Background: The South African government currently discusses further expansion of eligibility for antiretroviral treatment (ART), including to all adults with CD4 cell count <500 cells/micro (500), the positive partner in serodiscordant couples (DC), all HIV-positive pregnant women (PW), or all HIV-positive individuals (UTT). Methods: We modelled 20-year cost and cost effectiveness of each of the four eligibility options under up to three different patterns of testing and care uptake: status quo, SQ; increased testing and immediate ART initiation amongst 80% of the (eligible) population (universal expansion, UE); and, for the DC and PW scenarios, coverage of 80% of the targeted sub-population, with the rest of the population continuing to access testing and care at the current level (prioritized expansion, PE). Data on the number of HIV tests, HIV infections, and patients in care came from EMOD-HIV, a microsimulation model of HIV transmission and treatment; data on the cost of testing, outpatient and inpatient care from our own bottom-up cost analyses in South Africa. We varied staff, space, equipment and overhead cost with the expected size of each clinic and included the cost of creating and sustaining the requisite incremental demand for testing and treatment. Results: Under CG.SQ, 3.4 million adults and 135,424 children will be on ART by 2033; under CG.UE, this increases to 5.3 million and 103,789, resp. All SQ scenarios cost around 33 billion USD, all UE scenarios around 45 billion USD, over 20 years. DC.SQ, PW.SQ and PW.PE are least costly, followed by 500.SQ and UTT.SQ. PW.PE is the most cost-effective scenario per infection or DALY averted. All UE scenarios have both greater cost and greater impact, and are more costly per infection averted. CG.UE has the highest cost per outcome of all scenarios. None of the scenarios is cost-saving over CG.SQ, though most SQ sub-scenarios might become cost-saving over more than 20 years. Conclusions: Because of the long-term prevention benefits of earlier treatment initiation, increasing ART uptake under the current eligibility threshold of 350 cells/ microl will cost more over twenty years than would similarly improved uptake under a threshold of 500 or with no threshold at all.
Conference: International AIDS conference 2014, Melbourne, Australia