CROI 2017 Abstract e-Book

Abstract eBook

Poster and Themed Discussion Abstracts

Poster and Themed Discussion Abstracts

1038 THE POTENTIAL IMPACT AND COST-EFFECTIVENESS OF MULTIPURPOSE PREVENTION TECHNOLOGIES Ide Cremin 1 , Jennifer Smith 1 , Rebecca Geary 2 , Timothy B. Hallett 1 1 Imperial Coll London, London, UK, 3 London Sch of Hygiene & Trop Med, London, UK

Background: Multipurpose prevention technologies (MPTs) to prevent both HIV and unintended pregnancy could create important opportunities in terms of cost efficiencies, higher acceptability and a reduced adherence burden for users. One such product in development is an intra-vaginal ring (IVR), and analyses of potential impact and cost- effectiveness will help guide development and delivery. Methods: We developed a mathematical model to examine the impact and cost-effectiveness of four delivery strategies for an IVR MPT in South Africa. The first strategy assumes broad access to women of reproductive age (15-49 years) through existing public healthcare facilities. The second assumes enhanced coverage among younger women (15-29 years). The third and fourth prioritize an IVR for young women only and female sex workers, respectively. In accordance with acceptability data, coverage was assumed to vary between 5-10% of women in these target groups. Costs were assumed to range from US$16-163 per woman per year, depending on the assumed cost of the product and delivery strategy. We summarised HIV-related and reproductive health impacts as DALYs averted over a ten-year intervention period. Results: HIV infections averted among women range from 2.7% (low efficacy ring) to 5.5% (high efficacy ring), when an IVR is prioritized to younger women (15-29 years old). Maternal deaths averted range from 1.7% (low efficacy ring) to 5.2% (high efficacy ring), assuming 50% displacement from other methods of contraception, and from 5.2% (low efficacy ring) to 6.5% (high efficacy ring) assuming all IVR ring users are new contraceptive users, when an IVR is provided to 15-49 year old women with enhanced coverage among younger women. Assuming a high-efficacy low-cost IVR, the cost per DALY averted would range from $176 to $1089 (depending on delivery strategy) with 50% displacement, and from $169 to $1070 if all IVR users are new contraceptive users. Conclusion: The use of an MPT IVR has the potential to cost-effectively reduce new HIV infections and unintended pregnancies and their sequelae among women in South Africa, if low costs can be achieved through integration with existing services. New and forthcoming data on the efficacy and cost of delivery of dual use IVRs, and women’s and couples’ behaviours and preferences will be critical for optimising the use of dual use products for the prevention of HIV and unintended pregnancy. 1039 THE IMPORTANCE OF A LOCAL PERSPECTIVE IN MONITORING THE SUCCESS OF HIV PREVENTION Sarah-Jane Anderson 1 , Joanne Enstone 2 , Geoff P. Garnett 3 , Timothy B. Hallett 1 1 Imperial Coll London, London, UK, 2 Univ of Nottingham, Nottingham, UK, 3 Bill and Melinda Gates Fndn, Seattle, WA, USA Background: Setting and monitoring progress towards targets for HIV prevention should ensure responsive programmes and the appropriate use of resources. However, because of the considerable diversity in the burden and characteristics of the epidemic it is unclear how to apply global targets for HIV control to local settings. Furthermore, a large number of indicators have been recommended for monitoring progress but it is unclear which are most informative. Methods: We use location specific models, tailored to reflect available epidemiological data in each of the counties and major cities in Kenya, to project the future trajectory of each local epidemic. Because both the epidemic and response may change in many ways, multiple future trajectories were considered through varying future behaviours, treatment coverage and available prevention interventions. From these we are able to (1) look at the change in incidence in each location between 2015-2030 across simulations, and (2) assess which indicators are most indicative of a change in incidence. Results: The future trajectory of the epidemic varies considerably between modelled locations, thus the impact of specific targets will differ according to local setting (only 10 of 48 locations experience a median reduction in incidence of >80% by 2030 across the scenarios considered). The strength of each indicator in predicting changes in incidence is highly dependent on local patterns of transmission and the extent to which the epidemic is concentrated or generalised (Figure 1). Most single indicators demonstrate only limited association with changes in incidence as changes in incidence depend upon the combined effect of changes in sexual behaviours and other intervention successes. Measures of the population prevalence of viral non-suppression show the most consistent associations with long-term changes in incidence across locations. Conclusion: Targets and indicators need to be appropriate for the local epidemic and what can feasibly be achieved. If they are to be used to assess progress in controlling HIV they need to be informed by a good understanding of how they relate to HIV incidence. There is no one universally reliable indicator to predict future changes in HIV incidence, but population viral load may offer the most consistently predictive metric in some of the settings with large generalised epidemics. 1040 PAST AND FUTURE IMPACT OF ART SCALE-UP IN BRAZIL: PROGRAMME DATA ANALYSIS 1996–2014 Juan F. Vesga 1 , Ana R. Pati Pascom 2 , Tara Mangal 1 , Marcelo Freitas 2 , Juliana M. Givisiez 2 , Ronaldo de Almeida Coelho 2 , Mariana Veloso Meirelles 2 , Clarissa Habckost Dutra de Barros 2 , Adele Benzaken 2 , Timothy B. Hallett 1 1 Imperial Coll London, London, UK, 2 Ministry of Hlth, Brasilia (DF), Brazil Background: Universal free access to ART is available in Brazil since 1996 and from 2013 it is offered to all living with HIV disregarding of their CD4 cell count. We combined mathematical modelling and extensive data analysis to assess, for the first time in the country, the impact of ART expansions and the prospects of future mortality and incidence reductions. Methods: A network of models was developed to represent HIV transmission and the flow through stages of care in Brazil’s five major regions. Four datasets (SISCEL, SICLOM, SINAN, and SIM) encompass Brazil’s records on laboratories, treatment, surveillance, and mortality respectively with records dating as early as 1981 and updated up to 2014. These

CROI 2017 449