential participants were purposively sampled through an initial brainstorming process with the research team and other contacts within the National Institute for Medical Research, Mwanza centre. This included discussion about the most important sectors in Mwanza (for example Manufacturing, Construction, Wholesale and Retail Trade, Transportation and Storage, Accommodation and Food Services, Information and Communication, Financial and Insurance activities, Professional Scientific and Technical activities (UN Department of Economic and Social Affairs 2008), businesses that would likely be locally owned compared to businesses that would be managed on a regional basis but might be owned either by pan-Tanzanian or International firms, and the key public institutions that employ large numbers of workers. This list served as a starting point for approaching potential participants, although it was added to as the fieldwork progressed. Participants were contacted in a range of ways, such as through personal networks within the research community in Tanzania, the personal networks of friends of the research community, institutional relationships (such as local public financial institutions or private sector firms that had engaged commercially with the host research institution), and turning up unannounced at business and institutional premises. Secondly, additional participants were identified though the initial sample via snowballing through their networks.
Throughout the data collection period, a total of 36 potential participants were approached, but due to the sensitivity of the research and the general challenges of recruiting elite participants (Harvey 2011), not all those approached agreed to be interviewed 2 . The final sample was composed of nine private sector business owners, seven regional/zonal managers of private sector companies, and seven public institution regional/zonal managers. The average age of the sample was 48.5 years old, reflecting the fact that those in senior positions in the sampled organisations were more likely to be older. Whilst we are confident that all participants would have been comfortably within the highest wealth quintile, it is suspected that income levels within the sample were highly differentiated.
Ethical clearance was granted by the Medical Research Coordinating Committee (MRCC) of the National Institute for Medical Research, Tanzania and also by the University of Northampton (UK) Research Ethics Committee. Further, written permission to conduct the fieldwork was obtained from local governmental and community leaders and all participants gave oral and written consent prior to the interview.

General barriers to and enablers of HIV testing
Findings from the study can be divided into factors that reflect themes identified in the broader literature and that are applicable to most men regardless of their income level, and factors that are more specifically related to wealth and high(er) socioeconomic status. The focus of this article is on the latter. However, it is worth noting that participants reported a wide range of barriers to, and enablers of, HIV testing that have been previously documented in the literature (see Musheke et al. [2013])

Being wealthy and incentives for HIV testing -Enablers
There were several factors reported by participants that were related to their wealth and status which encouraged them to test for HIV. Firstly, they noted there were incentives for them to protect their health and test for HIV so that they could enjoy their wealth in future years. This reflects perspectives in which health is viewed as something to be invested in. It also reflects the idea of the opportunity cost from neoclassical economics as applied to health behaviour (Oster 2012), with wealthy men having greater incentives to engage in protective health behaviours so that they can live longer to enjoy future consumption that their wealth makes possible: If you do not take care of your health you will not enjoy the benefit of your job … You make sure that at the end of the year the company has gained profit in terms of expanding, you need to be alive in order to witness the fruits (Zonal Manager, Private Sector, 54) These are people who have plans for their life … the big capital is health because if money they have already, authority and power, they have good living environment, And their plans mostly are not for three years, they plan for fifteen years, thirty years, so you can have everything but if your health is not fine you cannot achieve that (Zonal Manager,Public Sector,41) A second theme was the awareness that not only was treatment more accessible to wealthy men who did not have to make repeated trips to the local clinic to pick up medication, ongoing adherence to drug regimes was also easier: Therefore a person with money, a person with good economic status is easier for him to be convinced to go for testing than a person with economic problems (Zonal Manager,Private Sector,34) A final theme was that for some participants, HIV was viewed (or was attempted to be viewed) as a disease or health issue similar to others, suggesting changing attitudes amongst wealthy men towards HIV: It's an issue of going to get tested just the same way we test for Malaria and Typhoid or other diseases … As I have said, it's the need to educate people to take it as a disease just as any other (Private Sector Owner, 33) When every person go for general testing they should be tested for HIV, it should be seen as one thing instead of separating as a more special disease than others, I believe there are other diseases that are needed to be tested and they are more dangerous than HIV, so the campaign should be general health testing then from that general testing HIV should be included (Regional Director,Public Sector,58) This potential 'normalisation' of HIV is reflective of advances in treatment and expanded availability that mean that a HIV diagnosis is no longer always the death sentence that it used to be.

Being wealthy as a barrier to HIV testing
In terms of barriers that relate to wealth, the primary concern reported by study participants was the potential loss of social status especially if an HIV result was positive and became known to others. Wealthy men were very sensitive to this issue, noting that most of them were well known in their communities due to their status, with a number of the more elite participants known (and recognisable) across the entire city. Therefore, protection of this status, or 'fame' as one participant put it, was paramount. Whilst those with high incomes were aware that they would receive the necessary support if they were HIV positive, their main fear was the potential loss of social status due to a positive diagnosis: So, there is a problem there in that the infection is seen as a bad thing … .that disease my friend is a bad omen because it exposes you as one who is sick, so you face that fear of been known as positive, fear of how you will be understood. But then if you can be willing to get help you will get very good assistance, so about the other guys because of their higher income levels and higher education levels I think their biggest fear is losing that fame (Business Owner,60) This need to preserve social status was also reflected in enhanced concern about being recognised went attending an HIV clinic and the confidentiality of testing services: Wealthy men have popular names in the communities, When people see him going at the HIV centre, they will not think that he is going for testing people will think he is already infected and he is going for treatment Therefore to maintain respect in the community he will not want to be seen at those centres, even if he is not HIV positive, just to be seen at those places can make the community believe that he is infected (Regional Director,Public Sector,58) Because I have told you we have no secret, no confidentiality, if you come in our offices here and have HIV testing or you say, we have tested [name of well-known man in Mwanza], you tested another one maybe is infected, you find it leaks but we were two. Have you understood, we remained two of us then surprisingly this information has leaked (Business Owner, age unknown) Whilst concern around confidentiality has been reported in the broader literature, wealthy men experience this as a heightened riskthey think they are more wellknown and therefore more likely to be recognised or spotted at a clinic should they go for a test, or be gossiped about by clinic staff who were present at the test. Furthermore, they have more to lose vis-a-vis a member of the general population.
A related concern was what it might mean in the workplace, either as an owner or a senior manager, if men were seen at a clinic or if a positive HIV test result were to become common knowledge. In most cases, it was reported that this would lead to loss of status and authority, leading to increased instability, and it was also suggested that a positive diagnosis would also have consequences for business operations and access to finance: It is possible, the challenges that people face, maybe for people with a big position like this, The way the community sees them, they would have worries about their area of work.For anything that would come up, the perception of a leader at their place of work, could be a big challenge to leaders … For example, there are two hundred people at an institution. So up here they depend on you, everyone looks at you and if you are infected people will start saying different things. So, if it happens in such environment, they think that the place of work will be unstable (Regional Manager, Public Sector, age unknown) Especially a wealthy person, first he is known by many people, and he know that he is respected by many people, again he knows many people depend on him therefore if he will be found has problems and has been tested and found HIV positive it is possible that thing will be known by many people, who will give him loans, because a businessman has to take loans, thus why many of them are hiding themselves when s/he want to go for testing (Business Owner,59) A final barrier to HIV testing reported by participants was simply that they were busy and did not have time to take a HIV test: These senior level management people are always very busy, You see they don't have that time to go and stand in lines waiting, they are impatient, So that is also another problem, in that they are so busy that they lack that time to do those things (Zonal Director, Public Sector, 54)

Wealth and testing practices
Concerns about loss of social status and negative business impact that could result from being recognised at a clinic, result in wealthy men deploying a number of strategies to avoid testing at the local clinic or anywhere they thought they might be recognised. This often involved them undergoing testing in a different city, primarily Dar Es Salaam, or in a different country, such as Kenya (Nairobi is as accessible as Dar Es Salaam from Mwanza) or South Africa. Destinations in Europe and the Middle East were also mentioned as testing locations.
He has his own routine of going to Europe to take his health testing, not here, after every six months he must go to Europe, for testing or in Nairobi, or wherever, A wealthy person … .wealthy person can do whatever s/he wants (Business Owner, age unknown) Many of them are hiding themselves when s/he want to go for testing, he goes to Nairobi where there is no one who knows him and the results remain with himself (Business Owner,59) These measures reflect a range of factors, including the lengths that wealthy men go to in order to minimise the likelihood that they will be seen, a degree of opportunism through the utilisation of HIV related services while they are away on business trips, and the options that wealth enables that are s out of reach of the vast majority of the population.
Alongside testing in other locations, wealthy men also reported testing in a range of settings including private hospitals: They test at hospitals many hospitals are private hospitals because if you to the government hospitals there is a long queue, so people see it is better if they go to the private hospitals to test (Regional Officer,Public Sector,54) Linked to this, some participants noted that they were tested for HIV as part of routine health checks. This reduced concerns around confidentiality and privacy, as these appointments did not take place in a HIV-specific clinic: More of the people with these positions have their system. Routine check-up, that I must get health testing every time … therefore in his routine check-up he does not test for HIV only, but he tests HIV and other things (Branch Director, Private Sector, age unknown) Such routine check-ups are not an option for the general population, reflecting enhanced access to healthcare that higher income levels bring. Closely related, tests may be accessed via the private health insurance schemes provided to senior employees: Most of them test because also, they have an insurance cover, an insurance cover if your company offers such insurance services, automatically you'll test, so you can test by choice or by force [laughter]. If you have an insurance scheme, then they would also like to know of your health (Zonal Manager,Private Sector,58) Testing could also occur in conjunction with the investigation of another health problem: Most of the wealthy men have [a] big age too. Therefore, when he gets the age-related diseases such as blood pressure, diabetes, cholesterol and other things of that nature, when he go for testing he is also convinced to test for HIV (Regional Director,Public Sector,58) Finally, and in contrast to most of the reported testing practices that enable wealthy men to mitigate concerns around privacy, confidentiality and HIV stigma, there was one setting in which wealthy men reported testing more publicly: namely, the workplace.
He went and got tested, such acts indicate that we as CEOs we should be on the forefront so that we don't find ourselves telling the employees to go and get tested but not us, it would be better when telling them to go and get tested you confirm to them that you as the director you have already done so, and you show them, that would really help (Zonal Director,Public Sector,54) Most times I just like to test because I want to know my status but also to encourage others because by the time you can insist them to test HIV. If you don't test yourself they cannot test too. Therefore in order to convince people to go for testing and to make them agree and see the importance of testing you must be a good tester (Zonal Manager,Private Sector,34) In these examples, wealthy men noted that their organisational leadership roles required them to take a HIV test more visibly to set an example to their employees in the hope it would encourage them to also test. This example of testing in a public domain did not act as a barrier to testing and was in stark contrast to the avoidance of public clinics and concerns about confidentiality noted above.

Discussion
The evidence presented above sheds light on HIV testing practices and attitudes towards HIV testing among members of a previously rarely studied population and enables reflection on a range of conceptual and theoretical issues. It also raises questions about HIV research and intervention priorities and provides insights with implications for policy makers. Firstly, our findings suggest that wealthy men should not necessarily be a group to prioritise for public health interventions focused on HIV testing, given the many different options and opportunities they have in comparison to the general population. The higher rates of HIV testing reported in the epidemiological data for the wealthiest quintile are plausible, reflecting the privileged access members of this group have to ongoing health check-ups, private clinics and private health insurance.
However, our findings signal a high degree of internalised and anticipated HIV stigma among members of this group. To some extent this reflects the average age of participants, many of whom lived through the early days of the epidemic when HIV stigma was pervasive. Whilst there were reports of changing attitudes towards HIV as 'just another' disease, wealthy men go to great lengths, such as travelling to other cities and countries, to protect their privacy and/or confidentiality, with local public HIV clinics being avoided at all costs. Central here is the maintenance of social status and the potential problems that being seen taking a HIV test or a positive diagnosis might have for relationships with partners, families, the general community and employees. Whilst this is an issue that could apply to all socio-economic groups, concerns about loss of social status are amplified within this group in the context of high levels of income quality and the desire to protect income levels and/or lifestylessimply put, there is more at stake for wealthy men. Furthermore, whilst wealthier men report less stigmatised views related to HIV in PHIAs, this is at odds with the HIV testing behaviours they report, and so it is unclear whether attitudes are changing. To some extent, a tension exists between the high levels of stigma reported by wealthy men and their enhanced ability to avoid this, which may help to explain the generally slow progress of reducing HIV related stigma in Tanzania and other settings in sub-Saharan Africa. If stigma can be avoided, then there is less need to address it directly as a public health priority or a social issue, priorities that are set by those in power at key public institutions. Many strategies that aim to expand HIV testing such as self-testing kits are designed to enable the avoidance of stigma given the challenge of changing these deep-seated attitudes towards HIV (Conserve et al. 2018). Furthermore, the stigmatising attitudes held by influential community members and people in positions of power may also serve to assist in the perpetuation of stigma. Therefore, while our findings suggest that promoting interventions targeted at wealthy men to enhance HIV testing should not be a central concern for policy makers, evidence suggests that there is still much work to be done to address HIV stigma across all levels of Tanzanian society Our evidence points to the role of income inequality in shaping divergent experiences of HIV testing across different social classes. While wealthier men can avoid the public glare through the mechanisms noted above, the ignominy of having to queue at a public HIV clinic is an experience to be suffered by the poorer classes of society. In this respect, HIV is intimately related to poverty, with HIV-testing related stigma primarily reserved for the poor. Our findings also emphasise the different meanings associated with 'access', so often seen as a key issue in explaining low testing rates. In the general literature this term is used as a proxy for factors such as distance to a clinic and the availability of services, the cost of getting there, and the lost income from attending an appointment (McIntyre, Thiede, and Birch 2009) . However, our findings emphasise that what is being accessed is highly dependent on socio-economic status, with parallel health systems serving different classes through private health insurance schemes, private clinics and testing in the context of more general healthcare appointments. Therefore, 'access' to testing is often presented in conceptual frameworks in a depoliticised way which obscures these inequalities and experiences. Furthermore, whilst previous work on the social determinants of HIV testing has identified socioeconomic status, education and employment as key social determinants, our findings enable some initial mapping of the pathways through which these operate. For example, formal employment can enhance access to testing through the provision of health insurance and access to private, preventive health care, as well as through HIV testing programmes that are sometimes run in formal workplace settings. High socioeconomic status operates through the affordability of not only access to private health care but also HIV testing services in other cities or even countries. Conceptual frameworks that address HIV testing need to more explicitly incorporate these social determinants of health if they are to provide a more comprehensive picture of how and why HIV testing rates are positively correlated with increased income.
Our findings also enable reflection on testing-related decisions by emphasising the role of social context and the multiple social roles that wealthy men play. As noted above, one exception to the stigma reported by wealthy men was related to the workplace. In this setting, testing for HIV was viewed in a positive light. Rather than a potential acknowledgment of risky sexual behaviour, it was seen as setting a good example for others through demonstrating leadership. This example illustrates how different socio-economic contexts and the roles that these confer on wealthy men are of central importance in shaping testing decisions and emphasises the importance of understanding how 'class formation and consolidation … frames both the transmission and response to HIV/AIDS in Africa' (Bujra 2006). Whilst the observation that context shapes behaviour is not new in the social sciences, a political economy lens enables a reflection on how these contexts and settings are constituted, directing attention to class and class dynamics rather than context as a product of culture or identity (Helle-Valle 2004). In this example, this positive experience is reserved for wealthy men and is directly related to their social status and role as owners and/or employers. The role that political economy plays in shaping HIV testing behaviours also enables a critical reflection on rational choice and behavioural models of health seeking behaviour. These models are unable to engage with the way in which social and political context frames decision making, and how key barriers to testing such as HIV stigma are experienced in different ways across different settings. Furthermore, the improved health-seeking behaviours often ascribed to wealthy men may not provide as convincing an explanation for higher rates of testing as the wildly divergent access to testing services that they have. This analysis has important implications for policy makers. Firstly, acknowledging the role that social relations play in creating contexts where those in leadership roles feel able to test in a more visible and public way, policy makers should focus on identifying and targeting the social and political spaces in which testing (as a wealthy man) can be cast in a positive light and does not attract the degree of stigma it usually would. One example from our research is that of the workplace where wealthy men can be viewed as demonstrating leadership, but this principle could extend to a range of different settings in which wealth(ier) men are also in positions of power and influence. One recent example of this took place in Tanzania, where over 300 MPs publicly tested to help promote HIV testing (The East African 2018), and public testing by presidents, members of parliament and other political figures has been a strategy to encourage HIV testing employed across a number of sub-Saharan African countries (Karan, Hartford, and Coates 2017). Applying these principles to other contexts in which men who are in positions of relative power and influence, for example to religious leaders and community leaders (Jobson et al. 2019), may be able to test more publicly could encourage and normalise HIV testing. In this way, the power relations within the workplace and other settings can be utilised to challenge HIV related stigma.
A second key policy issue relates to the debate about whether HIV services should continue to be provided at separate clinics, as this is a context often avoided due to concerns around stigma and being seen. Whilst recent research suggests that there may be a range of challenges associated with integrating HIV services into the general health system (Zakumumpa et al. 2018), our findings suggest that overcoming these challenges could significantly enhance access to HIV testing. The wealthy men in our sample reported testing alongside treatment and care for other health issues in private clinics which enabled them to test in relative privacy. This suggests that the integration of services is possible (as it is already happening in the private sector), and so given a healthcare setting in which testing poses relatively fewer stigma-related risks, men are willing to test. Reform of the health system to enable the fuller integration of HIV within mainstream services is perhaps the most significant barrier to expanding coverage of HIV testing.
A final reflection relates to the value of 'studying up'. The general focus of the majority of social science work on the poor skews knowledge about the social aspects of the epidemic and can also implicitly stigmatise poor people for their health-related decisions, a view that our findings have helped to challenge. The HIV stigma widely observed and reported across sub-Saharan Africa is not just purely a product of lessenlightened attitudes towards HIV amongst the poor, it also reflects the attitudes of those in powerful positions who are conversely in the best position to avoid experiencing it. This raises questions about the extent to which stigma is a priority for those in power, and how poverty-related narratives can serve to deflect attention from wealthier classes. More research that focuses on higher income groups is needed if the dynamics of the epidemic are to be fully understood.

Notes
1. This may not be the case during the COVID pandemic 2. For a full review of the challenges of recruiting participants for this study, see Deane et al. (2019)