CROI 2017 Abstract e-Book

Abstract eBook

Poster and Themed Discussion Abstracts

model was calibrated to race-specific HIV incidence estimates for the Atlanta population in 2012. Then, in an “idealized” prison HIV care continuum scenario, all incarcerated agents were tested for HIV, initiated ART (for those testing positive), and were permanently maintained on ART post-release. Next, we compared projected community HIV incidence 10 years after implementation of the “idealized” correctional care continuum to that of the base case model. Results: Compared to the base case, the idealized care continuum scenario reduced 10-year community HIV incidence by 23.7% (from 48.8 to 37.3 per 100,000 by 2022). The percentage reduction was larger among blacks (25.4%, 97.5 to 72.8 per 100,000 by 2022) than among whites (16.7%, 17.2 to 14.3 per 100,000; Figure 1). The overall reduction in community HIV incidence was larger (30.7%) in a sensitivity analysis in which HIV risk was doubled for 6 months post release. Conclusion: Universal HIV testing and treatment in prisons, and programs to ensure retention in care post-release, may reduce HIV incidence in non-incarcerated populations. An optimized HIV care continuum in the prison setting and post-release may also significantly reduce racial HIV disparities, particularly if effective, culturally tailored linkage to care interventions can be implemented. 1045 THE COST-EFFECTIVENESS OF FINANCIAL INCENTIVES FOR VIRAL SUPPRESSION IN HPTN 065 Blythe J. Adamson 1 , Deborah J. Donnell 2 , Dobromir Dimitrov 2 , Louis Garrison 1 , Geetha Beauchamp 2 , Theresa Gamble 3 , Wafaa M. El-Sadr 4 1 Univ of Washington, Seattle, WA, USA, 2 Fred Hutchinson Cancer Rsr Cntr, Seattle, WA, USA, 3 FHI 360, Durham, NC, USA, 4 ICAP at Columbia Univ, New York, NY, USA Background: HIV viral suppression (VS) is associated with individual and societal health and economic benefits. The HPTN 065 study found providing $70 financial incentives (FI) for VS to HIV patients in Bronx, NY, and Washington, DC, resulted in a significantly higher proportion of patients with VS (revised as 3.7%) at sites randomized to FI compared to standard care. We developed a mathematical model to evaluate the cost-effectiveness of FI in HPTN 065. Methods: A two-part semi-Markov model was used to simulate the cohort of HIV patients in care at study sites and their sexual partners during the two-year intervention. The effect on VS was assumed to diminish to zero over six months when FIs end, with patients and partners followed over a lifetime horizon. The number of HIV transmissions during the study period was estimated with transmission risk equations. Study budgets and staff time informed cost parameters; self-reported sexual activity informed the number of partners; patient utility projection relied on literature-based utility by CD4 count. Lifetime total costs, HIV transmissions, and quality-adjusted life years (QALYs) are predicted from a health-care sector perspective and discounted 3% annually. We assumed a US willingness to pay of 3xGDP per capita threshold of $150,000/QALY for cost-effectiveness. Results: FIs for VS are likely to be highly cost-effective for HPTN 065 with an overall incremental cost-effectiveness ratio (ICER) of $7,371/QALY. Over two years, FI had a fixed cost of $167,714 per clinic for administration plus an average variable cost of $337 per patient for gift cards (Table 1). The resulting improvement in VS is projected to gain 19 patient QALYs and 20 partner QALYs in the population, and prevent 3 HIV infections for an average clinic with 456 patients on ART. Outpatient visits and ART costs increased 8.7% for FI patients; however, the estimated marginal increase in health care costs ($119 and $3,089 per patient respectively) are offset by savings from fewer HIV transmissions. A sensitivity analysis projected FI for VS cost-effective in all pre-specified sub-groups based on clinic size, type, baseline VS, and city with ICERs ranging from cost-saving to $79,471/QALY. Conclusion: Financial incentives offer substantial value for money to improve the length and quality of life for HIV patients and their partners. This analysis provides evidence supporting the likely cost-effectiveness of an intervention to strengthen the clinical care continuum and reduce HIV transmission.

Poster and Themed Discussion Abstracts

1046 WITHDRAWN

1047 POPULATION-LEVEL IMPACT OF REACHING UNAIDS HIV PREVENTION TARGETS IN CÔTE D’IVOIRE Mathieu Maheu-Giroux 1 , Juan F. Vesga 2 , Souleymane Dabate 3 , Michel Alary 3 , Stefan Baral 4 , Douada Diouf 5 , Kouame Abo 6 , Marie-Claude Boily 2 1 Imperial Coll London, London, UK 2 Laval Univ, Quebec, Canada, 3 The Johns Hopkins Univ, Baltimore, MD, USA, 5 Enda Santé, Dakar, Senegal, 6 Ministère de la Santé et de l’Hygiène Pub, Abidjan, Cote d’Ivoire Background: National responses need to be markedly accelerated to achieve UNAIDS’ ambitious targets. We aimed to estimate the impact of various scale-up of antiretroviral treatment as prevention scenarios in Côte d’Ivoire. Methods: An age-stratified dynamic model was developed and calibrated to epidemiological and programmatic data using a Bayesian framework, following a comprehensive review of the published and grey literature and site visits. The model represents sexual and vertical HIV transmission in the general population, female sex workers (FSW), and men who have sex with men (MSM). We estimated the impact of scaling-up interventions to reach the UNAIDS targets (90-90-90 by 2020 and 95-95-95 by 2030), and that of eight other scenarios, on HIV transmission in adults and children, compared to our baseline scenario that maintains 2015 rates of testing, antiretroviral therapy (ART) initiation, retention, treatment failure, and levels of condom use. Results: In 2015, we estimated that 51% (95% Credible Intervals: 45-58%) of HIV positive individuals were aware of their status, 75% (59-85%) of those aware were on ART, and 77% (74-79%) of those on ART were virologically suppressed. Reaching the 2020 and 2030 UNAIDS targets on time would avert 49% (41-57%) of new adult HIV infections over 2015-2030 compared to 30% (25-36%) if the 90-90-90 target is only reached in 2025. Attaining the UNAIDS targets in FSW, their clients, and MSM - but not in the rest of the population - would avert a similar fraction of new infections (30%; 23-39%). Only 37% (23-49%) of all HIV infections would be averted if scaling-up of ART was accompanied by a 25% points drop in condom use from their 2015 levels among FSW and MSM. Conclusion: Rapid scale-up of interventions, particularly HIV testing, ART initiation, and retention in the next five years is needed to halve HIV incidence by 2030. Reaching UNAIDS 90-90-90 targets with a five year delay reduces impact by two fifth (40%) and is not more effective than reaching UNAIDS targets in time among vulnerable key populations (MSM,

CROI 2017 452