Background
South Africa has the largest human immunodeficiency virus (HIV) epidemic in the world [
1]. In 2012, 6.4 million South Africans were living with HIV; 203,000 individuals had lost their lives to it and another 395,000 South Africans had acquired the infection [
2,
3]. South Africa’s life expectancy was understandably adversely affected by the considerable burden of HIV disease [
4]. However, life expectancy had since increased from 53 years in 2006 to 61 years in 2012, and ensuring its continued improvement remains a priority of the national department of health [
5]. The gains made in improving life expectancy are in no small part attributable to ‘the largest antiretroviral (ART) rollout in the world’ that South Africa has managed to achieve [
6]. To sustain this achievement is no mean feat. The growing number of patients previously initiated on ART need to be retained in care. While the public sector retention rate approximates 75 % after one year on treatment, South Africa needs to continuously enroll in excess of 500 000 new patients onto ART annually to maintain an ART enrolment ratio exceeding 1.3 [
4]. This brings into question the long term sustainability of the ART program considering the massive financial and human resource implications the expansion of ART program entails [
7].
Data suggests that close to 25 % of all new HIV infections occurred among young women aged 15–24 years, emphasizing this group as a major driver of the epidemic [
2]. The HIV prevalence in this age group is important as it serves as a proxy for HIV incidence. HIV prevalence declined by 18 % in this age group from 2008 to 2012, from 8.7 % to 7.1 %, however there remains a need for intensified prevention efforts [
8]. Despite massive accomplishments made in establishing the ART program, the women aged 15 – 24 years persist as the group with the poorest access to this life-saving treatment. The barriers that young people face in accessing public health services has been well documented [
9]. Issues concerning lack of confidentiality and privacy, unfriendly and judgmental attitudes of health care staff and inaccessible clinic hours persist [
10,
11]. It was against this backdrop that the re-engineering of primary health care in South Africa targeted the development of a school-based sexual and reproductive health service as a priority [
12].
The current HIV prevention program has enjoyed limited success in tackling the high rate of new infections in South Africa, highlighting the need for an alternative intervention. Vaccines are regarded at the most cost-effective prevention intervention in the world [
13]. Rerks-Ngam et al tested the first HIV vaccine regimen (RV144/Thai Trial) to show moderate vaccine efficacy in humans in Thailand (2009) [
14]. The study evaluated a prime-boost strategy, priming with a recombinant canarypox vector (ALVAC-HIV[vCP1521]) administered at baseline, then at week 4, 12 and 24 with recombinant glycoprotein 120 subunit vaccine (AIDSVAX B/E) boosts given with the ALVAC at weeks 12 and 24.
The prime-boost HIV vaccine regimen used resulted in modest efficacy of 31 % over 3.5 years [
14]. While the effects were not durable, they were indeed promising. After undergoing modifications to optimize the HIV vaccine regimen by making it Clade C specific and changing the protein and adjuvant, a potential vaccine regimen was entered into Phase I/IIb clinical trials at six major South African centers to assess safety and immunogenicity (HIV Vaccine Trial Network (HVTN) 100 study) [
15]. Additionally, a pivotal phase IIb/III HIV vaccine efficacy trial is planned to take place in South Africa designated HVTN 702, which will evaluate the same regimen [as HVTN 100], should HVTN 100 prove to be immunogenic.
The aim of this analysis was to guide decision makers in assessing the value of national implementation of a potential HIV vaccine among school-based adolescents in South Africa. The work determined the impact of vaccination on HIV disease burden and associated health costs, and evaluated the cost-effectiveness and potential changes in life expectancy based on the premise that school-based care would address the issues of equity and accessibility in health care that adolescent South Africa faces.
Discussion
The study aimed to assess the cost-effectiveness of national rollout of the hypothetical HIV vaccine to school-based adolescents. The South African HIV epidemic is widely acknowledged to be generalized, with adolescents and young adults disproportionately at risk for HIV [
37]. In 2013, South Africa reported 16 % of the global HIV incidence despite concerted efforts at the national level ranging from increasing ART distribution by 75 % between 2009 and 2011 to boasting the largest and most established condom distribution program in the world [
2,
38]. This earmarked adolescents as a key population to be reached if HIV prevention strategies are to impact incidence and if HIV mortality rates are to be significantly curtailed [
37]. While the introduction of a potential HIV vaccine in schools represents a significant financial investment, the health outcomes in terms of improved life expectancy, markedly decreased potential years of life lost and decreases in HIV mortality and incidence are substantive. Life expectancy was equally influenced by vaccine coverage rates, while the assessment of cost-effectiveness was found to be sensitive to the vaccine efficacy.
The life table findings together with the conventionally accepted thresholds for cost-effectiveness being met demonstrate the financial plausibility of HIV vaccine implementation [
19]. Importantly, the vaccine remained cost-effective even at higher prices per dose examined but at substantially greater programmatic costs. Annual HIV vaccination represents a substantial increase in costs per capita at base coverage of 60 % of HIV negative adolescents. This constitutes a significant investment considering the intense competition of several competing burdens of disease on a constrained South African health budget [
39]. As much as the long term financial sustainability of the burgeoning ART program has been brought into question, the implementation of a HIV vaccine program over several decades may prove equally daunting. It is important to bear in mind that the comparator cost reflects those currently on treatment (excluding the treatment shortfall of approximately 58 % [
1]) and thus represents a gross underestimation of what we should be paying if those unable to access treatment were indeed able to access it. Another major consideration is that the upscaling of ART may not impact the HIV incidence as definitively as a primary preventative strategy may. It must be remembered that while averting infections has a cost attached from a government perspective, it may also give rise to the substantial financial gains of reducing the demand for ART [
40].
South Africa has successfully negotiated reduced pricing for ART and HPV vaccines in the past, and this bodes well for future procurement of HIV vaccines [
23,
24], as the price is undetermined at this point. If vaccine development fails to reduce the number of annual boosters required to maintain protection, then the pricing represents a key factor in deciding the cost-effectiveness of the intervention. Apart from the economic impact, HIV vaccine implementation has the capacity to influence long term health outcomes. The mean cumulative gain of LYG could support efforts to improve life expectancy in the country, an area identified as a strategic output of the National Service Delivery Agreement [
5].
The South African epidemic is predominantly heterosexual. This work represents an over-simplification of the rather complex sexual networking structures at play in the South African HIV epidemic. Nonetheless, those individuals at high risk may still acquire infection ascribed to repeated risk exposures despite the protection conferred by the vaccine compared with those at low risk. At a population level, the premise remains that a partially effective vaccine may still avert or delay infection even if it is unable to completely prevent an infection from establishing [
41]. Assessment of a partially effective vaccine in the United States of America (USA) emphasizes that even modest and temporal reductions in HIV infections have important benefits at the population level [
42]. Andersson et al demonstrated similar health benefits to the USA study when modelling the RV144/Thai trial vaccine in South Africa, but cautioned that a vaccine of limited duration could only be effective with high coverage levels, which translated to millions of doses [
43].
Adolescents are a critical target for this intervention. Apart from being a key population identified in the transmission of HIV, adolescents in a school environment appear more easily accessible as a target group considering that more commonly identified high risk groups such as commercial sex workers are often harder to reach due to stigma and marginalization [
43]. However, adolescents have historically encountered barriers in trying to access health services in South Africa from confidentiality issues to the judgmental attitudes of staff. It is not surprising that they often do not return for follow up care [
9]. The school environment could be deemed a “safe space” for peer discussion and accessibility of relevant health services. Neglecting the comprehensive health needs and barriers to care of this adolescent population has the potential to undermine the success of HIV prevention initiatives [
44]. Further, low social acceptability of HIV vaccines fueled by the fear of vaccines and poor side effect profiles present potential deterrents to uptake and coverage [
45]. It is understandably difficult for hypothetical scenarios to emulate real-life behavioral changes but knowledge of these factors underscores the need for comprehensive sexual education and risk reduction counselling; which could prove more plausible in the school environment [
46].
This study had several limitations. Firstly, it is unclear to what degree behavioral disinhibition may occur following vaccination as this was not assessed in the model. Changes in sexual risk behavior post HIV vaccination are poorly understood in the African setting [
46]. In high HIV prevalence communities like South Africa, a decrease in condom use even with stable partners would likely result in an increase in HIV rates [
46]. In fact, South African data has inferred that poor comprehension of the ‘low-efficacy’ concept was associated with a reported potential decrease in condom use. It is further postulated that the degree of behavioral disinhibition may depend largely on the manner in which the vaccine effects are marketed to the public and vaccine recipients alike [
40]. The impact of risk compensation becomes critical when considering the low efficacy displayed by the candidate vaccines thus far [
46]. Secondly, the study was unable to assess the effects of herd immunity. Notably, Long et al. alluded to partial efficacy vaccines providing some benefits to the unvaccinated population through herd immunity [
42]. This is particularly important considering the low coverage rates of childhood vaccinations in South Africa as it speaks directly to the country’s capacity to introduce and implement a HIV vaccine [
22]. At 60 % coverage, this program calls for an unprecedented 5.9 million adolescents to be vaccinated. Given this, it is not surprising that implementation costs are high. Thirdly, the provider perspective was considered as the largest burden of direct medical program costs will be borne by the healthcare sector. Although the societal costs were not analyzed, its contribution would be substantial and could improve the overall cost-effectiveness of the vaccine. Fourthly, booster vaccinations were not assessed in the original RV144/Thai trial work [
42]. Therefore the assumption that booster vaccination would provide the same protective effects as the initial vaccination was hypothetical. There has been limited description of this in the literature [
46]. Additionally, administration costs would drastically increase the program costs with the need for annual boosters. This is the key cost factor implicated in the difference between the comparator and intervention cost. However, it is hoped that attrition rates of vaccine recipients would be minimized by targeting the relatively stable school population. Lastly, this study has considered HIV vaccination as an isolated intervention apart from the ART rollout and condom distribution. In the clinical setting, this intervention would probably work synergistically with other prevention strategies such as male medical circumcision and an optimal combination of strategies should be better defined once data becomes available [
40,
42]. As discussed earlier, the limited success achieved in curbing the national HIV incidence by the current public sector HIV prevention strategies warrants the evaluation of strategies on its individual merits.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
NM, MYB and GEG contributed to the conception of the study. MYB provided guidance and supported the co-ordination of the study. NM and MYB contributed to the statistical analysis. NM was responsible for the overall drafting of the manuscript. All authors contributed to critically revising its content. All authors read and approved the final manuscript.