‘Ending AIDS’ between comparison and commensuration and the role of global health indicators

ABSTRACT

The use of targets and indicators in global health has become ubiquitous within global health and disease elimination programmes. The drive to ‘end AIDS’ has become a global flagship endeavour, including nation-states, donor organisations, NGOs, pharmaceutical companies, medical researchers, and activists. Almost synonymous with the campaign of ending AIDS is UNAIDS’ 90-90-90 targets. Beyond indicators’ role in neoliberal global health, an essential aspect of indicators and quantitative metrics is their ability to provide a basis for measurements and comparability across time and between different actors and entities. These processes are based on what has been called, commensuration, visual simplification, and serialisation. This article seeks to provide an account of how we can think about indicators in the drive to end AIDS as doing work that is contingent upon commensuration, simplification, and serialisation. The argument is that by attending to issues of commensuration, visual simplification, and serialisation we are better able to see how we risk erasing and foreclosing other forms of conceptualising what the end of AIDS could be. Logics of quantification risks erasing and foreclosing other qualitative aspects of the HIV epidemic as well as obscuring various epistemological tensions inherent in counting towards the end of AIDS.

KEYWORDS:

• Metrics
• HIV/AIDS
• commensuration
• epidemic
• comparison

Introduction

The use of targets and indicators in global health has become ubiquitous within global health and disease elimination programmes (Adams, 2016). This is not to say that target setting is entirely new in global health; as Randall Packard shows in his historical work (Packard, 2016, 1997), targets have a more extended history in disease elimination efforts than we often think. Another example would be the work of Gorsky and Sirrs on the history of ‘the politics of international health system metrics’ (Gorsky & Sirrs, 2017) and, to a certain degree, the work of Iris Borowy on the League of Nations Health Organization and its use of indicators and statistics in global health (Borowy, 2003).

The use of indicators in global campaigns to ‘end’ various diseases is part of this history of global health indicators. Riding on the tails of what often seems to be an exaggerated techno-optimism and faith in biomedical advances, we find indicators in campaigns aiming to end tuberculosis, hepatitis, and malaria (Waheed, 2021; WHO, 2015, 2021). The drive to end AIDS is no different; subsequently, indicators and targets are at the centre of the discourse on the end of AIDS. The drive to ‘end AIDS’ has become a global flagship endeavour, including nation-states, donor organisations, NGOs, pharmaceutical companies, medical researchers, and activists (Kenworthy et al., 2018a, 2018b). A pivotal moment in the genealogy of the end of AIDS discourse was findings that showed that people living with HIV and who adhered to daily antiretroviral treatment (ART) significantly reduced the risk of transmitting HIV (Cohen et al., 2011; Cohen, 2011). Clinical studies combined with modelling showed that by scaling up ART, there was a possibility of bringing about a ‘phase change’ in the HIV epidemic, which would significantly reduce HIV mortality and incidence (Granich et al., 2009). This led to a paradigm shift wherein the discourse changed from treatment and prevention to treatment as prevention (Nguyen et al., 2011). In the aftermath of this shift, UNAIDS launched a massive campaign that projected the end of AIDS within 2030 through its ‘Fast Track’ Strategy (UNAIDS, 2014a) – alongside this strategy followed three important numerical targets which were to be reached within 2020, the 90-90-90 targets. These stipulated ‘that by 2020 90% of all people living with HIV will know their HIV status; 90% of all people with diagnosed HIV infection will receive sustained antiretroviral therapy; and 90% of all people receiving antiretroviral therapy will have viral suppression’. While the 2020 deadline for the 90-90-90 targets was not met except for a few countries globally (UNAIDS, 2020), UNAIDS, in their latest strategy, has doubled down and launched a new set of metrics where the target goals now are set to 95-95-95 to be reached within 2025 (UNAIDS, 2021).

Nevertheless, the global HIV response is currently witnessing increased pressures, such as ongoing budget cuts and political pressure to justify its existence vis a vis various other global health issues (Wexler et al., 2022). In this climate, data-driven initiatives have become crucial to securing funding and convincing donors to provide more economic support. As Sara Davis states, ‘the pressure on global health agencies to demonstrate their impact to increasingly reluctant bilateral donors, who directly finance global HIV programmes as well as their own bilateral HIV programmes in developing countries, has made granular HIV data an urgent imperative’ (Davis, 2017, p. 1149). Through indicators, recipient actors vying for funding can show that they are ‘on track’ towards the end of AIDS and thus use the indicators as proof of their work in progress and cost-effectiveness. Conversely, donors can use indicators as tools to audit the recipients and ensure that the funds are used according to the mandates set by the donors. Such an analytical optic aligns with the framework of ‘audit culture’ (Shore & Wright, 2015) and the neoliberal turn in global health.

Indicators in global health derive part of their power from three interlinked elements; ‘(a) commensuration (comparing performances on a common metric), (b) visual simplification (presenting performances in an appealing format), and (c) serialisation (framing performance as a continuous developing property)’ (Ringel, 2023, p. 189). These three elements are also found in the 90-90-90 targets and the discourse around the end of AIDS. Nations and other actors can compare their performance along the 90-90-90 continuum while at the same time offering neat visual representations of said performances through the numbers, which can be adapted to graphs, statistics, charts, tables, and other visualisation tools. Finally, the 90-90-90 targets depict the ‘road towards the end of AIDS’ as a continuous developing journey through annual reporting. This article seeks to provide an account of how we can think about indicators in the drive to end AIDS as doing work that is contingent upon commensuration, simplification, and serialisation. Indicators have their uses but might risk reducing complexity if not followed by more nuanced qualitative analysis that connects the target indicator with a social world wherein context becomes visible rather than erased. Vincanne Adams has argued that evidence-based medicine in global health and its dominance might risk erasing other ways of knowing and understanding health (Adams, 2016). My argument is akin to Adams in that indicator-driven efforts to end AIDS, while laudable and in many cases beneficial, might risk erasing and foreclosing other forms of conceptualising what the end of AIDS could be. These risks of erasure and foreclosure are linked to issues around commensuration, simplification, and serialisation.

This article seeks to provide a conceptual analysis of the role of performance indicators in the drive to end AIDS. As such, my empirical departure is not a set of clearly defined materials for analysis. Instead, I will focus on two actors, UNAIDS and PEPFAR, and then use material from UNAIDS reports, secondary research, and news sources focusing on PEPFAR's recent strategic shift to build an analytical relief. In this way, the onus is on analytical terms’ theoretical and conceptual usage, which takes centre stage and not an empirical set of materials per se. I do this to supplement earlier perspectives on the role and function of indicators in global health.

Prior scholarship on indicators and the end of AIDS

Prior research on the role of indicators and metrics within the drive to end AIDS has focused on a select set of problems. First, scholarship has focused on how indicators in the drive to end AIDS are plagued by epistemic uncertainty and what Sara Davis calls ‘the politics of the uncounted’ (Davis, 2020, 2017). In this line of research, the critique against indicator-driven and data-run campaigns to end AIDS, while appearing to be objective and transparent, is often patchworked and partial (Montoya, 2013). This line of scholarship has focused on how the reliance on data-driven efforts to end AIDS might risk erasing those communities that are most vulnerable to HIV and ill health in part through what has been called the ‘data paradox’ for key populations (Baral & Greenall, 2013). As Sara Davis states, key populations within the HIV epidemic often evade participating in data collection efforts out of fear of being identified by friends and family and exposed to potential discrimination and even prosecution (Davis, 2022, p. 209). This then compounds the data paradox, which Baral and Greenall describe as when ‘decision-makers deny that most affected populations exist, or that they are relevant to the epidemic; so, no research gets done on these populations; the lack of data feeds the denial; and so on’ (Baral & Greenall, 2013). One example of this, and how it feeds into the indicators on the drive to end AIDS, is the association between criminalising same-sex behaviour and implausibly low-key population estimates or no key population estimates at all (Davis et al., 2017). Small size or no size estimates might erase and deny the existence of LGBTQI + communities by governments and, at the same time, produce an image of performing better vis a vis the 90-90-90 targets. In addition, prior research has focused on the failure of indicator-driven HIV campaigns to address other issues, such as quality of life and human rights within indicators, as well as how data-driven indicator campaigns might lead to human rights violations and a politic of erasure (Davis, 2015; Amon et al., 2018).

Another stream of research has focused on the neoliberal aspects of indicator-driven discourses on the end of AIDS. This line of research focuses on how indicator and target-driven programmes to end AIDS are built on a logic of audit culture (Shore & Wright, 2015). Indicators and targets, as used in the drive to end AIDS and control the epidemic, are analyzed as disciplinary technologies used by governments and donors who use indicators as tools to steer recipients and their actions according to progress along the indicators (Montoya, 2013; Owczarzak et al., 2016). Such scholarship has focused on how reaching performance indicators in the drive to end AIDS produces specific dynamics between donors and recipients. Tim Rhodes and Kari Lancaster have recently argued that looking at how missing targets within the 90-90-90 targets to end AIDS is also essential (Lancaster & Rhodes, 2021). They argue that since most targets in global health are never met, looking at what missing targets do within the drive to end AIDS is as important as looking at other aspects of indicators in global health (Lancaster & Rhodes, 2021, p. 220). Their overarching claim is that global health is generally governed not by success but by failure and asks, ‘How might missed targets be seen as performative actors with governing potential in the constitution of health?’ (Lancaster & Rhodes, 2021, p. 221). As such, scholarship on indicators and the end of AIDS has stressed how reaching and even missing targets have become part and parcel of the power of global health indicators to end AIDS.

One might here get the sense that prior scholarship on indicators used in the drive to end AIDS have seen little room for local actors to resist indicator and target settings from influential donors, that indicators somehow are so powerful that no local forms of navigation or critique are left difficult if not impossible. However, indicators often become sites of critique and contestation (Davis, 2022, p. 286). As such, scholarship has pointed out that while indicator-driven efforts to end AIDS might use indicators as tools for audits and discipline, they also leave room for resistance and creative efforts to work with and through indicators at the local level (Montoya, 2013; Park, 2015).

However, only some scholars have focused on my explicit conceptual framework in this article. This article will supplement the above scholarship by focusing on three underlying epistemic aspects of indicators, which can yield important insights into how indicators act, what they do, and what they obscure. In the following, I will account for the analytical usage of the term’s commensuration, simplification, and serialisation.

Conceptual inspiration and theoretical extensions

The concepts of commensuration and comparison, as David Nelken notes, emerged in the early modern period (Nelken, 2021b). A comparison shows that something can be compared, while commensuration indicates that there is a similar standard for measurement. As Nelken states, ‘Comparison requires knowing as much as possible about the matters being compared to highlight both the similarities and differences, whereas this will be less important when commensuration is only interested in demonstrating what is or should be similar despite differences’ (Nelken, 2021b, p. 177). In practice, comparison and commensuration often overlap, and both concepts feed into each other. In global health, numerical indicators have become powerful tools that offer comparison and ways of making the world commensurable, i.e. imposing common yardsticks and measurements for different actors to be evaluated and ranked against. Espeland and Stevens’ argument about the power of commensuration is that such processes are at their core about ‘reducing and simplifying disparate information into numbers that can easily be compared. This transformation allows people to quickly grasp, represent, and compare differences’ (Espeland & Stevens, 1998, p. 316). The upside of commensuration is that it allows for comparability and standardisation; the downside is that it reduces complexity and flattens differences, which might be essential to note.

Finally, commensuration produces an abstraction of the world, simplifying complexity and subsuming heterogeneity into homogeneity under the new rubric of a commensurable category (Davis et al., 2012a, 2012b). As I argue later, the 90-90-90 targets (now the 95-95-95 targets) make disparate epidemic settings commensurable and comparable by focusing on a set of targets for the world to reach within 2020.

A particular critique levied against the use of social indicators both in global health and elsewhere is that processes of commensuration and comparison done solely through metrics fail to capture local contexts, a particular concern for anthropologists working in global health (Davis et al., 2012a). Another consequence of using indicators as tools for comparison and commensuration is that they almost always imply ranking and power. As Davis et al. states, ‘all indicators are fundamentally comparative, and some element of ranking is a feature of the indicators we are studying’ (Davis et al., 2012a, p. 8). These elements of ranking through commensuration and comparison ultimately allow ‘certain actors to exercise influence over the conduct of large numbers of geographically dispersed actors, that are readily adapted to forms of governance outside or reaching across distances beyond the state’ (Davis et al., 2012a, p. 11). The most essential critique against social indicators in global health as seen through the lens of ethnography and local HIV activism, is how processes of comparison and commensuration through indicators efface context and locality. As Merry and Wood state ‘to be globally commensurate, they [indicators] cannot be rooted in local contexts, but to accurately reflect local situations, they need to be’ (Merry & Wood, 2015, p. 217).

Moreover, ‘context becomes moot because global social indicators are involved both in the practice of comparison, learning about similarities and differences, while also being linked, at the same time, to the goal of commensuration, seeking to rank performance and make matters come into line’ (Nelken, 2021a, p. 216). However, this is not to say that indicators and targets within the global HIV effort are unnecessary. This point will be revisited later in this article as I will try to argue that the end of AIDS, enabled through global health indicators such as the 90-90-90 targets, enables possibilities while foreclosing others.

A related issue to commensuration is what Ringel calls visual simplification (Ringel, 2023). Numbers do not exist independently of the shape or format they are given. Indicators are often visualised in particular manners, giving them an aesthetic form often omitted from global health indicators (Werron & Ringel, 2017). Visualisation of numbers is a key part of global politics (Freistein & Gadinger, 2022). In global health, mobilising graphs, charts, and tables are potent tools in producing what counts as evidence and serving as powerful technologies for convincing actors to join a cause while arguing for action to be taken (Gerrets, 2015). Indeed, scholarship has been done on this in campaigns aimed at ‘ending hepatitis’ (Lancaster & Rhodes, 2020), and the work of Leclerc-Madlala, Broomhall, and Fieno (Leclerc-Madlala et al., 2018) have demonstrated how the use of graphs and charts have been integral as evidence which have been mobilised for what they call a ‘final biomedical triumph’ within the campaign to end AIDS. A typical manner of visualising indicators and rankings is through using league tables (Ringel et al., 2021) usually ordering numbers from top to bottom, with the best at the top and the worst at the bottom. Through the visualisation of indicators, ‘the often-complex calculative practices that undergird rankings are transformed into a simplified and easy-to-understand “full picture” that invokes a competitive spirit and is therefore designed to pit the ranked against each other’ (Ringel et al., 2021, p. 8). League tables end up erasing ambiguities and nuances through clear-cut ranks and world maps, often shaded in different colours based on indicators and metrics of performance, disease incidences, or mortality rates, producing a simplified yet easy-to-understand world (Ringel, 2023).

However, Bandola-Gill, Grek, and Ronzanio have recently argued that the visualisation of indicators and their rankings more and more have turned away from the value-laden ‘winner and losers’ paradigm. They have argued that in the digital era of digital dashboards and interactive data displays, we are seeing a “move away from league table formats toward multivocal interactive layouts that seek to mitigate the competitive and potentially dysfunctional pressures of the display of ‘winners and losers” (Bandola-Gill et al., 2021, p. 28). Moreover, they state that visualised indicators in the age of the Sustainable Development Goals (SDGs) serve as “alignment devices’ that entice country buy-in and seek to align actors around common global agendas” (Bandola-Gill et al., 2021, p. 28). This is a point I will return to later in the analysis of league tables and indicators in the drive to end AIDS but suffice to say that visual simplification of indicators produces effects which on the one hand might risk erasing nuances and local context while on the other hand, such simplifications might do productive work in rallying actors to act in the name of ending AIDS.

The final analytical perspective I want to draw on is what is called serialisation, which is an integral part of indicator-driven global health. A key part of indicator-driven global health is regular reporting and publication of ‘performance’ as measured against a specific indicator or numerical goal. Regular publication of indicator performance produces a serialised narrative wherein nation-states can be tracked across time and show how they either ‘climb’, ‘stagnate’ or ‘fall’ as measured by the indicator (Ringel, 2023, p. 193). Ringel and Werron have proposed the concept of serialisation to conceptualise the connection between time and social order (Ringel & Werron, 2021) and thus between normative understandings of ‘good’ and ‘bad’ performers in global health. The foundation of producing serialised reports on indicator performance starts with ‘constructing a narrative of performance as volatile and elusive property’ (Ringel, 2023, p. 193) that needs repeated measuring.

Furthermore, the need to know ‘trends’ across time to give predictive power to the indicators in global health is an integral part of the serialisation of global health indicators. Serialisation of regular performance indicators in global health allows for establishing the past, the present, and a potential future to come. In the case of the drive to end AIDS, UNAIDS and PEPFAR both publish annual reports mapping progress on key indicators in the drive to end AIDS, producing a serialised narrative showcasing trends over time while, at the same time, visualising nation-states and their performance in league tables, a point I will return to later.

Eliminating, controlling or ending AIDS: Counting towards a future without AIDS

So, what does ‘ending AIDS’ entail when seen through the lens of the 90-90-90 targets? The traditional way epidemiology has conceptualised disease endings is often refracted through three principal terms: eradication, elimination, and disease control (Dowdle, 1998). Eradication is defined as a ‘permanent reduction to zero of the worldwide incidences of infection caused by a specific agent as a result of deliberate efforts’ (Dowdle, 1998), while elimination is defined as a ‘reduction to zero of the incidences of infection caused by a specific agent in a defined geographical area as a result of deliberate efforts; continued measures to prevent re-establishment of transmission are required’ (Dowdle, 1998). The end of AIDS, however, is much more ambiguously framed and is perhaps much more reminiscent of ‘disease control’ as defined as ‘the reduction of disease incidence, prevalence, morbidity or mortality to a locally acceptable level as a result of deliberate efforts; continued intervention measures are required to maintain the reduction’ (Dowdle, 1998).

Indeed, ending AIDS as it is formulated both within UNAIDS and PEPFAR, two of the most important actors in the drive to ‘end AIDS’ in global health, is much more akin to epidemic control than elimination and a far cry from eradication.

In a footnote in UNAIDS’ latest strategy report it is noted that ‘ending AIDS’ “is used to refer to the full term ‘ending AIDS as a public health threat by 2030’ and that this is ‘defined as a 90% reduction in new HIV infections and AIDS-related deaths by 2030, compared to a 2010 baseline” (UNAIDS, 2021, p. 7). As such, the end of AIDS is less about ending and more about controlling AIDS in that the ‘end’ does not signify a reduction of cases and deaths to zero but instead focuses on reducing cases below a threshold benchmarked against 2010. PEPFAR, for their part, states that their strategies are aligned with UNAIDS’ 90-90-90 targets yet operate within a paradigm that explicitly focuses on ‘epidemic control’. This is defined as ‘the point at which the total number of new HIV infections falls below the total number of deaths from all causes among HIV-infected individuals’ (PEPFAR, 2017). Ryan Whitacre has written succinctly about the impact of the science behind treatment as prevention, its subsequent biomedical logic, and the ways in which this has influenced PEPFAR’s strategies for funding and, ultimately how indicators play a role in this (Whitacre, 2021). Whitacre argues convincingly that not only has the science behind treatment as prevention shifted PEPFAR's donor policies and programming, but perhaps more importantly, in the context of my work here, it has shifted which indicators PEPFAR views as the most important to track in order to achieve ‘epidemic control’ (Whitacre, 2021, p. 189). The indicator pair that now attracts the most attention is the one listed above, mortality among people living with HIV and population-wide HIV incidences. Whitacre has argued that PEPFAR’s shift towards achieving epidemic control through the logic of treatment and prevention shifted PEPFAR’s focus from one which in earlier times had focused on providing clinical services and clients reached through these to one which focused on health outcomes, such as suppressed viral loads, new cases of HIV averted and lives saved (Whitacre, 2021, p. 191). I want to highlight that as with UNAIDS’ definition of ‘ending AIDS’, PEPFAR's explicit focus on ‘epidemic control’ as a form of proxy for ‘ending AIDS’ relies on indicators solidly embedded within a biomedical logic and measurable health outcomes. While this in and of itself is not inherently bad, it can risk narrowing down the health programmes that receive funding from PEPFAR and others by allowing only one dominant logic and subsequently also only a narrow set of indicators to define what ‘success’ is and what the ‘end of AIDS’ is (Whitacre, 2021).

Moreover, as Assefa and Gilsk note, the ‘end of AIDS’ project as it is currently formulated, ‘focuses on HIV incidence, but ignores HIV prevalence’ and ‘therefore, [we] argue that there is fuzzy public health thinking and planning, and imprecise language use, which will lead to a dangerous threat to a sustainable response, in particular beyond 2030’ (Assefa & Gilks, 2020, p. 274). These are important remarks to remember when we talk about ‘ending AIDS’ as it is unclear what sort of ending, we are talking about in many ways. Part of this ‘fuzzy public health thinking’ and imprecise language can also be seen when we look at how the 90-90-90 targets operate through commensuration and comparability concepts.

Commensuration and comparability: The role of the 90-90-90 targets

The 90-90-90 targets, when introduced, were seen as a new, final, ambitious, but achievable target’ (UNAIDS, 2014a). Moreover, the 90-90-90 targets were lauded to end AIDS since ‘modeling suggests that achieving these targets by 2020 will enable the world to end the AIDS epidemic by 2030, which in turn will generate profound health and economic benefits’ (UNAIDS, 2014b). The 90-90-90 targets became the common yardstick against which all nations could track progress regarding the end of AIDS. Through a set of metrics, ‘ending AIDS’ is represented as a “reality in a universal format which allows it to circulate and be further calculated and formatted. Through numbers, specific places are turned into abstract calculable spaces, which can be compared, ranked, variously organised, and governed ‘from a distance” (Samiolo, 2012, p. 382). Indeed, the end of AIDS as mediated through the 90-90-90 targets allows for governing the end of AIDS at a distance from what has been called ‘centers of calculation’ (Latour, 1987), such as Geneva and Washington D.C.

Many matters complicate the comparison of progress along the 90-90-90 targets; data collection can vary from year to year, and the granularity and processing of data can differ depending on various factors such as funding, political priorities, and technical systems, factors researchers have already pointed to (Levi et al., 2016). As Granich et.al warns, the ‘heterogeneity in country-specific methods for monitoring the continua of care creates complex issues when comparing progress towards targets across countries’ (Granich et al., 2017, p. 17). Levi et.al states that ‘there has been an absence of standardised reporting methodologies, stage definitions, and agreed structures to unify cascade data, making the comparison of HIV treatment cascades difficult’ (Levi et al., 2016, p. 2). Finally, and equally important is how ‘data invisibility’ become part of the issue here. Sara Davis notes, ‘While all UN member states agree to report on HIV data, as of 2016, only 20 out of 193 UN member states had ever reported any data on HIV among transgender people to UNAIDS. None of these 20 countries is in Sub-Saharan Africa’ (Davis, 2017, p. 1154). This is also reflected in data concerning people who inject drugs; Davis notes that ‘according to UNAIDS, only 15 out of 55 African states has a current size estimate for people who inject drugs’ (Davis, 2017, p. 1154). Without the collection of such data, the indicators that should be able to track the progress toward the end of AIDS will be missing many of the most vulnerable and key populations, leaving them behind on the road to the end of AIDS. However, collecting such data is not without its own risk, as scholars have noted; many of the most vulnerable communities in the HIV epidemic are also communities that are criminalised, prosecuted, and stigmatised, and as such, have good reason to be hesitant of being inscribed into various granular data collection systems and HIV surveillance programmes (Kavanagh et al., 2020; Davis, 2020).

Another difficulty can be found within the last 90, viral load suppression. Here, the issue of defining what constitutes viral load suppression muddles the indicator. Levi et.al. found that definitions of viral load suppression vary across nations, with France defining viral load suppression as less than 50 HIV RNA detected per mL. In contrast, Russia defined this as less than 1000 HIV RNA per mL (Levi et al., 2016, p. 8). The point here is not to comment on the clinical significance of these differences but rather to note that these differences have an impact on how one can define the end of AIDS through indicators and progress towards it. By using different metrics for viral load suppression, progress towards achieving the 90-90-90 targets might look different. Brazil, as Levi et al. notes at the time of their writing, had 250.000 people living with HIV (35% of people living with HIV in Brazil) who had achieved viral load suppression of less than 50 HIV RNA per mL; however, by using the measurements of less than 1000 HIV RNA per mL, this would have increased the number of people in Brazil with viral load suppression to 40% or 293 000 people (Levi et al., 2016, p. 8). Such issues show how the last 90, ‘viral suppression’, is itself plagued by various epistemological uncertainties which problematise the indicator itself. This is well known to both UNAIDS and others working in global health, but the point here is to look at how the 90-90-90 targets are used as heuristics towards policy and the public. Davis et.al, have noted the same problem stating that ‘the degree of uncertainty beneath the surface of many of the most influential simplifying indicators in global governance is quite intensively scrutinised, but usually only in specialised scientific literature’ (Davis et al., 2012b, p. 9) alluding to the fact that while many of the weaknesses of indicators are well-known, they are often discussed in more specialised forums and less so when seen from the perspective of the public and policy. This perspective of the end of AIDS is, at the same time, a political act in that it produces a set of political priorities: a focus on HIV testing, access to ART, and viral load reductions across populations. These are all important facets of the end of AIDS, but it also means that how we come to understand what the end of AIDS could be is very much guided by a highly ambivalent biomedical discourse (Gaspar et al., 2022; Kippax & Stephenson, 2016) with a particular emphasis on health outcomes rather than other aspects of the HIV epidemic. These transformations produce various political effects that impact how the HIV effort unfolds. A case here would be a recent audit of PEPFAR, which showed that several respondents working in Kenya, Malawi, Tanzania, and Uganda reported that programmes and planning processes funded by PEPFAR were plagued by a culture of dictatorial, directive, and autocratic leadership (Igoe, 2020). Some of these issues related to target setting and indicator-driven programming wherein local and national staff stated that The Office of the U.S. Global AIDS Coordinator and Health Diplomacy (OGAC) more and more are ‘telling country teams what they must do and even what targets they must set for themselves. It is very much a one-way communication’ (Igoe, 2020). Such findings show how indicator-driven HIV efforts to end AIDS can become a top-down and donor-driven process, with little attention paid to local contexts and priorities. Moreover, as the audit in question found, ‘OGAC has set the targets using  …  estimates. It is “take it or leave it.” The targets have not been negotiable. [We] have hardly achieved the targets because they are unrealistic’ (Igoe, 2020). Such statements show the need for local insights into what targets are realistic based on local knowledge and contextual information, a weakness of top-down indicators. Another case could be drawn from PEPFAR's 2017-2020 strategy (PEPFAR, 2017), which marked a pivot away from its prior strategy to focusing on 13 priority high-burden countries to achieve epidemic control by 2020. As Devex noted, this shift in strategy was seen by many as problematic and raised a range of questions, some of which are linked to the issues of governing the end of AIDS by numbers. First, this new pivot was chosen based on ‘epidemiology – that they had a chance of reaching epidemic control by 2020, with the exceptions of Côte d’Ivoire and Haiti, which were added because PEPFAR felt it needed a West African country and a country in the Western hemisphere’ (Saldinger, 2018). In such a funding climate, showing that one is ‘on track’ to reach epidemic control to ensure further funding becomes paramount. Conversely, for countries who are ‘off track,’ governing by numbers can become a way of penalising nations that are not ‘doing their part’ and ‘reaching the numbers’. Such strategic shifts are enabled precisely through processes of commensuration and comparability using indicators. Indeed, part of the decision from PEPFAR was motivated by giving countries that were excluded from the new strategy a ‘wake-up call’ as per PEPFAR head Deborah Brix (Saldinger, 2018). Commensuration and comparability through indicators become disciplinary and power-laden rankings, which can become part and parcel of political decision-making processes with far-reaching consequences. Commensuration and comparison based on target setting are important parts of this analytical optic and remind us of the dangers of what might happen when indicators fail to be supplemented by qualitative and local data and knowledge.

Visual simplification and serialization: Indicators and the production of being ‘on track’ or ‘off track’ to end AIDS

Visualising progress along the 90-90-90 targets is a popular form of communicating to different stakeholders how progress toward the end of AIDS is going. A key format here is the traditional league table format wherein nations are listed according to their performance towards the 90-90-90 targets.

As they are called in UNAIDS reports, these’ scorecards’ report on different nations’ performance along the 90-90-90 targets. In these tables, nations are made ‘accountable’ in a very literal sense as these tables represent a world wherein ‘territories are refashioned as parts in relation to an imagined whole constituted by an ‘acceptable’ level of reduced infection’ (Rhodes & Lancaster, 2020). Moreover, such tables, as argued by Ringel ‘erase ambiguities and nuances by visualising an entire field of observation in clear-cut ranks’ (Ringel, 2023, p. 192). The power of such league tables is that in ‘contrast to publications such as statistical yearbooks, which require high levels of numerical competency on the part of the reader, global health indicators, when produced in this format, use aesthetic appeal to attract both expert and lay audiences’ (Ringel, 2023, p. 192). UNAIDS' produce scorecards visualizing each countries progress along the 90-90-90 cascade. The scorecards showcases how different nations are performing and list how close each country is in reaching the 90-90-90 targets. However, contrary to more traditional league tables, which explicitly rank performers according to their progress, the UNAIDS scorecards does not explicitly rank nations. The UNAIDS scorecards list each country in alphabetic order, and then the indicators are listed horizontally. Rather than listing nations according to ‘good’ or ‘poor’ performance by numerical ranking, the tables seems to produce a more ‘rank neutral’ visualisation of national progress towards the end of AIDS as measured by the 90-90-90 targets. Bandola-Gill et.al argues that league tables such as the one above, try to foreground ‘issue-based’ messaging (in this case, ‘ending AIDS’) instead of highlighting competition and peer pressure among nations (Bandola-Gill et al., 2021). Ranking visualisations, such as UNAIDS' scorecards, do not explicitly rank performers numerically. Rather they can be seen as aligning devices that seek to align actors with ‘diverse interests and interpretations of performance by allowing for the co-existence of multiple, often contradictory interpretations of one ranking’ (Bandola-Gill et al., 2021, p. 30). Akin to the arguments made by Leclerc-Madlala et al. (2018) on how the slogan of ‘ending AIDS’ produced a common language which diverse actors could rally around and take action to end AIDS, so too do these league tables produce a visualisation which aligns actors towards achieving the same goals, reaching the 90-90-90 targets and subsequently the end of AIDS. Yet while the UNAIDS league tables does not explicitly rank countries according to their performance nor engage in any ‘naming and shaming’ that can be seen in other league tables, the colouring of the indicators signals an implicit ranking and hierarchy based on the country's performance. The colours shaded in green signal ‘good’ performance, while orange and pinkish light red colour signal ‘poor’ performance. As such, indicators are almost always inserted into a power hierarchy, one which always implies some sort of value judgment on performance. Visual simplification in the above performs several things; it provides an easy-to-see representation of the world of different countries and their progress towards the 90-90-90 targets while at the same time using numerical indicators as neutral devices to align different nations towards a common goal, the end of AIDS. However, the colouring of the various indicators shows that performance is evaluated, if not explicitly, then at least implicitly. Visual codes here come to connote social, moral, and political values as per Bandola-Gill et al. (2021, p. 35). Moreover, the visual simplification of the world in the above table obscures the ‘often-complex calculative practices that undergird rankings and transforms these into simple to understand visualization’ (Ringel et al., 2021, p. 8) of a complex epidemic. This is not to say that UNAIDS or PEPFAR and others, for that matter, do not know this, but this is not the point here. The issue here is instead what these visual simplifications do when they travel into policy and the public, i.e. the performative work that charts, tables, and graphs do (Rhodes et al., 2020). The focus on visual simplification of the league tables within the politics of ending AIDS also draws our attention to another important and related aspect of the power of indicators and data-driven reporting: the issue of serialisation. Regular publication of UNAIDS’ and PEPFAR reports might seem like a trivial matter. However, serialising indicators in tables, charts, and graphs often takes a long time to establish the social, technical, and discursive infrastructure (Ringel & Werron, 2020). By producing regular, serialised reports on the progress towards the end of AIDS, visualised by the 90-90-90 targets, UNAIDS’ represents the world where certain nations ‘climb’, ‘stagnate’ or ‘fall’ as per Ringel’s vocabulary (Ringel, 2023, p. 193). While UNAIDS never uses vocabulary such as climbing or falling, words such as ‘lagging’, ‘on track’, or ‘falling behind’ have been prevalent throughout the reports since the introduction of the 90-90-90 targets (UNAIDS, 2018, 2020, 2017). Once again, and in line with Bandola-Gill et.al’s observation on using more politically neutral language in ranking visualisation, UNAIDS’ language does not rank nations per se by following progress year by year and then explicitly calls out nations who climb or fall on rankings. The language is not of ‘winners and losers’ but rather nations who are ‘off track’ or ‘on track’ in the race to end AIDS. Kari Lancaster and Tim Rhodes have highlighted how such language and visualisation of performance in the case of hepatitis C elimination efforts creates a form of race towards elimination as a concrete space in time to be reached as well as a race in time and against time (Lancaster & Rhodes, 2020). Such representation and usage of indicators and their visualisation can only be achieved by serialising indicators. It also is contingent upon establishing a narrative wherein performance is so volatile and elusive that measurements must be conducted frequently (Ringel, 2023, p. 193) to accurately track the epidemic's development. By serialising progress along the 90-90-90 targets, trends can be discerned and disseminated to a broad range of actors who can easily track indicator performance over time by following the visual simplifications found in numbers, graphs, and charts. Perhaps the most important aspect of the serialisation of indicators and their visualisation in tracking progress towards the end of AIDS is that annual performance reports produce a sense of hope or despair based on indicator progress. Moreover, the serialisation of indicators in UNAIDS reports produces a narrative of numbers (Espeland, 2015), which tells a story and call for action to be taken by relevant stakeholders. In this way, Sara Davis’ argument that the 90-90-90 targets can be read as communicative signs is apt (Davis, 2022). For all their visual simplification and their flattening of contextual information, the 90-90-90 targets can also be seen as visual signs where critique and contest play out as well as rallying points for diverse actors to act towards the end of AIDS.

Who and what is left behind between commensuration and comparison?

Global health indicators are often filled with embedded shortcomings; the 90-90-90 targets are no different. The targets anticipate that there will be those who do not get tested, those who do not access treatment, and finally, those who will not achieve viral load suppression. As such, the number of people who will not reach the targets grows from one metric to the next. If we here use the denominator of all people living with HIV, then the 90-90-90 target anticipated shortcomings translate to 90-81-73; thus, the targets themselves acknowledge implicitly that 10% will not be tested, 19% will not receive treatment and 27% will not be virally suppressed. As Judith Auerbach has succinctly noted, ‘What about the 10-10-10’ (Auerbach, 2019) and their role in the ‘end of AIDS’?

Auerbach's point is illustrative of several important points in my argument. First, it shows that the ‘end of AIDS’, when mediated through the 90-90-90 targets is an uneven end for different populations. Data shows that it is often those who are most marginalised who are also the ones who are lost in the ‘treatment cascade’ and who are left behind on the road toward the end of AIDS. Significant geographical variations exist regarding who is reached with testing, treatment, and retaining services along the HIV cascade. For instance, adolescents and young adults are less likely to know their HIV status (Ajayi et al., 2020). Other examples would be female sex workers in Asia and the Pacific who are less likely to know their status (UNAIDS, 2018) and that among gay men in the U.S., there are significant gaps in testing which leads to African American and Latinx gay men being less likely to get a prompt diagnosis (Sheehan et al., 2017). These patterns of inequality and gaps are also found in the second 90 and the third 90 and are often mediated by race, gender, sexuality, socioeconomic status, and geography (Hall et al., 2019).

The three indicators have a clear biomedical orientation, and as such, the end of AIDS becomes dominated by biomedical endpoints (Kippax & Stephenson, 2016). This set of indicators, however, seems to obscure other aspects of what life with HIV entails, in particular, ‘quality of life’. This ‘fourth 90’, as it has been dubbed (Lazarus et al., 2016), was proposed back in 2016 as a supplement to the three original indicators. Even though the idea of a fourth 90 garnered much attention, even gaining traction in The Lancet with a special report on ‘beyond viral suppression’ (Lazarus et al., 2016; Serwadda, 2019), the fourth 90 was never included in the official UNAIDS set of indicators. The issue, as Politico describes, was that the term ‘quality of life’ seemed too difficult to capture and agree upon when deciding what kind of data should be represented within the indicator (Wheaton, 2021). Quality of life for people living with HIV includes both biomedical factors, such as diseases; mental health factors, such as depression and anxiety; social, such as housing, work, and human relationships; and political, such as stigmatisation and criminalisation. As Wheaton states, ‘The more momentum the 4th 90 effort seemed to build, the more elements’ people wanted it to encompass. That made the task of measuring more and more complicated’ (Wheaton, 2021). This shows the complexities of rendering what constitutes quality of life amendable to indicator-driven global health. The complexities of agreeing on what data points should be included to capture what quality of life is across populations and local contexts is a difficult task, showing how the commensuration of ‘quality of life’ is complex. It should be noted that the new 95-95-95 strategy from UNAIDS does include what seems to be proxies for the ‘fourth 90’ such as ‘less than 10% of people living with HIV and key populations experience stigma and discrimination’ (UNAIDS, 2021), and while this is an important indicator, it must be said that this is not the same as ‘quality of life.’

Another point worth considering when it comes to commensuration and comparability is that while the 90-90-90 targets are framed as a linear arrow of progress, the data collected on the indicators are collected as cross-sectional data sets. Hence, they only reflect what is happening in a reporting year. However, the HIV treatment cascade is often far from linear for many people, as anthropological work has shown; people can fall in and out of care and hence viral suppression can fluctuate in much more erratic manners than a linear progression of testing – starting treatment – achieving viral load suppression (McGrath et al., 2014). As Rachel Chapman argues in the case of Mozambique, in what she calls the ‘therapeutic borderlands,’ initiating ART, retention in care, and adherence are all complicated by austerity measures at the systemic level (Chapman, 2021). Such austerity measures at the structural level complicate any linear notion of the end of AIDS and make problematic the idea that we can neatly track, through indicators, the road towards the end of AIDS.

Concluding remarks: ending AIDS between commensuration and comparison or what indicators enable

In many ways, establishing global standards such as the 90-90-90 targets might obscure many of the dynamic and fluid local epidemiological situations. As such, processes of commensuration and comparison should be critically examined. First, they provide a powerful rallying point for various actors to focus on. Indicators’ work through commensuration and comparison allows for a ‘common language’ to emerge, which various actors can rally around (Leclerc-Madlala et al., 2018). Urueñ Newa has made a similar argument and states that indicators can facilitate interaction between groups, organisations, or ‘regimes’ that otherwise would face difficulties interacting with each other and finding some common ground for action (Urueña, 2015). The global HIV effort is an excellent example of precisely this point: through the 90-90-90 targets, actors ranging from local patients’ organisations to international NGOs, to UNAIDS, PEPFAR, and private pharmaceutical companies have a standard frame of reference of what the end of AIDS is.

In the drive to end AIDS, the power of indicators such as the 90-90-90 targets might lie not in their fidelity to describe the world accurately but in their ability to produce a common framework for understanding what the end of AIDS is or can be: in this case reaching three global indicators. This also shows how commensuration can be linked to the pursuit of common interest (Samiolo, 2012), expressed through a common numerical value.

As such, while commensuration through indicators can obscure local contexts, they provide the global HIV effort with a form of ‘communication tool’ in the drive to end AIDS, as Sara Davis has argued. Davis notes that the ‘end of AIDS’ is a form of storytelling narrated through various models, projections, strategies, and reports to compel and persuade undecided people to care and act (Davis, 2020). As such, commensuration and comparison through indicators enable specific affective modes to be activated amongst stakeholders. Indeed, while commensuration of the world through standardised indicators might risk producing a less detailed and less localised view of the global HIV epidemic, commensuration does ensure a continual response to the HIV effort globally. If the universal ambition of the 90-90-90 targets is to spread and communicate that the end of AIDS is possible through standardised metrics, commensuration and comparison are inevitable at some abstract level. As Nelken states, ‘Standardization means making things similar and spreading supposedly universal values. Insofar as the goal of global social indicators is to spread such standards, faithfulness to local contexts may be neither necessary nor possible’ (Nelken, 2021b, my italics). Here, ‘values’ can have a double meaning; it can mean spreading the universal value of seeing the end of AIDS come, a form of moral and humanitarian value. However, it can also spread the numerical values of the 90-90-90 targets. Governing by values in this case, is a double entendre, both humanitarian (ethical) and technocratic (numerical/technical). In a sense, commensuration enables a continual, albeit somewhat biomedicalized, HIV effort to still compete for funding and political attention in a world that sees competition for funding for global health issues more and more spares.

Sally Engle Merry was one of the keenest critics of indicator-driven global health and global governance (Merry et al., 2015; Merry & Conley, 2011) and her calls for more qualitative and ethnographic work to supplement indicator driven governance is still an important call. Engle Merry herself did, however, also recognise the difficulties of disregarding indicators completely as she stated that ‘we face the paradox that indicators need to be in local terms to measure correctly but to retain universal meanings to make comparisons possible across borders’ (Merry, 2016, p. 212) and that ‘we need to accept, the process of commensuration means that the translation will never be without slippage, but it could be less distorting’ (Merry, 2016, p. 217). This last point is important regarding the end of AIDS and the use of indicators such as the 90-90-90 targets. How these targets render the world commensurable and comparable means that in comparing progress toward the end of AIDS, we also need to recognise how indicators can distort our view of the world and what the end of AIDS will look like. In this optics, the end of AIDS lies between a global call for action and local knowledge about needs and wants, between commensuration and comparability.

Acknowledgements

The author would like to thank the editors of Global Public Health for their comments and thoughtful remarks on the manuscript. The author would also like to thank the two anonymous reviewers for their engagement with the manuscript and the final article.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.