CROI 2017 Abstract e-Book

Abstract eBook

Poster and Themed Discussion Abstracts

1030 COST-EFFECTIVENESS OF DIFFERENT DELIVERY APPROACHES FOR HIV SELF-TESTING Valentina Cambiano 1 , Cheryl Johnson 2 , Carmen Figueroa 2 , Paul Revill 3 , Rachel Baggaley 2 , Elizabeth L. Corbett 4 , Tsitsi Apollo 5 , Karin Hatzold 6 , Frances Cowan 7 , Andrew Phillips 1 1 Univ Coll London, London, UK, 2 WHO, Geneva, Switzerland, 3 Univ of York, York, UK, 4 London Sch of Hygiene & Trop Med, London, UK, 5 Zimbabwe Ministry of Hlth and Child Care, Harare, Zimbabwe, 6 Pop Services Intl, Harare, Zimbabwe, 7 Cntr for Sexual Hlth and HIV/AIDS Rsr, Harare, Zimbabwe Background: HIV self-testing (HIVST) has been shown to be highly acceptable and able to reach people at high risk who might not otherwise test. This study assesses the cost- effectiveness of introducing HIVST in Zimbabwe for specific populations, considering different delivery models. Methods: A dynamic model (HIV Synthesis) is used. The base case cost per HIVST kit is $4.8. Alternative strategies are compared to a reference scenario (RS) of no HIVST: a) secondary distribution of HIVST to partners of pregnant women (uptake in the population eligible in each year 40%; fully loaded cost/test $5.0); b) pharmacy-based distribution (PBD) of HIVST to young people (15-24), female sex workers (FSW) and adult men (25-49) (5%; $6.0); c) community-based distribution (CBD) to young people (65%; $7.2 for all CBD, as it is based on data including supervision), d) FSW (42%), e) adult men (55%). The incremental net health benefit (difference between health gains and health opportunity costs, calculated as costs divided by the cost-effectiveness threshold [CET]) of each strategy is compared to the RS. Alternative CETs are used: $1,000, $500, $150. A health care payer perspective is taken using 20 year time horizon. Results: In the context of Zimbabwe, where we projected 85% of people with HIV know their status in 2016, the introduction of HIVST is likely to be cost-effective (CET of $500- 1,000) when considering secondary distribution, PBD and CBD for FSW. Reductions in the cost of HIVST kit, which are believed to be possible, improve the cost-effectiveness of HIVST. However, higher cost of HIVST and lower linkage to care for people whose diagnosis is a consequence of a reactive HIVST result could lead to situations in which HIVST is not cost-effective. Conclusion: In settings with high levels of HIV status awareness, interventions involving additional HIV tests (at the current cost) are unlikely to be cost-effective; our analysis suggests that HIVST strategies most likely to be cost-effective are secondary distribution, PBD and CBD for FSW. The most cost effective strategy is likely to involve a combination of distribution approaches and this will be evaluated as we move forward. In settings with lower testing coverage or if individuals found to be HIV-negative through HIVST were to link to HIV prevention (e.g. pre-exposure prophylaxis and voluntary medical male circumcision), it is likely that even other forms of HIVST distribution could become cost-effective, similarly if the CBD had to be performed less frequently.

Poster and Themed Discussion Abstracts

1031 THE COST OF NOT RETESTING: HIV MISDIAGNOSIS IN THE ART “TEST-AND-OFFER” ERA Jeffrey Eaton 1 , Cheryl Johnson 2 , Simon Gregson 1 1 Imperial Coll London, London, UK, 2 WHO, Geneva, Switzerland

Background: WHO recommends re-testing before starting ART, but this practice is not widely implemented. New recommendations for immediate ART initiation of all HIV- positive adults increase the risk that misdiagnosed HIV-negative persons will be initiated on lifelong ART. We compared the anticipated costs of re-testing persons before ART initiation to the expected cost of providing ART to misclassified HIV-negative persons. Methods: We created a model to estimate the cost and outcomes of testing 10,000 persons using WHO-recommended serial HIV testing strategies in settings with 1% prevalence (3-test strategy) and 10% prevalence (2-test). The model calculated the expected number of misclassified HIV-negative persons initiated on ART assuming 98% test specificity (equating to 99.6% specificity for the two-test strategy and 99.9% for the three-test strategy), consistent with the real-world testing algorithm performance from a multi- country CDC study. Costs were ‘fully-loaded’ (including commodities, personnel, supply chain, and management) typical of LMIC settings. The first test and associated counseling cost US$8; each confirmatory test cost $6. ART provision to misclassified persons cost $450 per annum, for a total discounted (6%) lifetime cost of $6300 assuming a 30-year life expectancy after ART initiation. In the re-testing scenario, re-testing was assumed to occur just before ART initiation by an independent healthcare worker. Results: In the 1% HIV prevalence setting, testing 10,000 cost $83,000, and 9 HIV-negative people would be misdiagnosed and initiated on ART; costing $58,000 in unnecessary ART costs. Re-testing all diagnosed HIV-positives cost $2,000, providing a net saving of $56,000. For 10% prevalence, testing 10,000 persons cost $87,000. 39 HIV-negative people are misdiagnosed, costing $243,000 in unnecessary ART. Re-testing cost $14,000, providing net savings of $226,000. Savings from averted ART costs were greater than expenditure on re-testing within 0.5 years and 0.8 years for the 1% and 10% prevalence scenarios, respectively. That averted ART costs quickly overtake re-testing costs was robust to varying test specificity from 92% to 99%, suggesting re-testing will be cost-saving even as HIV testing performance improves.

CROI 2017 445