You are here
HIV Mortality by Care Cascade Stage and Implications for Universal ART Eligibility
Eran Bendavid1; Anna Bershteyn2; Andrew Boulle3;Jeffrey W. Eaton4; Timothy Hallett4; Daniel J. Klein2; Jack J. Olney4; Andrew N. Phillips5; Emma Slaymaker6; for the HIV Modelling Consortium Writing Group on ART Eligibility Guidelines
1Stanford Univ, Stanford, CA, USA;2Inst for Disease Modeling, Bellevue, WA, USA;3Cntr for Infectious Disease Epi and Rsr, Cape Town, South Africa;4Imperial Coll London, London, UK;5Univ Coll London, London, UK;6London Sch of Hygiene & Trop Med, London, UK
A decade after the scale up of ART in southern and eastern Africa, mortality rates among HIV-positive adults remain 3-6 times higher than in HIV-negative adults. Prioritising interventions for improving HIV care requires information about the care stages where most deaths arise, and thus where the greatest gains could be made. Immediate ART eligibility may fundamentally reshape the care cascade by removing the ‘pre-ART care’ stage where high dropout has been documented.
We reviewed empirical data and mathematical modelling estimates about mortality across stages of HIV care. Empirical estimates came from linked clinical and vital registration data from Western Cape, South Africa, and population cohorts in Uganda, Malawi, and South Africa. We used four mathematical models calibrated to HIV epidemics and care and treatment utilization in Rwanda, Kenya, Malawi, and South Africa. Models estimated the distribution of HIV deaths occurring at each stage of care and projected this over the next decade assuming continuation of current patterns of HIV care uptake and retention. Three models simulated the effects of changing guidelines to immediate ART initiation for all patients linked to care, assuming that retention would be similar to current levels.
Only 10–30% of HIV-related deaths are estimated to occur among patients who are continuously on ART for 6 months or more. At present, the majority of HIV deaths occur among patients who did not initiate ART (Figure). Patients disengaging from ART have a high mortality rate, and models show that this will account for an increasing and substantial share of HIV deaths (21–44% in 2025). Assuming continuation of current care patterns, models projected that between 9% and 22% of HIV deaths from 2016–2025 would occur among patients who had linked to care but never initiated ART. Immediate ART initiation could reduce HIV deaths by between 6–14% over 2016–2025, mostly due to removing the opportunities to disengage before treatment initiation.
Even in settings with high ART coverage, the majority of HIV-related deaths are likely to continue to be among patients who are not on ART, rather than patients who are stable on ART. Programmes should continue to prioritise interventions to link and retain patients to and on ART. Universal eligibility for ART initiation may bring substantial benefits through the simplification in the care cascade.