CROI 2016 Abstract eBook

Abstract Listing

Poster Abstracts

897 Longitudinal Analysis of Sexual Networks of US Black MSM in HPTN 061 Hong Van N. Tieu 1 ; Beryl A. Koblin 1 ;Vijay Nandi 1 ; Pich Seekaew 2 ; Sophia Hussen 3 ; DeMarc A. Hickson 4 ; Kenneth H. Mayer 5 ; Hyman Scott 6 ;Vanessa Cummings 7 ; Carl Latkin 8 ; for the HPTN 061 StudyTeam 1 New York Blood Cntr, New York, NY, USA; 2 Columbia Univ Mailman Sch of PH, New York, NY, USA; 3 Emory Univ, Atlanta, GA, USA; 4 Univ of Mississippi Med Cntr, Jackson, MS, USA; 5 The Fenway Inst, Fenway Hlth, Boston, MA, USA; 6 Univ of California San Francisco, San Francisco, CA, USA; 7 Johns Hopkins Univ Sch of Med, Baltimore, MD, USA; 8 Johns Hopkins Univ, Baltimore, MD, USA Background: Black men who have sex with men (BMSM) have been disproportionately affected by HIV in the U.S., which is not explained by differences in individual behavioral risks. Sexual networks of BMSMmay place them at increased HIV risk compared to non-BMSM. Sexual networks are rarely static, and longitudinal data are lacking on how changes in BMSM sexual networks over time influence HIV and sexual risk behaviors. Methods: In this longitudinal investigation of BMSM sexual networks, we utilized self-reported egocentric network data collected over 12 months from BMSM in HPTN 061, a feasibility study of a multi-component HIV prevention intervention in 6 U.S. cities. At months 0 (M0), 6 (M6), and 12 (M12), BMSM completed a network inventory by describing their social and sexual network members from the prior 6 months. Sexual network composition, size, density (extent to which members had sex with one another), and serodiscordant/serostatus unknown condomless sex fromM6 and M12 at the participant and partner levels were compared to M0 data using paired t-tests or McNemar’s test. Results: Of 1,553 men at M0, 348 (22%) were HIV-positive by HIV testing. At M0, mean sexual network size was 3.3 (SD 2.2) and mean sexual network density was 4.6% (SD 16.4%). 56% reported having exclusively Black partners in the last 6 months, 86% had overlap of social and sexual networks, 22% reported having an HIV-positive male partner, and 48% reported serodiscordant/serostatus unknown condomless sex. Using partner-level M0 data, 27% of partners had ≥2 age category difference between partner and participant. Compared with M0, the following significantly decreased over time: mean sexual network size (2.4 at M6, 2.0 at M12), mean sexual network density (2.5% at M12 only), overlap of social and sexual networks (76% at M6, 70% at M12), and condomless sex (26% at M6, 23% at M12) (all p-values<0.001). No significant changes in having exclusively Black sex partners, having HIV-positive male partners, and age difference between partners and participants were noted over time. Conclusions: The sexual networks of BMSM in HPTN 061 were dynamic over 12 months. While decreases in sexual network size and density may reflect decreased HIV transmission and acquisition risk, decrease in overlap between social and sexual networks over time, with decreased social support, is concerning. Future studies should examine how BMSM sexual networks change over time and how these changes influence HIV infection and sexual risk behaviors. 898 Does Size Really Matter? Sensitivity Analysis of Number of Seeds in an RDS Study Nathan J. Lachowsky 1 ; Henry F. Raymond 2 ; Zishan Cui 3 ; Paul Sereda 3 ; Ashleigh Rich 3 ; Eric A. Roth 4 ; Robert S. Hogg 5 ; David M. Moore 1 ; for the Momentum Health Study 1 Univ of British Columbia, Vancouver, BC, Canada; 2 Univ of California San Francisco, San Francisco, CA, USA; 3 BC Cntr for Excellence in HIV/AIDS, Vancouver, BC, Canada; 4 Univ of Victoria, Victoria, BC, Canada; 5 Simon Fraser Univ, Burnaby, BC, Canada Background: Respondent-driven sampling (RDS) is an increasingly used peer chain-recruitment method to sample “hard-to-reach” populations for whom there are no reliable sampling frames. Implementation success of RDS varies; one potential negative factor being the number of seeds used. We conducted a sensitivity analysis on estimates produced using various RDS-weighting procedures and sample cuts based on data from a study of men who have sex with men in Vancouver, Canada. Methods: Participants were recruited from 2012-2014 using RDS, and had to be aged ≥16 years of age and report recent sex with another man. The study protocol included a self-completed questionnaire on demographics, sexual behavior and substance use. To conduct this analysis, we used various sample cuts, starting with all participants and subsequently removing unproductive seeds (0 recruitment waves), chains of ≤1 recruitment waves, and chains of ≤2 recruitment waves. Using the RDS Analyst 0.52 and SAS 9.4, we calculated estimates for three different outcomes for each sample cut and three different RDS weighting procedures: RDS-I (Salganik-Heckathorn), RDS-II (Volz-Heckathorn), and RDS-SS (Giles). We also assessed seed dependence with bottleneck analyses and convergence plots. Results: Overall, 719 participants were recruited, which included 119 seeds and a maximum of 16 recruitment waves (mean recruitment chain length = 1.75). The sample of >0 recruitment waves removed all unproductive seeds (n=50/119, 42.0%), resulting in 69 chains (mean length = 3.0). The sample of >1 recruitment waves removed 125 seeds or recruits (17.4% of overall sample), resulting in 37 chains (mean length = 4.8). The final sample of >2 recruitment waves removed a further 182 seeds or recruits (25.3% of overall sample), resulting in 25 chains (mean length = 6.1). Based on these various samples, Table 1 provides estimates for three key study outcomes: HIV serostatus, high risk sex (condomless anal intercourse with HIV discordant/unknown status partner), and injecting drugs. Convergence plots and bottleneck analyses were satisfactory. Conclusions: For each outcome and within each sample cut, the crude proportions fell within 95% confidence intervals of all RDS-weighted estimates. All RDS-weighted estimates were similar and fell within the 95% confidence intervals of each other. Although potentially costly and time consuming, our results indicate that RDS studies are not negatively affected by large numbers of unproductive or lowly productive seeds.

Poster Abstracts

899 Respondent-Driven Sampling: An Epidemiological Tool With Interventional Potential Sunil S. Solomon 1 ; Aylur K. Srikrishnan 2 ; Santhanam Anand 2 ; Allison M. McFall 1 ; Pachamuthu Balakrishnan 3 ; David Celentano 1 ; Suresh Kumar 2 ; Gregory M. Lucas 1 ; Shruti H. Mehta 1 1 Johns Hopkins Univ, Baltimore, MD, USA; 2 YRG Cntr for AIDS Rsr and Educ, Taramani, India; 3 YRG Cntr for AIDS Rsr and Educ, Chennai, India Background: To achieve the UNAIDS 90-90-90 target, key populations, such as men who have sex with men (MSM) and people who inject drugs (PWID) in low- and middle- income countries, must be reached. Respondent-driven sampling (RDS), a chain-referral strategy, is an established epidemiologic tool to characterize epidemics in hidden populations. We explore the ability of RDS to identify unaware HIV-positive persons over successive recruitment waves and its ability to reach people in all regions of medium to large cities in India. Methods: We conducted RDS surveys in 27 Indian cities (12 MSM, 15 PWID). Individuals were eligible if they were age ≥ 18 years and reported recent MSM behavior (MSM sites) or drug injection (PWID sites). We initiated each RDS with 2-3 “seeds” per site, gave participants two coupons to recruit network members, and continued recruitment until 1000

377

CROI 2016