How Can We Get Close to Zero? The Potential Contribution of Biomedical Prevention and the Investment Framework towards an Effective Response to HIV

Background In 2011 an Investment Framework was proposed that described how the scale-up of key HIV interventions could dramatically reduce new HIV infections and deaths in low and middle income countries by 2015. This framework included ambitious coverage goals for prevention and treatment services resulting in a reduction of new HIV infections by more than half. However, it also estimated a leveling in the number of new infections at about 1 million annually after 2015. Methods We modeled how the response to AIDS can be further expanded by scaling up antiretroviral treatment (ART) within the framework provided by the 2013 WHO treatment guidelines. We further explored the potential contributions of new prevention technologies: ‘Test and Treat’, pre-exposure prophylaxis and an HIV vaccine. Findings Immediate aggressive scale up of existing approaches including the 2013 WHO guidelines could reduce new infections by 80%. A ‘Test and Treat’ approach could further reduce new infections. This could be further enhanced by a future highly effective pre-exposure prophylaxis and an HIV vaccine, so that a combination of all four approaches could reduce new infections to as low as 80,000 per year by 2050 and annual AIDS deaths to 260,000. Interpretation In a set of ambitious scenarios, we find that immediate implementation of the 2013 WHO antiretroviral therapy guidelines could reduce new HIV infections by 80%. Further reductions may be achieved by moving to a ‘Test and Treat’ approach, and eventually by adding a highly effective pre-exposure prophylaxis and an HIV vaccine, if they become available.


Introduction
In 2011 a new Investment Framework for HIV/AIDS was proposed to guide efforts in the coming years towards the rational use of resources to confront the AIDS epidemic [1]. The Investment Framework called for all low-and middle-income countries to focus on a set of Basic Programs of proven effectiveness: i) prevention of mother-to-child transmission (PMTCT), ii) condom promotion and distribution, iii) programs for key populations (in particular sex workers and clients, men who have sex with men, people who inject drugs), iv) treatment, care and support for those living with HIV, v) voluntary medical male circumcision (in countries with low prevalence of male circumcision and high HIV prevalence), and vi) targeted behavior change programs. The Framework also called for country-specific decisions about the implementation of additional programs, called Critical Enablers, including program enablers (management, procurement, distribution, research, and program communications) and social enablers (such as outreach for HIV testing and counseling, advocacy, mass communications, community mobilization, and activities aimed at stigma reduction and the realization of human rights). The Framework also recognizes that programs should align with broad development objectives and, therefore, also support key development areas where synergies will be high. These include social protection for children, education, legal reform, gender equality, reduction of gender-based violence, poverty reduction, health system strengthening, community systems, and workplace programs.
The annual resources needed to implement this approach in 139 low-and middle-income countries were expected to increase to about US 24 billion by 2015 and decline thereafter due to increased efficiencies and a progressive reduction in disease burden due to decreasing morbidity and significantly reduced numbers of new infections. The full implementation of the Investment Framework would be expected to avert at least 12?2 million new infections and 7?4 million AIDS deaths by 2020, and thus provide a cost-effective means to achieve the goals of the 2011 United Nations General Assembly Political Declaration on HIV/ AIDS 2 for 2015, such as reducing sexual transmission by 50%, reducing transmission among those who inject drugs by 50%, and virtual elimination of mother-to-child transmission.
Substantial progress has taken place since the investment framework was first put forward in 2011. Key relevant areas include the potentially large secondary preventive benefit of treatment on HIV transmission, and the individual level benefit associated with earlier initiation of therapy. These finding are now reflected in the 2013 WHO treatment guidelines which recommends ART for all HIV+ adults with CD4 counts below 500 cells/ mm 3 , all HIV+ children below the age of 5, all HIV+ pregnant women and all HIV+ adults with active TB disease, co-infected with HBV with severe liver disease, or in serodiscordant partnerships [2]. Furthermore, there has been a renewed interest in the potential impact of 'Test and Treat', and eventually other new highly effective prevention technologies including preexposure prophylaxis (PrEP), and an HIV vaccine. Therefore, we undertook the present analyses to examine the potential impact of these strategies on the HIV epidemic through 2050.
Previous modeling work has investigated the potential impact of scaling up ART coverage [3][4][5], pre-exposure prophylaxis (PrEP) [6][7] and potential HIV vaccines [8][9]. The modeling results differ somewhat across different models but conclusions about magnitude and timing of impact have generally been similar. The Goals model, used for this analysis, produces results that are comparable to those produced by other models [10]. This study builds on this previous work by examining the impact of achieving high coverage of all existing HIV prevention interventions and three new approaches on the HIV epidemic in all low and middle income countries.

Methods
Similar to the work undertaken for the 2011 Investment Framework, we used the Goals model [11], part of the Spectrum software package, to model the potential impact of different interventions in 24 countries accounting for 85% of new infections in low-and middle-income countries as classified by the World Bank in 2012 (Table 1). Goals is an HIV transmission model that simulates the spread of HIV by modeling sexual contacts and needle sharing behavior. It is a compartment model that divides the population into different categories based on behaviors. Its structure is similar to other compartment models although the specific population groups included differ across models. Its structure is different from microsimulation models that create discrete populations of individuals and characterize each individual with randomly assigned characteristics based on population data. Both types of model can be used to investigate the impact of HIV interventions. The results are expected to be broadly similar when comparable inputs are used.
The adult population aged 15-49 is divided into 11 main risk groups, six for men and five for women: those who are not sexually active, those in stable partnerships with a single partner in the last year, those with multiple partners in the last year, female sex workers and their male clients, people who inject drugs and, men who have sex with men. The annual probability of a susceptible person becoming infected in a risk group is modeled as a function of the probability of encountering an infected partner, the number of acts per partner, the number of different partners per year, the proportion of acts protected by condoms, the prevalence of other sexually transmitted infections, the stage of infection of the infected partner, whether the male partner is circumcised, whether the infected partner is using antiretroviral therapy (ART) and whether the susceptible partners is using PrEP or has been vaccinated against HIV. Most sexual contacts take place between males and females in the same risk group but some proportion of contacts are between higher risk individuals and their low risk partners. Biomedical interventions (ART, male circumcision, condom use, STI treatment, PrEP, vaccination) act directly on the probability of transmission per act and the effect sizes are based on randomized control trials and other scientific studies. Behavioral change interventions influence key behaviors (proportion of acts covered by condom use, number of partners, age at first sex, needle sharing behavior) and the effects are based on studies of behavior change interventions. Those who are newly infected are tracked over time by CD4 count, age and ART status and are subject to non-AIDS mortality as well as CD4-dependent risk of HIV-related mortality ( Figure 1). Mother-to child transmission is determined based on fertility rates, HIV prevalence among women of reproductive age and the coverage of prophylaxis to prevent transmission. HIV-infected children are tracked by CD4 count, CD4 percent and ART status. Additional information on the Goals model and its application to these countries is described in File S1.
To model the impact of existing interventions we used the same coverage targets as in the Investment Framework (Table 2 and File S1 ''Assumptions about the Existing Investment Framework'').  While some countries have already met or exceeded some of these levels the assumption that all countries would achieve them by the target dates is very ambitious. We have used them in these scenarios in order to maintain comparability with the earlier Investment Framework calculations. In addition to the interventions in the Investment Framework, we now also model the impact of the implementation of the latest WHO guidelines regarding initiation of ART with high coverage targets (Table 2 and File S1 ''Assumptions about an enhanced Investment Framework'').
Several new approaches and technologies could contribute to future prevention efforts. First, we consider the provision of ART to all people living with HIV regardless of CD4 counts or clinical criteria (Test and Treat), which is currently recommended in North America and elsewhere [12][13][14] and has been shown to reduce infectiousness dramatically [15]. (While the term 'Test and Treat' has been used in very different ways by different experts and communities, for this paper we use the term to mean the provision of ART to people living with HIV who are not eligible under the new 2013 WHO treatment guidelines, with CD4. 500 cells//ml and not part of a defined population group.) Additionally, we consider several hypothetical scenarios where we explored the potential additional impact of a future highly effective Pre-Exposure Prophylaxis (PrEP) and, a preventive HIV vaccine. PrEP is available today both as oral pills and as vaginal gel but trials have shown widely varying efficacy rates, which have been attributed to suboptimal adherence [16][17][18][19][20][21]. Here we use the term PrEP to refer to all forms of ARV-based prophylaxis including the oral PrEP and vaginal gel formulations as well as new forms that may become available in the future, including long-lasting vaginal rings and injectable forms that are expected to improve adherence, and therefore overall effectiveness. We have assumed that wide scale implementation would wait until these new forms are available. Effective HIV vaccines are still under development and may not be available until around 2030, assuming current technological challenges will be overcome [22][23][24]. A cure may become an important component in the future response to AIDS. However, since this is still in early stages of development, we did not include this option in the current work. The assumptions for each prevention technology are shown in Table 3. For each technology we modeled a 'low' or pessimistic scenario using conservative assumptions about coverage for each technology and the introduction date for a vaccine, and a 'high' or optimistic scenario with higher coverage and earlier availability for vaccines. Full descriptions of the new technologies and the research behind the assumptions are given in File S1 (under Assumptions about New Prevention Technologies).
The impact of these new technologies was illustrated by modeling several different scenarios:

Results
If no progress is made towards meeting the coverage targets of the Investment Framework and current coverage of prevention and treatment interventions remains constant at 2012 levels we can expect the total annual number of new HIV infections in lowand middle-income countries to increase slowly mainly due to  Next, we consider the impact of adding each new approach by itself to the Investment Framework Enhanced scenario. Adding a 'Test and Treat' strategy might reduce new infections in 2050 by an additional 6-10% compared to the IFE scenario to 490,000-510,000 in the low and high coverage scenarios. Implementing a future highly effective PrEP intervention could reduce overall new HIV infections by 16-47% to 290,000 to 450,000 in the low and high coverage scenarios. Finally, under the hypothetical scenario where an effective preventive vaccine becomes available for widespread use by 2025 to 2030 new infections in 2050 would be reduced by 37-75% compared to the Investment Framework Enhanced to 140,000-340,000.
The combination of the IFE with 'Test and Treat', plus a future highly effective PrEP and a vaccine, could reduce new HIV infections in 2050 by 86% in the high scenario and 50% in the low scenario resulting in just 80,000 (50,000-110,000) in the high scenario to 270,000 (200,000-380,000) annual new infections in the high scenario. (The ranges represent the sensitivity analysis on the magnitude of reduction in infectiousness due to ART with 96% as the high estimate which corresponds to the HPTN 052 trial [15] and a low value of 60% assuming lower adherence in the general population.) The most optimistic scenario combining all present and future prevention technologies could reduce the cumulative number of new infections from 2013 to 2050 by 96%.
The number of AIDS-related deaths will also be reduced dramatically by these scale-up scenarios (Figure 3a Table 4. The costs range from 390 per QALY gained for the Investment Framework Enhanced compared to the original Investment Framework to 11,000 for the low vaccine scenario compared to the Investment Framework Enhanced. The cost per infection averted (not shown) is 6800 when comparing the Investment Framework Enhanced to the Investment Framework and 7300 when comparing the combined scenario with all three new prevention technologies to the Investment Framework Enhanced scenario.
Our results are highly sensitive to the assumptions we have made about the levels of coverage that could be achieved for each new technology and the effectiveness that would be possible in programs. We have thus included a sensitivity analysis that may give a broad indication on how the different assumptions may play out in terms of impact on new HIV infections. As mentioned above the estimated number of new infections in 2050 varies from 80,000 in the most optimistic scenario to 270,000 in the low or pessimistic scenario.

Discussion
The present analysis explored the magnitude of incidence and mortality reductions that might be possible by 2050 as a result of the implementation of the 2013 WHO treatment guidelines, and newer available and hypothetical biomedical interventions. We find that expanding treatment coverage expeditiously based on the 2013 WHO treatment guidelines is a critical foundational step of an effective strategy. While such an approach would be challenging to implement, if it can be successfully achieved it would result in marked decreases in progression to AIDS and premature deaths among HIV infected persons and reductions in new infections. Additional resources would be required to achieve this result but at US 390 per QALY gained this investment would be very cost-effective.
Embracing 'Test and Treat' could further contribute to a decrease in new HIV infections and to a lesser extent a decrease in progression to AIDS and premature deaths among HIV infected persons. The addition of 'Test and Treat' to the Investment Framework Enhanced may avert an additional 6-10% of new infections by 2050. Similarly, the addition of a future highly effective PrEP in a targeted fashion could prevent an additional 22% and a hypothetical highly effective preventive vaccine could prevent an additional 30%.
The impact could be much lower if high quality in program is not attained, underlying the sensitivity of our estimates to key assumptions about what can be achieved and which reinforces the need to main quality in programs. Although we have seen important success in preventing new infections over the past decade, the lack of greater progress partially reflects limited growth in funding and human resources constraints but also indicates the difficulty of achieving higher coverage of current interventions. The leveling off in international resources [25] since 2008 is of concern in this regard. Successful implementation of current approaches (such as outreach to high risk populations, condom use, male circumcision) requires demand for services among the populations at risk, political will to implement these programs, and continuing efforts to reach key populations year after year.
Expanding treatment in line with the 2013 WHO guidelines and Treatment as Prevention requires a substantial expansion in the number of people on ART while at the same time maintaining quality for both new and existing patients. It also requires a large expansion in testing to identify people living with HIV earlier in their infection and it requires willingness on the part of those with high CD4 counts to start treatment before symptoms of HIV infection appear. These factors have previously been identified as major drivers of both the cost and impact of expanded treatment programs, although the costs of outreach are not captured here [26]. If it is not possible to identify such a high proportion of the HIV-infected population or achieve high adherence to treatment and PrEP then these results cannot be achieved. Efforts need to be focused both on expanding coverage of the intervention as well as on ensuring high quality and adherence. Of note, over time the actual number of individuals who need ART converge for all strategies evaluated here; however, the more aggressive the ART roll out, the lower the number of AIDS related deaths.
Our study is limited by only considering a small selection of the possible strategies available. The modeling approach we have used makes a number of simplifying assumptions about patterns of risk behavior and the impact of interventions, but the model has previously been shown to produce results in qualitative agreement with other models making alternative sets of assumptions [27][28] which provides reassurance that these simplifications will not interfere materially with the conclusions drawn. Overall, our aim is to provide a useful estimate of the magnitude of impact that may be expected for a set of scenarios. The relationship between the chosen intervention scenarios and how real programs will evolve in the coming years is the most important source of uncertainity.
Our study shows that additional investment now would have a dramatic and lasting impact on the epidemic over a short period of time and on a sustained basis. Under ambitious assumptions, this could be expected to prevent 63 million AIDS related deaths, and 88 million new HIV infections in low and middle income countries by 2050 compared to no expansion of prevention and treatment beyond today's coverage.

Supporting Information
File S1 Contains further information on the assumptions for the Investment Framework and Investment Framework Enhanced, more detailed descriptions of Test and Treat, PrEP and HIV vaccines and the sources we used in defining their characteristics, and a full description of the Goals model, including the model equations. Figure S1. Risk Structure of Goals. Figure S2. Characteristics determining transmission of HIV. (DOCX)