CROI 2018 Abstract eBook

Abstract eBook

Oral Abstracts

while the strategies with the lowest yield are expected to include home-based and school-based testing (Figure). Considering only testing costs associated with the potential new strategies, the incremental cost per life year saved is lowest for MSM testing ($182) and highest for self-testing scenarios ($1935 for partners of pregnant women and $1108 for home-based testing). Incremental costs dropped substantially when considering the impact on the cost of the entire HIV programme, rendering a number of strategies marginally cost saving over 20 years. In the absence of any change to current strategies, levels of HIV diagnosis in adults are expected to increase from 84.8% in 2016 to 93.3% in 2025. Biennial home-based testing including a self-testing kit offer would have the greatest impact, increasing this fraction to 96.7% in 2025, while home-based testing without the self-test kit offer would increase the fraction to 96.2%. Conclusion: There is a trade-off between achieving substantial population- level increases in diagnosis rates and pursuing the most cost-effective HIV testing modalities. Community-based testing and self-testing are likely to be important in reaching diagnosis targets but may be relatively inefficient.

per night in the direct peer delivery arm (mean difference -0.78 clients, 95% CI -1.28 to -0.28, P=0.002) and facility-based coupon arm (-0.71, 95% CI -1.21 to -0.21, P=0.005) compared to standard-of-care. Similarly, they reported fewer non-client partners in the direct peer delivery arm (-3.19, 95% CI -5.18 to -1.21, P=0.002) and in the facility-based coupon arm (-1.84, 95% CI -3.81 to 0.14, P=0.07) compared to standard-of-care. Conclusion: In contrast to previous concerns, expansion of HIV self-testing may have positive spillover effects on HIV prevention efforts among FSW in Zambia. 149 HIV SELF-TEST DISTRIBUTION INCREASES TEST FREQUENCY IN SOUTH AFRICAN MSM Sheri A. Lippman 1 , Tim Lane 1 , Oscar Rabede 2 , Hailey Gilmore 1 , Yea-Hung Chen 3 , Nkuli Mlotshwa 2 , Kabelo Maleke 2 , Alexander Marr 1 , James A. McIntyre 2 1 University of California San Francisco, San Francisco, CA, USA, 2 Anova Health Institute, Johannesburg, South Africa, 3 San Francisco Department of Public Health, San Francisco, CA, USA Background: South African men who have sex with men (MSM) have a high burden of undiagnosed HIV infection and HIV-testing rates incommensurate with their risk. HIV self-testing (HIVST) may increase testing uptake, frequency, and earlier HIV detection and treatment. We implemented a longitudinal HIVST study among South African MSM in Mpumalanga Province, in order to explore acceptability, feasibility, utilization and distribution patterns, and to understand how HIVST might expand testing frequency in this high prevalence area. Methods: We recruited 127 HIV-negative MSM between June 2015 and May 2017 in Gert Sibande and Ehlanzeni districts. Participants received 5 self-test kits of their choice - oral fluid or blood fingerstick – at baseline and an additional 4 kits at a three month visit. Participants were asked to use the kits themselves at least one time and to distribute the other kits to their networks. Surveys were conducted at baseline, three months, and six months post-enrollment to elicit information on HIVST experiences, preferences, acceptability, utilization of HIVST and clinic-based testing, and test distribution to others. We used generalized estimating equations to assess changes in regular (every six months or more frequent) testing. Results: Ninety-one percent of all participants self-tested, all of whom reported being likely to self-test again, with over 80% preferring HIVST to clinic-based testing. Fingerstick tests were preferred: 45% ever choose oral fluid tests and 80% ever choose blood. Returning participants distributed 728 tests to sexual partners (18.5% of kits), friends (51.6%), and family (29.8%). Among those testing, 32% of the cohort reported testing with someone else present and 24% reported concurrent testing (testing at the same time as another). Six participants (5% of those returning for follow-up) seroconverted during the study; 40 new diagnoses were reported among network test recipients. Regular testing increased from 37.8% prior to the study to 84.5% at follow-up (p<0.1), and participants reported anticipated regular testing of 100% if HIVST were available compared to 84% if only clinic-testing were available in the coming year (p<.01). (Figure 1) Conclusion: HIVST is highly acceptable and feasible to distribute through MSM networks in South Africa. Newly quarterly testing guidelines are unlikely feasible in a clinic-based environment alone, however our data suggest that HIVST is key to meeting regular testing goals and improving early detection.

Oral Abstracts

148 EFFECT OF HIV SELF-TESTING ON SEXUAL PARTNER NUMBERS FOR ZAMBIAN FEMALE SEX WORKERS Catherine E. Oldenburg 1 , Michael M. Chanda 2 , Katrina F. Ortblad 3 , Magdalene Mwale 2 , Steven Chongo 2 , Nyambe Kamungoma 2 , Catherine Kanchele 2 , Andrew Fullem 4 , Guy Harling 5 , Till Bärnighausen 6 1 University of California San Francisco, San Francisco, CA, USA, 2 John Snow, Inc, Lusaka, Zambia, 3 Harvard University, Cambridge, MA, USA, 4 John Snow, Inc, Boston, MA, USA, 5 University College London, London, UK, 6 Heidelberg University, Heidelberg, Germany Background: HIV self-testing is a promising HIV testing technology that may reduce some traditional barriers to HIV testing among female sex workers (FSW). However there are concerns that self-testing may lead to behavioral compensation amongst those who test negative. Here, we assess the effect HIV self-tests distribution modalities on the number of FSW client and non-client sexual partners in a randomized controlled trial of HIV self-testing among FSW in Zambia. Methods: Peer educators recruited participants via their social networks. Peer educator-participant groups were randomized in a 1:1:1 fashion to one of three arms: 1) delivery of HIV self-tests directly from a peer educator, 2) free facility-based delivery of HV-self tests in exchange for coupons, or 3) referral to standard HIV testing (standard of care). All participants also completed four peer educator intervention sessions, including condom distribution. Quantitative assessments were completed at baseline, one, and four months. Participants were asked their average number of client partners per night at baseline, one and four months, and their number of non-client partners in the past 12 months at baseline and in the past month at one and four months. The mean change in number of client and non-client partners was calculated separately using a mixed effects generalized linear model, with fixed effects for study arm, study site, and baseline average number of client or non-client partners and a random effect for peer educator group. Results: From September-October 2016, 965 women were enrolled and randomized in 160 peer educator groups. Participants were a median of 25 years of age (interquartile range 21 to 30). The majority of participants (89.3% at one month and 79.6% at four months) reported testing for HIV during the study period. At four months, participants reported significantly fewer clients

54

CROI 2018