Ethics of pursuing targets in public health: the case of voluntary medical male circumcision for HIV-prevention programs in Kenya

Abstract

The use of targets to direct public health programmes, particularly in global initiatives, has become widely accepted and commonplace. This paper is an ethical analysis of the utilisation of targets in global public health using our fieldwork on and experiences with voluntary medical male circumcision (VMMC) initiatives in Kenya. Among the many countries involved in VMMC for HIV prevention, Kenya is considered a success story, its programmes having medically circumcised nearly 2 million men since 2007. We describe ethically problematic practices in Kenyan VMMC programmes revealed by our fieldwork, how the problems are related to the pursuit of targets and discuss possible approaches to their management. Although the establishment and pursuit of targets in public health can have many benefits, assessments of target-driven programmes tend to focus on quantifiable outcomes rather than the processes by which the outcomes are obtained. However, in order to speak more robustly about programmatic ‘success’, and to maintain community trust, it is vital to ethically evaluate how a public health initiative is actually implemented in the pursuit of its targets.

BACKGROUND

A comprehensive history of the use of targets in public health initiatives has yet to be written, but there are hints scattered in the existing literature. Some scholars claim that the idea of using targets to better organise health promotion activities ultimately originated in the manufacturing sector, particularly through the idea of ‘management by objectives’ described in Peter Drucker’s The Practice of Management.¹ Others state that Drucker’s conception itself owes much to the work of earlier mechanical engineers and managers, particularly Fredrick W. Taylor’s The Principles of Scientific Management.² The use of targets and associated performance indicators, has become ubiquitous in many societies throughout the world, and has altered practices from policing³ to education.⁴

It is unclear precisely when the target approach entered into the health domain. According to Gunning-Schepers,⁵ the creation of universal healthcare systems after the Second World War was a major impetus for the use of health targets as access to health services needed to be radically increased, and plans for expansion were organised around outcomes such as levels of vaccination coverage and numbers of health professionals per population. Decades later, as healthcare expenditures increased during periods of economic decline, health targets (eg, reduction of emergency services waiting times) were used to cut costs and increase efficiencies. Healthcare system reform and policy implementation commonly understands improvement in terms of setting and reaching targets. In addition, targets have been a prominent feature of polio, smallpox eradication and family planning campaigns worldwide over the last decades.⁶

The use of targets by healthcare systems and programmes has become so pervasive that it may be difficult to imagine how health promotion was ever conducted without them. Global health initiatives (eg, Sustainable Development Goals) are strongly marked by a ‘management by objectives’ approach, with distinct targets, outcomes, benchmarks and deadlines.⁷ Setting and pursuing health targets at national, regional and local levels has become normalised, and to some extent, globalised. Over time, the target-oriented approach to health promotion has also been strengthened by the development of new epidemiological tools (eg, mathematical modelling) and technology-backed improvements in health surveillance and reporting (for an example, see Houben et al⁸).

Health targets have their virtues. As Wismar et al⁹ state, “Health targets express a commitment to achieve specified outputs in a defined time period, and enable monitoring of progress towards the achievement of broader goals and objectives.” Having explicit targets set in advance can enhance transparency, timelines can stimulate productivity, benchmarks in function of targets can increase efficiency, and monitoring can increase accountability and enable course correction. The use of targets can also have an ethical dimension, for example when they are directed towards the reduction of health inequalities. Importantly, advocates of targeting in health promotion can quantitatively show how uptake of an intervention increases beyond base-line over time, and use the increase as a marker of success. The latter is important for advocacy and funding requests.

However, the target-oriented approach to health promotion has also been subject to criticism from its inception. Critical studies have focused on the use of performance targets in public healthcare systems, especially the National Health Service in the UK. Many shortcomings and ethically troubling features have been identified, offered with varying levels of evidence. For example, the use of targets has been associated with ‘gaming the system’, that is, achieving targets by inappropriate means. Bevan gives the example of patients being forced to wait in ambulances in order for hospitals to meet emergency department waiting-time targets.¹⁰ Hood similarly speaks of ‘gaming in target world’ to describe a range of questionable strategies to meet health-related targets that take a number of forms, from changing the interpretation of the target (and what meeting it means) to outright data fabrication.¹¹ Mannion et al¹² describe how the introduction of star ratings for acute hospital trusts in England led to beneficial changes as well as distortions of clinical priorities, bullying and intimidation, erosion of public trust and ghettoization. Lockwood et al describe how targets of lower prevalence in some leprosy programmes may have been ‘met’ by simply reducing active case finding, diagnosis and reporting.¹³ Critical analyses of the use of targets in global health initiatives are relatively rare, with some exceptions, such as Hendrixson’s recent discussion of the adverse effects of using contraception targets in developing countries¹⁴ and discussions surrounding India’s population-control initiatives.¹⁵

In this paper, we reflect on the ethics of using targets in voluntary medical male circumcision (VMMC) for HIV-prevention programmes across eastern and southern Africa (ESA). Our ethical discussion is inspired by our prior empirical social science research in Kenya on VMMC among adolescents in Kenya,^16,17 as well as the relevant scientific literature. Our reflections raise a crucial general question: is the now common public health practice of using targets to increase uptake of a health intervention fundamentally at odds with other important ethical values and concerns?

VMMC programs in ESA

Three randomised controlled trials conducted in Africa indicated that medical male circumcision (MC) significantly reduces female-to-male HIV transmission during sexual intercourse.^18–20 Since the publication of these findings, VMMC has become an important part of HIV-prevention efforts, particularly in regions of ESA with high HIV incidence and low circumcision prevalence. VMMC is currently central to the Joint United Nations Programme on HIV/AIDS (UNAIDS) strategy for ending AIDS by 2030.²¹ It is also a top priority of the United States Agency for International Development and the United States President’s Emergency Plan for AIDS Relief (PEPFAR) to combat HIV in Africa.^22,23

To date, PEPFAR has been the leading funder of VMMC for HIV prevention in Africa and has sponsored almost 22.8 million MCs since 2004.²⁴ To optimise VMMC’s potential impacts on population-level HIV prevalence rates, epidemiological model-informed targets are annually set for how many MCs are to take place among specific age groups within a given region.¹⁶ With nearly 2 million MCs,²⁵ Kenya has been a front-runner in VMMC programme scale-up. To achieve these results, Kenya has consistently met or exceeded its VMMC targets^22,23,25; and is generally considered a VMMC ‘success story’ in ESA.²⁶ For example, in fiscal year (FY) 2018, Kenya achieved 101% or 304 576 of the targeted 300 744 MCs,²⁵ though these numbers may be inflated by reporting issues.²⁷ While early on Kenya’s VMMC programme sought to circumcise men aged 15 and older, in 2014 providers began offering services to adolescents from age 10.²⁸ For FY 2017, 227 272 PEPFAR-sponsored MCs took place in Kenya. Of these, 60% of these were among boys aged between 10 and 14 years, 23% were among those aged 15–19 years and 14% were among those aged 20–29 years.²⁹ However, while PEPFAR provides the funding for Kenya’s VMMC programmes, it is its independent implementation partner organisations (IPs) (generally non-governmental organisations) who work on the ground to educate the public, increase demand for VMMC services, recruit potential clients and provide MC clinical services.^30,31

Ethical considerations

Agencies involved in VMMC promotion for HIV prevention recognised from the beginning that mass male circumcision campaigns in sub-Saharan Africa could be ethically challenging. The targeted ‘high HIV/low male circumcision’ communities are commonly marked by poverty and other forms of social vulnerability. In 2008, UNAIDS issued guidance on VMMC implementation focusing on human rights, ethics and legal considerations.³² This document grew from an earlier WHO/UNAIDS consultative workshop in March 2007 which concluded that, ‘a human rights-based approach to the development or expansion of male circumcision services requires measures that ensure that the procedure can be carried out safely, under conditions of informed consent and without coercion or discrimination’.³³

However, despite these early calls for the ethical implementation of VMMC services, studies have shown that in practice, VMMC IPs continue to be dogged by both practical and ethical issues.³⁴ These concerns include, but are not limited to, how youth are recruited for VMMC,³⁵ problems with parental consent and adolescent assent, a lack of understanding related to VMMC and its benefits and risks,^36–38 as well as the necessity of continuing HIV-protective behaviours³⁹ and care access following MC.^40,41 Indeed, a 2014 assessment of VMMC programmes in Kenya, South Africa, Tanzania and Zimbabwe found that IPs are commonly plagued by adverse events, insufficient equipment and supplies, and failed adherence concerning approved guidelines for best clinical practice.⁴²

During a discussion of our study on the responsible inclusion of adolescents in HIV-related research in Kenya, a county health authority mentioned ethical concerns about local VMMC programmes, particularly how adolescents were being recruited and the integrity of consent processes. Following up on these concerns, in 2016–2018 we conducted a mixed-methods study of VMMC implementation practices and experiences in Kenya.^16,17 Our goal was to gain a better understanding of how VMMC practices unfold ‘on-the-ground’ in Kenya. We conducted indepth interviews and observation sessions in a region of Kenya with high HIV prevalence, low circumcision prevalence and where VMMC programmes are actively implemented.¹⁶ In-depth interviews involved 29 VMMC stakeholders including community-based client recruiters known as ‘mobilisers’, HIV counsellors, clinical service providers, schoolteachers and policy professionals. The observation sessions took place at 14 VMMC clinics alongside mobilisation teams at 13 different venues.¹⁶ In addition, we conducted a survey of 1939 Kenyan men aged 15–19 years about compliance of VMMC programmes with parental permission and adolescent assent processes.¹⁷ We summarise our findings below, focusing on the tension between the public health use of targets in VMMC programmes and the ethical repercussions in the field when these targets are pursued. More details on the qualitative and quantitative study methods can be found elsewhere.^16,17

A summary of our empirical findings from Kenya: unintended consequences of efforts to meet VMMC targets

In Kenya, male adolescents are primarily recruited for MC at schools during recruiterled informational ‘health talks’. For practical reasons, schools do not allow VMMC mobilisation to take place during examination periods or other demanding times during the school year. This means that in order to efficiently recruit school-age clients to meet VMMC targets, mobilisers must focus their efforts during the ‘high seasons’ of client availability, such as the few days following examinations at the end of the school term.¹⁶ Data we collected in Kenya suggest that the pressure to reach VMMC targets may have unintended consequences for programme implementation. This includes compromised VMMC service delivery and adherence to approved protocols.¹⁶ Discussed below, we found ethical concerns related to adolescent client recruitment, informed consent and clinical care.

Client recruitment

Mobilisers are crucial to VMMC programmes as they are responsible for engaging with communities, educating the public about VMMC, increasing demand for VMMC services and facilitating the arrival of clients at clinics. Mobilisers are paid, but their performance is evaluated based on whether they attract enough clients. Overall, our interviews and observations revealed that some mobilisers turn to misleading or otherwise ethically questionable practices to recruit young clients.

For instance, some mobilisers downplay circumcision risks, while emphasising potential benefits and the offering of incentives to circumcise (eg, sodas and new underwear). Although monetary incentives were not employed in the programmes we studied, some public health professionals advocate for their use to increase VMMC uptake,⁴³ which raises concerns about coercion when applied to adolescents who largely live in poverty. Additionally, some mobilisers encourage the stigmatisation, ridicule and shaming of uncircumcised youth as means to increase client demand. During observation sessions, we noted how some mobilisers make statements suggesting that those who choose to remain uncircumcised should be blamed for the spread of HIV.¹⁶ As a result, some adolescents may opt for MC, not necessarily on the basis of weighing the risks and benefits of the procedure, but simply from social pressure to conform. Such strategies for recruitment may contribute to poorer health outcomes among youth: our survey of Kenyan men aged 15–19 years found higher levels of psychosocial distress among uncircumcised adolescents compared with their circumcised peers, possibly due to MC-associated social pressure and stigma.¹⁷ It should also be noted that pressure to circumcise from other ethnic groups has been experienced by the traditionally non-circumcising Luo tribe for many years, including incidents of forced circumcision, and this pressure has only increased as the Luo have become the primary focus of VMMC efforts in Kenya.^44,45

Furthermore, some mobilisers recruit underage boys whom providers may or may not turn away when they arrive at the clinic. Current VMMC guidelines for HIV prevention and national protocols approved for use in Kenya require that adolescent clients must be at least 10 years of age.⁴⁶ Our interviews and observations revealed that children under 10 were being circumcised in VMMC programmes, a practice which may be tolerated by mobilisers and other stakeholders out of concern about meeting VMMC targets.

Informed consent

Attaining quality informed assent and parental permission for research or health interventions involving adolescents is known to be challenging in the best of circumstances.^47,48 VMMC programmes involving adolescents in rural communities oriented around the attainment of ambitious targets have less than ideal assent and parental permission processes. The survey data we collected among men aged 15–19 years revealed that up to 10% of adolescents underwent MC without parental or guardian consent.¹⁷ While neither the qualitative nor quantitative aspects of our study were designed to detect possible family-related social harms associated with unpermitted circumcision, our qualitative study revealed other concerns. These include missing consent forms, forged parent/guardian signatures on consent forms, lack of awareness among parents about the nature and significance of the procedure and lack of awareness that VMMC programmes involve HIV testing (and potential HIV status disclosure).¹⁶

‘Cutting corners’ in the clinic

We found evidence that, especially during the high season, VMMC mobilisers and providers may work to circumcise many more boys than the approved number per VMMC clinical team per day. Strategies to increase the number of adolescents who undergo MC per day include ‘cutting corners’ in clinical care by skipping steps (eg, opting not to complete preoperative or post-operative patient examinations), operating on multiple clients in the same room and rushing to complete large numbers of circumcisions with few breaks. We observed that while safety protocols required two providers to work together on each patient as a surgical team, individual providers worked alone in order to increase the number of circumcisions conducted.¹⁶

Reflections on VMMC targets in Kenya

On the face of it, the use of targets in VMMC programmes has been a success: the number of Kenyan men who are circumcised has risen dramatically since 2007.⁴⁹ On the fundamental assumption that the increase in MC will have an important impact on HIV incidence in Kenya, the use of targets can be justified in public health terms. On the other hand, as noted above, part of the success may have involved departing from a ‘human rights-based approach’. To some extent, our study suggests that the epidemiological imperative to reduce HIV incidence by increasing circumcision prevalence has taken precedence over other issues such as safety, voluntariness, informed consent/assent and privacy. How should one react to this? A number of distinct approaches are possible: elimination, utilitarianism, harm reduction and contextualisation.

Elimination

One possibility is to advocate against the use of targets altogether. One could conceivably continue to promote and offer MC services without setting/pursuing specific numerical targets. This might alleviate some of the ethical problems our study revealed, while yielding far fewer circumcisions than a target-based model. This was also suggested by one of our interviewees, who drew an analogy with opt-in HIV or tuberculosis testing integrated within Kenyan medical facilities. Infectious disease and public health professionals, as well as funders, would likely reject this alternative as unrealistic and even unethical, since fewer MCs mean more HIV transmission events. According to this objection, the cost of rejecting the use of targets (including economic costs of maintaining HIV-infected persons on antiretroviral treatment for life) would be unacceptably high. In addition, there are practical objections: how would budgets be structured for non-target-based VMMC programmes? How would IPs be held accountable by their funders and how would programme success be measured?

Utilitarian

Another view is that the target-based approach should be retained as is, and one should consider the ethical concerns with the approach as harms and burdens outweighed by the public health benefits. The overall consequences of not having a target-based approach are (far) worse than taking that approach, and one should rationally and morally choose the approach with the better overall outcomes. One implication of this view is that agencies such as UNAIDS, WHO and PEPFAR would have to, in all honesty, drop aspirational language about VMMC programmes being significantly constrained by human rights and ethical considerations, as this does not fully reflect realities on the ground.

Harm reduction

This approach would retain use of targets, but take robust practical measures to minimise the sort of unintended consequences and undesirable behaviours noted in our study. For example, one could provide additional material and human resources to VMMC teams and local health facilities in order to make the achievement of targets easier, and thereby lessen clinician fatigue, deviations from protocol and temptations to cut corners. Those obtaining consent could use alternative approaches (ie, other than simply reading a form) that could improve comprehension, and mobilisers could be trained to avoid recruitment approaches that leverage stigmatisation or peer pressure. However, as harm reduction efforts are made more robust, the more likely they are to impede the attainment of ambitious targets. As long as programmes (like those in Kenya) continue to meet or exceed their VMMC targets, there may be little motivation on the part of funders and local healthcare systems to invest in harm-reduction strategies.

Contextualisation

This approach would also retain the use of targets, but make them less demanding by lowering the number of circumcisions to be conducted and/or increasing the time for reaching their targets. This would adapt VMMC targets to the socio-economic realities of programme implementation. For example, some respondents in Kenya suggested retaining annual targets but allowing monthly targets to depend on availability of adolescents or other intervening factors (such as the rainy season). Setting and pursuing these ‘smart targets’, on this view, would involve more than a top-down approach that sets national targets on the basis of mathematical models (ie, HIV infections averted by circumcisions over time); it would also mean engaging with a wide range of community stakeholders and taking local constraints into account.¹⁶ This approach, by lowering the pressure of time and numbers, could potentially reduce the unintended consequences and undesirable behaviours that our qualitative study exposed. But again, as long as VMMC programmes are reaching their targets, many in the public health community may consider the aggressive pursuit of ambitious targets a success story that is not to be tampered with.

The harm reduction and contextualisation approaches are attempts to ‘pursue targets responsibly’. However, if these approaches in practice continue to override concerns about safety, consent and coercion in the pursuit of VMMC targets (ie, if they are utilitarian positions in disguise), this strengthens the case for elimination as a non-utilitarian option. However, our purpose here is not to defend any particular approach as ethically superior. It is rather to raise awareness that the setting and pursuit of targets in VMMC programmes is both efficacious and problematic, and any way that the inherent tension between public health goals and other important considerations is managed constitutes an ethical—and not a purely public health—decision.

Finally, we should mention the ethics of pursuing VMMC targets has been further complicated by the emergence of two other HIV-prevention approaches: treatment as prevention and pre-exposure prophylaxis (PreP). As these uses of antiretroviral drugs to prevent HIV transmission become more widespread in countries like Kenya, the specific population-level effect of MC on HIV incidence may be increasingly diluted and/or difficult to measure. In addition to the fact that HIV prevalence is decreasing in some areas of sub-Saharan Africa (while rising in others),⁵⁰ increased access to antiretroviral treatment and the use of PreP will add fuel to an already ongoing debate about the true significance and impact of VMMC programmes. Over the last two decades, doubts have been raised about the implementation of VMMC as an HIV-prevention approach, including questions about the validity of the original randomised clinical trials^51,52 and how the trial results were translated into policy,⁵³ though efforts to question trial validity have also been disputed.^54–58 Recently, using 2013 Demographic and Health Survey data in Zambia, Garenne and Matthews conclude that a sharp rise in MC has had only a marginal effect on new infections among men aged 15–29 years.⁵⁹ Other studies suggest otherwise.⁶⁰ According to Brives, defenders of the clinical trials on male circumcision have largely sidestepped long-standing debates about the limits of evidence-based medicine and the distinction between the efficacy of an intervention in a controlled trial versus its effectiveness in real-world circumstances.⁶¹ Bell writes that ‘evidence alone does not inevitably inspire an appetite for intervention’ and explains the strong ‘appetite’ for implementing VMMC in terms of three non-scientific factors: (1) the rise of evidence-based medicine and a simplistic conception of randomised controlled trials as revealing incontrovertible ‘facts of nature’; (2) the limited success of behavioural and structural approaches to reduce HIV transmission in the past making a biomedical (surgical) intervention highly attractive to medical, public health and industry stakeholders, especially with male circumcision being considered less politically controversial than condom use; (3) underlying and long-standing assumptions about the nature of African sexuality, especially the idea that the Africans are promiscuous and therefore HIV in Africa can only be tackled by biomedical approaches that operate independently of behaviour.⁶² Uncertainties about the scientific basis, underlying motivations and assumptions, and overall effectiveness could weaken the fundamental public health and ethical justification for VMMC programmes as a whole. Particularly in Africa, trust in medical and public health authorities may also be seriously compromised if it turns out that millions of men were circumcised by global initiatives to little avail.

CONCLUSION

Our qualitative study was small, has limited generalisability, and cannot speak to the prevalence of ethical issues within VMMC implementation programmes in Kenya generally or in any other ESA country. We were fortunate that our interviewees were willing to disclose some less than savoury aspects of VMMC programmes and permitted us to observe the actual conduct of VMMC programmes. At a minimum, however, we believe our study affords a glimpse into what may be occurring on the ground in VMMC programmes, not only in Kenya, but also in other VMMC initiatives in other sub-Saharan countries with IPs that are similarly driven to meet VMMC targets. There are still many unknowns. More information is needed about the processes involved in the setting of national and regional circumcision targets, that is, who is involved and how such determinations are set. Our interviewees seemed to experience targets as descending to them from on high, without significant input from IPs, and without accompanying rationale and justification.

It is noteworthy that at least some departure from what Bevan provocatively calls a regime of ‘targets and terror’ during the early 2000s in the UK⁶³ has been taken in Europe in regard to health system reform efforts. The darker side of targets, and the need to use them wisely, is slowly gaining visibility. By contrast, a relatively uncritical use of targets, and a relative lack of attention to its effects, seems to strongly mark VMMC programmes largely funded by American agencies and conducted largely in sub-Saharan Africa. To some extent, a model where HIV-prevention interventions are contracted out to local agencies by international funders may reflect the weakness of public health services and the limits of global health governance, that is, it is possible that the problems with target-seeking practices might be less acute if such services were locally governed and embedded within well-functioning healthcare systems. In any case, we hope to stimulate a more critical stance towards the public health use of targets by publicising ethically significant phenomena that often lie buried underneath public health ‘success stories’. The ethics of target-setting goes beyond VMMC. Similar concerns are likely to arise when any public health initiative, structured predominantly by the achievement of quantifiable targets, is implemented. We hope that those in public health and social science interested in the neglected area of ‘implementation ethics’ will emulate, amplify and nuance our findings in the future.