This is a joint post with Mead Over and Denizhan Duran.
In mid-2011, one of the biggest developments in HIV/AIDS research took place. The HPTN 052 study found that early antiretroviral therapy treatment could reduce HIV transmission by 96% in couples where one partner is HIV positive and the other is HIV negative. The study was heralded as the breakthrough of 2011 by Science, and was hailed as a game changer by many others, including UNAIDS, The Economist and The Lancet. The World Health Organization wrote a comprehensive guideline for TasP, or treatment as prevention, in June 2012, asserting that “TasP needs to be considered as a key element of combination HIV prevention and as a major part of the solution to ending the HIV epidemic.”
Further, at the 2012 AIDS conference, the HPTN052 team presented their models assessing TasP in India and South Africa as “very cost-effective” as compared to delayed ART or no ART, using the WHO “rule of thumb” for cost-effectiveness (<1x or <3x an average GDP per capita). The strategy is already gaining traction on the ground: the Clinton Health Access Initiative announced plans to scale up TasP to 90% of the relevant population in Swaziland.
Yet even if TasP is cost-effective according to the GDP per capita rules of thumb, is it the most cost-effective use of resources to combat HIV/AIDS? When and how much TasP versus other interventions are needed? These critical policy questions can only be answered by mathematical models that optimize the mix of interventions and their coverage for HIV prevention for a given budget, not models that compare the cost-effectiveness of a new intervention to no intervention (no ART) or the status quo (delayed ART) with no budget constraint.
Till Barninghausen, David Bloom and Salal Humair have recently published the results of such a modeling exercise for South Africa, assessing “whether TasP is indeed a game changer or if comparable benefits are obtainable at similar or lower cost by increasing coverage of medical male circumcision (MMC) and antiretroviral treatment (ART) at CD4 <350/microliter” (see here, gated). The authors find that MMC is significantly cheaper than TasP in terms of cost per infections averted – $1,096 versus $6,790. Further, they find that most benefits result from high levels of ART coverage using the CD4 <350/microliter criteria.
Barninghausen et al conclude that the most cost-effective HIV prevention strategy in South Africa is first to increase MMC coverage, and then to scale up ART: this strategy would achieve comparable incidence reductions to TasP and cost $5 billion less, a figure comparable to total Global Fund expenditures on AIDS over a five year period.
While the Barninghausen et al study looks at all the benefits of TasP, there are certain benefits and costs they overlook. There are significant differences of benefits across different risk groups: for example, if MMC is targeted to the most at-risk men, including soldiers, policemen, truck drivers and MSM, it would yield higher benefits (for more on this by Mead Over and others, see here and here). Similarly, for costs, there are issues of diseconomies of scale in the production of TasP (South Africa; here). There are also certain tradeoffs, or production synergies, that have not received sufficient attention: a study from Burkina Faso shows that adding ART to existing health services did not increase the waiting time for non-ART services in Burkina Faso. Finally, there could be diminishing returns to program scale up due to the saturation of groups most willing to receive and use the intervention – this could apply both to TasP and MMC.
All in all, this evaluation shows that as new interventions become available, the need to assess and invest in the most cost-effective mix and coverage of interventions becomes more relevant. The UNAIDS investment framework is a start. However, if we want to get the most done with the money that we have available, reaching an AIDS-free generation, we need to maximize our impact on HIV incidence reduction – getting to the intervention mix and coverage levels that are optimal for that purpose, given resources available.
The Barninghausen et al approach is the way to go for the future. To maximize value for money, global health funders and technical agencies should support the development of these models –in close collaboration with country officials- for all of their major recipients.