Network Characteristics Associated with HIV Testing Conversations Among Transgender Women Living in Los Angeles, California
Cherie S. Blair, Ian W. Holloway, Jesse B. Fletcher, Cathy J. RebackBackground: Transgender women (TW) are disproportionately impacted by HIV, with prevalence among TW estimated to be over 10 times higher than other populations in the U.S. Yet, according to the U.S. Transgender Health Survey, only 62% of TW reported receipt of HIV testing in 2015. Given substantial social marginalization experienced by this group, there have been limited culturally responsive public health information geared towards TW, resulting in suboptimal HIV testing and linkage to healthcare. Understanding social network structures of TW can be utilized to inform the development of novel HIV prevention and treatment paradigms. This study seeks to describe network factors associated with HIV testing conversations among TW and their social network alters.
Methods: From July 2015 to September 2016, 271 self-identified TW, who reported alcohol, drug use and/or condomless anal intercourse in the past 6 months, enrolled in the study. Respondents completed an audio computer assisted self-interview survey that evaluated sociodemographic characteristics, HIV risk/protective behaviors and egocentric network characteristics. Data collected on network alters included: relationship type (friend, relative, partner, trans “mother”, caseworker, or other), length of relationship, method/frequency of communication, member of a LGBQ community, perception as a “confidant”, HIV risk/protective behaviors, and substance use. Mixed-effects logistic regression evaluated discussion of HIV testing with network alter characteristics, controlling for respondent sociodemographics, HIV status, network density, and network size.
Results: Of the 271 enrolled participants, 264 nominated 2,529 alters. Participants reported discussing HIV testing with 35.5% (n=897) of alters. Median age of participants was 32 years (IQR 26-43). Median network size was 11 alters (range 2-60) and median density was 0.35 (IQR 0.16-0.56). In adjusted analysis, discussion of HIV testing was positively associated among alters who were trans “mothers” (aOR 2.17; 95% CI 1.10-4.26; p=0.025), compared to a friend, and if the relationship duration was >10 years (4.41; 2.02-9.62; p<0.001), compared to relationships <1 year. Discussion of HIV testing was positively associated with the following alter characteristics: perception as a confidant (aOR 2.91; 95% CI 1.78-4.78; p<0.001), member of a LGBQ community (2.13; 1.29-3.50; p=0.003), discussion of condom use (29.66; 16.73-52.60; p<0.001), knowledge about HIV pre-exposure prophylaxis (PrEP) (4.04; 2.05-7.94; p<0.001), receipt of HIV testing (22.58; 11.70-43.57; p<0.001), and discussion of drug use (1.90; 1.09-3.29; p=0.023). Stimulant use with the alter was negatively associated with discussion of HIV testing (aOR 0.31; 0.12-0.81; p=0.017), compared to alters with no stimulant use.
Conclusions: HIV testing conversations were associated with close network relationships and among alters who had knowledge of and/or engaged in HIV preventative behaviors themselves. Conversely, stimulant use, which has repeatedly been associated with increased HIV risk behaviors, was negatively associated HIV testing conversations. These results indicate the importance of leveraging close relationship networks among TW to increase HIV testing and the potential role for network-based HIV prevention strategies for groups of TW. Popular-opinion leader models have been used successfully with MSM to increase HIV preventative behaviors and disseminate HIV prevention information within their networks, and these findings suggest the potential utility of adapting such models for use among TW.