Modeling the HIV Epidemic: Why the 95-95-95 Target and ART Effectiveness Parameters Matter

Abstract

HIV remains a major global public health threat with one million HIV-related deaths, two million new infections and over 1 million HIV-associated TB cases each year. However , population-based studies suggest marked declines in incidence , prevalence and deaths in the countries in East and Southern Africa that have expanded antiretroviral treatment (ART) at the fastest rate in the world. Previous research has indicated that most of the positive epidemiologic developments are indeed due to the rapid expansion of ART. However, many modeling groups' estimates of the impact of ART on these trends vary widely. This calls into question the ART efficacy, effectiveness and coverage parameters that modelers use to project HIV incidence and prevalence. We reviewed 2015-2016 global and national mathematical modeling studies regarding ART's impact on new HIV infections. We extracted ART and HIV transmission parameters (i.e., proportion diagnosed, proportion on ART, proportion on ART and virally suppressed, proportion on ART and not virally suppressed, percentage reduction in transmission for those on ART and virally suppressed, percentage reduction in transmission for those on ART but not suppressed, and retention). We then derived a model-specific ART effectiveness percentage that captured the aggregate impact of ART on transmission by 2020. We describe the available parameters and 2020 ART effectiveness calculations for 9 models and compared the two with the lowest and highest ART effectiveness. The ART effectiveness expressed as a percentage reduction in HIV transmission by 2020 ranged from 20% to 86%. ART effectiveness disparities between the highest (SACEMA) and lowest (GOALS) models for Mo-zambique are highlighted in Figure 1. The GOALS Mozam-bique model limits eligibility for ART initiation to 80% coverage of people living with HIV and with a CD4 + cell count below 350 cells/µL, assumes that ART reduces transmission by 80%, and that 70% of patients are retained, which yields a derived ART effectiveness figure of 20%. The SACEMA model, on the other hand, assumes 90-90-90 by 2020 (i.e., 73% viral suppression of estimated PLHIV), that ART reduces transmission by 96% in those on ART and virally suppressed, and by 88% in those on ART but not virally suppressed, yielding a derived ART effectiveness value of 86%. ART parameter selection and assumptions dominate, and low ART effectiveness translates into lower impact. The disparity between the models is striking, and the implications are significant: The more realistic parameters that yield higher ART effectiveness valuation suggest that the continued expansion of ART and support for sustainable viral suppression will make it possible to significantly reduce transmission and eliminate HIV in many settings.