key: cord-0028651-9762ax7z
authors: Kripke, Katharine; Eakle, Robyn; Cheng, Alison; Rana, Sangeeta; Torjesen, Kristine; Stover, John
title: The case for prevention – Primary HIV prevention in the era of universal test and treat: A mathematical modeling study
date: 2022-03-14
journal: EClinicalMedicine
DOI: 10.1016/j.eclinm.2022.101347
sha: 4a8e72a346f5ba1adaca5fb23686c625ce7c637a
doc_id: 28651
cord_uid: 9762ax7z

BACKGROUND: As antiretroviral therapy (ART) has scaled up and HIV incidence has declined, some have questioned the continued utility of HIV prevention. This study examines the role and cost-effectiveness of HIV prevention in the context of “universal test and treat” (UTT) in three sub-Saharan countries with generalized HIV epidemics. METHODS: Scenarios were created in Spectrum/Goals models for Lesotho, Mozambique, and Uganda with various combinations of voluntary medical male circumcision (VMMC); pre-exposure prophylaxis; and a highly effective, durable, hypothetical vaccine layered onto three different ART scenarios. One ART scenario held coverage constant at 2008 levels to replicate prevention modeling studies that were conducted prior to UTT. One scenario assumed scale-up to the UNAIDS treatment goals of 90-90-90 by 2025 and 95-95-95 by 2030. An intermediate scenario held ART constant at 2019 coverage. HIV incidence was visualized over time, and cost per HIV infection averted was assessed over 5-, 15-, and 30-year time frames, with 3% annual discounting. FINDINGS: Each prevention intervention reduced HIV incidence beyond what was achieved by ART scale-up alone to the 90-90-90/95-95-95 goals, with near-zero incidence achievable by combinations of interventions covering all segments of the population. Cost-effectiveness of HIV prevention may decrease as HIV incidence decreases, but one-time interventions like VMMC and a durable vaccine may remain cost-effective and even cost-saving as ART is scaled up. INTERPRETATION: Primary HIV prevention is still needed in the era of UTT. Combination prevention is more impactful than a single, highly effective intervention. Broad population coverage of primary prevention, regardless of cost-effectiveness, will be required in generalized epidemic countries to eradicate HIV.

As antiretroviral therapy (ART) for HIV treatment has expanded to reach the UNAIDS 90-90-90/95-95-95 diagnosis, treatment, and viral suppression targets and evolved into universal test and treat (UTT), 1 new HIV infections have declined, and the number of people living with HIV has increased. 2 However, these numbers have stagnated in the last 2 −3 years, according to global UNAIDS data. 2 The impact of HIV treatment on slowing new infections and improving life expectancy among those infected has been definitively proven, however, the role of primary HIV prevention in this era has not yet been clearly defined.

Mathematical modeling has supported the case for new HIV prevention modalities, such as voluntary medical male circumcision (VMMC) and HIV vaccine development. 3−6 These studies demonstrated the impact and cost savings that could be achieved by introducing these new interventions, leading to massive rollout of VMMC programs in 15 priority countries in Southern and Eastern Africa. Modeling of oral pre-exposure prophylaxis (PrEP) largely concluded that cost-effectiveness would be achieved when PrEP is delivered to those at the highest risk or in settings where UTT has not been fully achieved. 7−13 The WHO recommends that PrEP should be provided to populations at substantial risk of HIV, originally defined as those with incidence higher than 3% per year. 14 Practically, however, few geographic areas with such high incidence have been identified, and behavioral risk scores have demonstrated poor prediction of HIV incidence, 15 not to mention social harms and poor uptake associated with behavioral targeting. Even if precision targeting were possible, so few people would have access to PrEP, the resulting use would have little impact on epidemic control in the highest burden countries. 13 Modeling has previously shown that HIV prevention is needed alongside ART to reach epidemic control 16 ; however, cost-effectiveness analyses recommend high thresholds for investment. To date, newer prevention products have largely failed to reach the modeled levels of cost-effectiveness in the general population shown in the VMMC and vaccine modeling conducted prior to the expansion of UTT in resource-constrained settings. 4, 5, 17 This has led to a number of questions. What is the role of primary prevention in the era of UTT? Why isn't HIV prevention always cost saving, like VMMC was in the initial analyses published in 2011? What drives cost-effectiveness, in addition to unit cost? Previous modeling has examined the impact of scaling up biomedical prevention in the context of expanded treatment, 18 or has attempted to present an optimized package of interventions, including expanded ART, 19 but has not explored the role of prevention in the context of UTT, nor examined the specific characteristics of prevention interventions necessary for impact and costeffectiveness.

This paper seeks to explore responses to some of these questions using mathematical modeling of hypothetical scenarios in three sub-Saharan African countries. The scenarios presented here illustrate principles that can inform the rollout and further development of HIV prevention in the context of continued expansion and eventual maintenance of universal ART coverage.

No ethics approval was sought for this mathematical modeling study, which used only data that was previously published and did not involve any human subjects.

Lesotho, Mozambique, and Uganda were chosen as examples of countries with generalized HIV epidemics, with different characteristics in terms of HIV incidence and ART scale-up trends.

The Goals and AIM models within the Spectrum suite of models have been described in detail elsewhere, including equations, parameter ranges and fitting procedures. 18, 20, 21 Briefly, Spectrum is a suite of interacting dynamic, compartmental models, with an underlying demographic model. Spectrum/Goals (see Supplemental Figure 1 ) disaggregates the adult population ages 15−49 years into behavioral risk groups and estimates the impact and costs of scaling up HIV prevention and treatment interventions. The risk groups are defined as: (1) low-risk heterosexual (stable couples, defined as men and women reporting a single sexual partner in the last year); (2) medium-risk heterosexual

Evidence before this study

The authors searched PubMed in October 2020 for studies dating from 2005 to 2020. Search terms included "mathematical modeling" OR "cost-effectiveness" AND "HIV" AND one of the following: "HIV prevention," "voluntary medical male circumcision," "pre-exposure prophylaxis," and "vaccine."

Previous modeling has examined the impact of scaling up biomedical prevention in the context of expanded treatment or has attempted to present an optimized package of interventions, including expanded ART, but has not explored the role of prevention in the context of universal test and treat (UTT), nor examined the specific characteristics of prevention interventions necessary for impact and cost-effectiveness.

Primary HIV prevention is still needed in the era of UTT. Preventing infections now with available interventions is more impactful and cost-effective than waiting for the perfect intervention. Combination prevention is more impactful than a single, highly effective intervention. In settings where intervention targeting to the highest risk populations has minimal impact, one-time interventions such as voluntary medical male circumcision provided to the general population are substantially more cost-effective than those that need to be delivered recurrently.

(men and women with more than one partner in the last year); (3) high-risk heterosexual (female sex workers and their male clients); (4) men who have sex with men; and (5) male and female people who inject drugs. Spectrum is freely available, and the base Goals files are available upon request.

This set of analyses employs the Goals model calibrated to the nationally validated annual HIV prevalence and incidence estimates used by UNAIDS, produced within another module of Spectrum called AIM; the ART coverage and viral suppression rates also came from this source ( Table 1) . Behavioral data such as condom use, numbers of partners, and age of sexual debut are extracted from AIDS Indicator Surveys, Demographic and Health Surveys, and Population-based HIV Impact Assessment surveys. 22 , and viral suppression in infected partners) and comparing the estimated prevalence and incidence with survey and surveillance data, including their confidence intervals. Output uncertainty intervals around HIV prevalence and incidence are 95% plausibility bounds resulting from selecting alternative parameter combinations according to their goodness of fit to the data.

Three ART coverage scenarios were created for this analysis ( Table 2 ). The "2008" scenario was created to have similar assumptions to those used in earlier VMMC modeling. 4 In the "2008" scenario, ART and VMMC coverage were held constant at 2008 levels through 2050. In the "2019" scenario, ART coverage was held constant at 2019 levels through 2050, to represent what might happen if countries failed to achieve 95-95-95 targets. At the time this analysis was conducted, 2019 was the most recent year of validated program data for estimating ART coverage. In the "95-95-95" scenario, ART coverage was scaled up from 2019 levels to 81% in 2025 (representing 90% of HIV-positive people knowing their status and 90% of those people being on ART) and 90% by 2030 (representing 95% of HIV-positive people knowing their status and 95% of those people being on ART). All scale-up in this paper used linear interpolation between the indicated values. 

A selection of scenarios was developed for scaling up VMMC, oral PrEP, and a hypothetical HIV vaccine, then layered individually and in combination with the three ART scenarios. Table 3 shows the scale-up patterns for each prevention intervention. After scale-up, coverage of each prevention intervention was maintained at the indicated scale-up coverage value through 2050. VMMC was assumed to provide a 60% (51% −64%) reduction in HIV incidence among heterosexual adult males and was applied among males ages 15−49 years. 28−30 Oral PrEP was assumed to have 95% efficacy and 75% adherence, 31,32 resulting in an overall effectiveness of 71% (61%−81%). 33 Coverage levels for oral PrEP indicated in Table 3 were among medium-and highrisk heterosexual men and women, except for the "PrEP All" scenarios, in which coverage extended to all populations represented in the Goals model. Unlike vaccines tested to date, for the purpose of this analysis, the HIV vaccine was optimistically assumed to have 80% efficacy (reduction in acquisition) and a 30-year duration of protection. The vaccine was scaled up among all adults. The combination of prevention and ART scenarios is depicted in Table 4 . To calculate HIV infections averted and incremental costs, each prevention scale-up scenario was compared with the corresponding ART base scenario without any prevention scale-up. HIV infections averted and incremental costs were counted for 5, 15, and 30 years starting in 2012 for the 2008 scenarios and 2020 for the 2019 and 95-95-95 scenarios.

Fully loaded, provider side delivery costs of ART, VMMC, PrEP, and the vaccine were included in this analysis. All costs were converted to 2019 USD using Gross Domestic Product deflator values from the U.S. Department of Commerce Bureau of Economic Analysis website. 34 The unit costs and sources for each intervention for each country are listed in Table 5 . The PrEPit Cost Lite module, 35 which provides cost estimates based on an analysis of primary data from six countries, was used to estimate the per person cost for a full year of PrEP for each of the three countries. Default personnel costs available in the tool were used for each country. Lab costs from the PrEP Cost Model South Africa were available in the Cost Support tab of the tool. 36 The South Africa costs were converted from Rand to USD using the conversion rate for July 1, 2018 (mid-year) from www.xe.com, accessed November 3, 2020. Costs were averaged across the four populations listed in the tool. Other recurrent and capital costs were calculated from the default ratios available in the tool. Average per person per year antiretroviral drug costs (USD, 2017) for emtricitabine 200 mg + tenofovir 300 mg tabs for the Eastern and Southern Africa region were obtained from the UNAIDS HIV Financial Dashboard. 37 Both costs and HIV infections averted were discounted at 3% per year to facilitate comparison with similar analyses.

Intervention unit costs can vary substantially, depending on service delivery model, demand, implementation efficiencies, fluctuating commodity costs, and other factors. For the cost-effectiveness analysis, we varied the unit costs of each intervention − ART, VMMC, PrEP, and the hypothetical HIV vaccine − between 50% and 150% of the base costs listed in Table 5 . 43 Articles minimum and maximum cost per HIV infection averted across the cost sensitivity analyses.

Study sponsors played no role in the design or conduct of the analysis. All authors had access to the data and took the decision to submit the paper for publication.

What is the role of primary prevention in the era of UTT? Figure 1 shows how VMMC, oral PrEP, and a hypothetical highly effective HIV vaccine would have been projected to decrease HIV incidence when ART coverage was held constant at 2008 coverage levels in Lesotho, Mozambique, and Uganda. This set of scenarios is 42 Used the high end of the range of costs used in the study. similar to those used in the early VMMC and vaccine modeling exercises. In all three countries, HIV incidence was substantially impacted by each of the interventions, with the impact increasing in line with the assumed effectiveness of the intervention. VMMC scaled up to 80% coverage had higher impact than oral PrEP scaled up to 30% coverage, except in Mozambique, where the baseline male circumcision prevalence was substantially higher, and therefore the increase in male circumcision coverage when scaling up to 80% was less, compared with the other two countries. If both VMMC and PrEP were scaled up to 80% coverage, VMMC, with an assumed efficacy of 60%, had a lower impact than oral PrEP, with an assumed effectiveness of 71% (Supplemental Figure 2) . The HIV vaccine, with assumed efficacy of 80% and scaled up to 80% coverage, had the greatest impact of any of the individual interventions. Combining all three interventions brought HIV incidence down to nearly zero by 2050 in all three countries.

In Figure Table 4 . Figure 3 explores the additional projected impact attained by scaling up primary HIV prevention in the context of achieving the 90-90-90/95-95-95 targets for ART. Lesotho provides the most dramatic example, but the trends are similar in all three countries. Firstly, a highly effective vaccine, scaled up to 80% coverage of the general population by 2030, could reduce HIV incidence by an additional 70−75% by 2050, compared with the incidence in the 95-95-95 base scenario. Secondly, the combination of oral PrEP and VMMC could reduce HIV incidence by an additional 20% in Uganda, 22% in Mozambique, and 34% in Lesotho, compared with the 95-95-95 scenario. VMMC scale-up in the 95-95-95 scenarios was projected to have less of an impact in Mozambique and Uganda, since these countries are estimated to have already achieved relatively high male circumcision coverage by the end of 2019. Thirdly, the "PrEP All" scenario, in which oral PrEP is scaled up to 80% coverage of all populations, achieved the same HIV incidence reduction as the HIV vaccine scale-up scenario. This demonstrates that what is important for incidence reduction is not the specific platform (PrEP or vaccine), but rather how broadly within the population coverage can be achieved, along with the effectiveness of the intervention. And finally, the greatest impact was projected to be achieved by the PrE-P_VMMC_Vaccine scenario, demonstrating that a combination of different prevention options for different population segments is ultimately the most impactful strategy for ending the HIV epidemic.

To demonstrate the impact of different coverage levels, Supplemental Figure 3 shows the impact if PrEP were scaled up to 80% coverage in the PrEP, PrEP_VMMC, and PrEP_VMMC_Vaccine scenarios. Figure 4 demonstrates the impact of uncertainty around VMMC (51 −64%) and PrEP (61−81%) effectiveness, using Lesotho in the 2008 scenario (with the highest overall HIV incidence), and Uganda in the 95-95-95 scenario (with the lowest overall HIV incidence) as examples. Table 4 . Table 4 . Supplemental Figure 4 shows the magnitude of the uncertainty around the incidence estimate itself.

What if we do not reach 95-95-95?

The scenarios presented above assume that the 90-90-90 and 95-95-95 ART targets can be met and maintained. However, some countries are struggling to attain these goals, and disruptions such as the COVID-19 pandemic can cause treatment interruptions and delays in scaling up. Figure 5 shows that if the 95-95-95 targets are not met, primary prevention becomes even more important for attaining epidemic control. In Lesotho, continuing to scale up VMMC can mitigate increases in HIV incidence that would occur if ART were not scaled up beyond 2019 levels. In Lesotho and Uganda, if ART coverage is maintained constant at 2019 levels, scaling up both VMMC for men and oral PrEP for mediumand high-risk men and women can bring HIV incidence down close to what could be achieved from scaling up ART to 95-95-95 without further primary prevention scale-up, but in Mozambique, higher coverage of primary prevention would be needed to achieve the same HIV incidence reduction as scaling up ART to 95-95-95. As was demonstrated in the other two ART scenarios, adding a highly effective vaccine for the general population, either alone or in combination with PrEP and VMMC scale-up, can provide further reductions in HIV incidence in all three countries.

As shown in Table 6 , in all countries, across all time frames, one-time interventions (VMMC and the hypothetical highly effective, long-acting HIV vaccine) are substantially more cost-effective than oral PrEP, which needs to be delivered on a recurrent basis throughout an individual's period of risk. In all but two of the scenarios ( 

This modeling analysis demonstrates that primary prevention is still needed to reduce HIV incidence beyond what is achievable with ART scale-up, even in the context of significant ART contributions to decreased community transmission of HIV through progress toward Table 4 .

Articles the 95-95-95 targets. This is particularly important in countries that have been slow to scale up ART, as demonstrated in the 2019 scenario for Lesotho and Mozambique. Importantly, it is possible for primary prevention to decrease HIV incidence by an additional 70−75% compared with reductions that could be achieved by scaling up ART alone, particularly with highly efficacious prevention interventions delivered broadly to the general population at high coverage levels. In addition, it showed that combinations of prevention interventions with broad population coverage provide the greatest impact.

This analysis demonstrates that for one-time interventions, the cost per HIV infection averted decreases (cost-effectiveness increases) as the time horizon of analysis increases. Therefore, they are likely to remain cost-effective, or even cost-saving when the cost of the prevention is less than the cost of treatment over time. As incidence declines, higher cost interventions that need to be delivered on a recurrent basis, like PrEP, would only be cost-effective in individuals during phases of highest risk or in localities with high incidence. The size of the population in which PrEP would be considered cost-effective will shrink as overall incidence drops.

These findings provide important context for decision-makers and funders that are considering supporting novel HIV prevention interventions and planning for the future of large-scale HIV prevention and treatment programs. This analysis shows that cost-effectiveness findings from today cannot be compared with those published ten years ago, given significant changes in the epidemic where ART has expanded and HIV incidence has declined. Rather, it is critical to consider impact and cost-effectiveness of new HIV prevention interventions, and how they influence decision-making, differently than in the past. For instance, based on costeffectiveness modeling, the trend among funders in recent years has focused narrowly on HIV prevention among populations and geographies with the highest HIV incidence, resulting in limited reach and impact of interventions. While more expensive than other prevention modalities, oral PrEP has shown dramatic decreases in new infections, even controlling for contributions from ART, regardless of type of epidemic and imperfect adherence. 45 HIV infection today, even with an intervention that is relatively less cost-effective than other currently available interventions, is more cost-effective than preventing an HIV infection in the future, as prevention today will lead to fewer downstream infections in the future, as well as lower ART costs. With this in mind, if using theoretical cost-effectiveness thresholds as the only criterion for resourcing prevention, the field will never end HIV transmission in countries with generalized epidemics, even if the 90-90-90 and 95-95-95 targets are achieved. For countries in which a large proportion of HIV transmission is occurring among people who are not easily identifiable as "high risk," primary HIV prevention will need to be available generally, now, and with high coverage, to have a substantial, sustained impact on HIV incidence. Prevention provided to people with lower risk will be less cost-effective, but it may be necessary to achieve population-level impact.

A broadly available, comprehensive package of prevention options will provide the greatest impact, as roughly demonstrated in our scenarios combining VMMC, PrEP, and the hypothetical vaccine. For men and women this could include a method mix of condoms and lubricants, VMMC, varieties of PrEP (oral, topical, injectable), and eventually a vaccine. Even with a highly effective, long-acting vaccine available, additional impacts can be gained by providing multiple prevention interventions to different segments of the population, providing support for the concepts of choice and combination prevention.

Cost-effectiveness may vary for new biomedical prevention products depending on the product and mode of delivery. However, the cost of maintaining epidemic control, and even reaching eradication, must be considered. Even though the cost to prevent an HIV infection by any intervention may increase as HIV incidence decreases, discontinuing HIV prevention in the context of epidemic control would likely bring substantial risk for resurgence as people move around and contextual risks shift, potentially regressing to endemic levels of HIV in a given country.

This analysis highlights an important point for consideration as new HIV vaccine candidates are advanced in the pipeline: whether and how frequently they require periodic boosting. Recent trial results of longacting injectable cabotegravir foreshadow the possibility of PrEP delivery modalities that require less frequent administration, where a vaccine that needs continual boosting could be competing with long-acting PrEP. To be a game-changer compared to PrEP, a vaccine would need at least two of the following characteristics: (1) administered less frequently than the longest-acting PrEP delivery technology; (2) at least as effective as PrEP; (3) side effect, cost, and supply chain profile such that it can easily and affordably be provided to the general population.

As with any modeling study, the specific HIV incidence and cost-effectiveness values presented here are a function of the input data, which carry inherent uncertainties. Projections into the future are particularly uncertain, as both future HIV incidence and future costs depend on numerous factors that are impossible to predict. The cost-effectiveness values are sensitive to the relative costs of the various prevention interventions and ART. Since the costs of these interventions are not all derived from the same study, the methodology of determining the unit costs of the different interventions in different countries may not be comparable. Oral PrEP costs may go down as implementation efficiencies are identified, and its cost-effectiveness will also increase if, as was demonstrated in the SEARCH and other studies, 31,47−49 users correctly identify periods when they are at higher risk and use the intervention during these periods. 50 Despite these limitations, the overall trends identified in this study of three sub-Saharan African countries with different epidemic characteristics should be applicable to other sub-Saharan African countries with generalized HIV epidemics. Some of the conclusions, most notably the importance of providing primary prevention to people who are not "high risk," will not be applicable in settings with concentrated HIV epidemics.

Primary HIV prevention is still needed in the era of UTT. Preventing infections now is more impactful and cost-effective than waiting for the perfect intervention. Combination prevention is more impactful than a single, highly effective intervention. While it will be less cost-effective, primary prevention must be provided to lower risk populations in generalized epidemic countries for prevention to have an epidemiological impact. In these settings, one-time interventions such as VMMC are substantially more cost-effective than those that need to be delivered recurrently. Finally, the field will need to take treatment cost savings into account as it grapples with the cost of maintaining epidemic control and future eradication of HIV.

KK and JS are employed by Avenir Health and conducted this work under the above-cited grants from USAID through a subaward from FHI 360. RE, AC, and SR are employed by USAID. KT is employed by FHI 360 and contributed to this work under the above-cited grants from USAID. 

No primary participant data was utilized in this analysis; only publicly available data were used. The Spectrum suite of models is freely available at https://avenirhealth.org/. All model files used in this analysis will be available to anyone for any purpose upon request immediately following publication until five years post-publication. Requests can be directed to the corresponding author at kkripke@avenirhealth.org.

Fast Track: ending the AIDS Epidemic by 2030

UNAIDS. Global HIV and AIDS statistics-2020 factsheet. 2020

Commentary: disease modelling to inform policy on male circumcision for HIV prevention

Voluntary medical male circumcision: modeling the impact and cost of expanding male circumcision for HIV prevention in eastern and southern Africa

Modeling the impact of a partially effective HIV vaccine on HIV infection and death among women and infants in South Africa

Expected epidemiological impacts of introducing an HIV vaccine in Thailand: a modelbased analysis

What do we know about the cost-effectiveness of HIV preexposure prophylaxis, and is it affordable?

Evaluating the cost-effectiveness of pre-exposure prophylaxis (PrEP) and its impact on HIV-1 transmission in South Africa

Cost-effectiveness of pre-exposure prophylaxis for HIV: a review

The cost and impact of scaling up pre-exposure prophylaxis for HIV prevention: a systematic review of cost-effectiveness modelling studies

Comparative effectiveness and cost-effectiveness of antiretroviral therapy and pre-exposure prophylaxis for HIV prevention in South Africa

Cost-effectiveness of PrEP in HIV/AIDS control in Zambia: a stochastic league approach

Modelling impact and costeffectiveness of oral pre-exposure prophylaxis in 13 low-resource countries

Guideline on when to start antiretroviral therapy and on pre-exposure prophylaxis for HIV

Age-specific risk scores do not improve HIV-1 prediction among women in South Africa

The new role of antiretrovirals in combination HIV prevention: a mathematical modelling analysis

Evaluating the potential impact and cost-effectiveness of dapivirine vaginal ring pre-exposure prophylaxis for HIV prevention

How can we get close to zero? The potential contribution of biomedical prevention and the investment framework towards an effective response to HIV

Maximising HIV prevention by balancing the opportunities of today with the promises of tomorrow: a modelling study

Spectrum Model [software] v5.89 Beta 8: avenir Health

What is required to end the AIDS epidemic as a public health threat by 2030? The cost and impact of the fast-track approach

Lesotho: Ministry of Health and ICF International

and ICAP at Columbia University. Lesotho population-based HIV impact assessment (LePHIA)

Instituto Nacional de Estat ıstica (INE), and ICF. Inqu erito De Indicadores de Imunização, Mal aria e HIV/SIDA Em Moçambique

Agency For International Development (USAID) Uganda, and World Health Organization (WHO) Uganda. Uganda AIDS Indicator Survey. Kampala, Uganda: Uganda Ministry of Health

UBOS) and ICF. Uganda demographic and health survey

Uganda Population-based HIV Impact Assessment (UPHIA)

Male circumcision for HIV prevention in men in Rakai, Uganda: a randomised trial

Randomized, controlled intervention trial of male circumcision for reduction of HIV infection risk: the ANRS 1265 Trial

Male circumcision for HIV prevention in young men in Kisumu, Kenya: a randomised controlled trial

Lower than expected HIV incidence among men and women at elevated HIV risk in a population-based PrEP study in rural Kenya and Uganda: interim results from the SEARCH study

Uptake, engagement, and adherence to pre-exposure prophylaxis offered after population HIV testing in rural Kenya and Uganda: 72-week interim analysis of observational data from the SEARCH study

Preexposure prophylaxis for the prevention of HIV infection: evidence report and systematic review for the US preventive services task force

Table 1.1.7. Percent Change from Preceding Period in Prices for Gross Domestic Product

PrEP implementation planning, monitoring, and evaluation Tool (PrEP-it

In: Health Economics and Epidemiology Research Office (HE2RO) WHC, University of the Witwatersrand/Department of Global Health

UNAIDS. HIV financial dashboard

Economic evaluation of differentiated service delivery models for ART service delivery in Lesotho: cost to provider and cost to patient

Impact and cost of the HIV/AIDS national strategic plan for Mozambique, 2015−2019 −projections with the spectrum/goals model

Do DSD models for HIV treatment save money for health systems?

Projected economic evaluation of the national implementation of a hypothetical HIV vaccination program among adolescents in South Africa

Potential clinical and economic value of long-acting preexposure prophylaxis for South African women at high-risk for HIV infection

The anticipated clinical and economic effects of 90-90-90 in South Africa

Evidence of an association of increases in pre-exposure prophylaxis coverage with decreases in human immunodeficiency virus diagnosis rates in the United States

Lessons on PrEP from the SEARCH study in east Africa

High levels of preventioneffective adherence to HIV PrEP: an analysis of substudy data from the EPIC-NSW trial

HIV pre-exposure prophylaxis for men who have sex with men in west Africa: a multicountry demonstration study

Incorporating oral PrEP into standard prevention services for South African women: a nested interrupted time-series study

Potential impact and cost-effectiveness of condomless-sex-concentrated PrEP in Kwa-Zulu-Natal accounting for drug resistance

The authors would like to thank Jason Reed for insightful discussions over the years about the continued role of HIV prevention in the context of ART scale-up and Robert Glaubius for discussions of the sensitivity analysis and other helpful inputs. Editing was provided by Nicole Macagna at FHI 360.

Supplementary material associated with this article can be found in the online version at doi:10.1016/j. eclinm.2022.101347.