|Home | About | Journals | Submit | Contact Us | Français|
E. C. Wilson originated the study, managed its implementation, ran the analysis, and was the primary author of the article. R. Garofalo was the site principal investigator in Chicago and assisted with the study development, implementation, and writing. D. R. Harris primarily ran the statistical analyses and assisted in writing. M. Belzer assisted in the study development and implementation and in developing the analysis and writing the article. The Adolescent Medicine Trials Network for HIV/AIDS Interventions facilitated funding and coordination for the entire study. All authors helped to conceptualize ideas, interpret findings, and review drafts of the article.
We examined associations between partner types (categorized as main, casual, or commercial) and sexual risk behaviors of sexually active male-to-female (transgender female) youths.
We interviewed 120 transgender female youths aged 15 to 24 years recruited from clinics, community-based agencies, club and bar venues, referrals, and the streets of Los Angeles, California, and Chicago, Illinois.
Sexual risk behaviors varied by partner type. Transgender female youths were less likely to use condoms during receptive anal intercourse with their main partner and were less likely to use condoms with a main partner while under the influence of substances. Youth participants were also more likely to talk to a main partner about their HIV status. Our data identified no demographic or social factors that predicted condom use during receptive anal intercourse by partner type.
Research and interventions that focus on understanding and mitigating risk behaviors by partner type, especially those that tackle the unique risks incurred with main partners, may make important contributions to risk reduction among transgender female youths.
Transgender female youths (young persons born anatomically male who identify as female) are a group at high risk for HIV infection. A growing body of research shows estimates of HIV seroprevalence among transgender women that are higher than among other US at-risk populations.1-4 A meta-analysis of the HIV prevention literature focused on transgender women found that 27.7% of transgender women tested positive for HIV, while 11.8% reported being HIV positive.5 In the 2 studies to date, rates of HIV infection among transgender female youths were found to be comparable to those of transgender women.6,7
Some studies have shown that partner type (categorized as main, casual, or commercial) influences the sexual risk behaviors of groups at high risk for HIV infection.8-10 Partnership characteristics have been found to be important predictors of unprotected intercourse.11,12 Research with transgender women and youths has primarily focused on risks associated with commercial sex partners because of the high prevalence of sex work within this community, estimated to range from 24% to 75%.5,6 However, research points to a need to understand the behaviors of transgender female youths with their main partners as well. Studies have shown that men who have sex with men are more likely to have unprotected intercourse with main partners than with casual partners,9,10 and this may also be true for transgender female youths. One study of partner differences among transgender women of color in San Francisco found that significantly more transgender women had recently engaged in unprotected receptive anal intercourse with main partners than with casual or commercial sex partners.13 This finding is supported by other descriptive studies that found that transgender women reported being less likely to use condoms with their main partners than with their commercial partners.2,14
The need for intimacy with a main partner has been reported by transgender and other women as a reason for lower condom use with main partners.14,15 Research with heterosexual Black and Hispanic women has found that low risk perception within monogamous relationships, misinformation regarding HIV transmission and risky partners, stigma about condom use, and lower self-efficacy for condom use are common reasons for unsafe sex with steady partners.16 The most recent hypotheses about why high-risk individuals may be having more unprotected sex in steady relationships involve serosorting. Serosorting has been defined as “ the practice of preferentially having sex with partners of concordant HIV status or of selectively using condoms with HIV-discordant partners.”12(p212) Increases in serosorting have been observed among men who have sex with men, but no data exist on whether this phenomenon occurs among transgender female youths.17
Despite the growing literature on transgender women and other high-risk groups, data are needed concerning the influence of partner types on the sexual risk behaviors of transgender female youths. We examined associations between partner types and sexual risk behaviors among sexually active transgender female youths from Chicago, Illinois, and Los Angeles, California. Most participants in our study were ethnic minority transgender female youths who identified as something other than White. We hypothesized that transgender female youths would engage in riskier sexual behaviors with main partners than with casual and commercial partners because studies of other high-risk populations suggest that main partnerships may carry the highest risk for HIV transmission.
From January 2005 to March 2006 we conducted the Transgender Research Youth Project, a cross-sectional study of HIV risk behaviors among 151 transgender female youths aged 15 to 24 years in Chicago and Los Angeles. We analyzed data from 120 of these participants. We excluded 31 youths from our analysis because they did not provide information concerning sexual partners in the past 3 months (n=7) or reported that they did not have any sexual partners (n=24). All respondents were recruited from health clinics, community-based organizations, club and bar venues, the street, or word-of-mouth referrals. Eligible youths identified as a gender different from their male anatomical sex at birth, lived in the metropolitan area of 1 of the study sites, and spoke either English or Spanish. Written consent was obtained from all youths aged 18 years or older and written assent was given by younger participants (in accordance with a review board waiver of parental consent). All participation was voluntary and confidential.
All respondents were screened to determine eligibility for enrollment into the study. The screening instrument gathered demographic information such as age and race/ethnicity. The demographic section of the main questionnaire collected information not already elicited in the screening instrument, including estimated income, education status and attainment, and history of homelessness, foster care, and incarceration.
We asked participants about sexual behaviors in the past 3 months with main, casual, and commercial partners. We defined main partners as persons with whom participants had a regular or steady relationship at the time of the survey, such as boyfriends. Casual partners were considered to be the last partner with whom youths had a sexual relationship in the last 3 months, but who were neither steady partners nor paying partners. Commercial partners were considered to be anyone with whom respondents had intercourse in the past 3 months in return for money, shelter, food, or drugs.
For each partner type, we asked a standardized set of sexual risk behavior questions about whether the participants engaged in different sexual behaviors (e.g., insertive anal intercourse, receptive anal intercourse) in the past 3 months. A 5-point Likert scale, from always to never, assessed frequency of condom use. For most recent casual and current main partner, we also asked participants whether they had intercourse with their partner while under the influence of drugs or alcohol (responses were yes, no, don't know, or refuse to answer). For commercial partners, we asked participants to rate on the 5-point Likert scale how often they had intercourse with these partners while under the influence. Behaviors with commercial partners were dichotomized to never or ever having intercourse under the influence. For all partner types, if youths indicated they had intercourse while under the influence, we asked whether they were less likely to use a condom under those conditions.
We asked whether respondents ever talked to their main and casual partners about the partner's HIV status. We did not ask whether they discussed their commercial partners' HIV status, because our advisory committee suggested that it was too risky to ask about a client's HIV status. Because we were primarily concerned with decisions youths made about risks they were willing to take—not risks their partners were willing to take—we did not ask whether partners asked about the HIV status of our participants.
We used univariate statistics (number, percentage) to describe the distribution of demographic characteristics of participants overall and to describe HIV risk behaviors by types of sexual partners reported in the past 3 months. To assess whether the frequency of a particular sexual risk behavior differed significantly by partner type, we fit logistic regression models to the data with a generalized linear model capable of dealing with correlated data arising when participants provided multiple outcome responses (i.e., risk behavior responses associated with different sexual partner types).
We used logistic regression modeling to explore the effect of possible mediating factors for always using condoms during receptive anal intercourse with a casual partner versus a main partner and a commercial partner versus a main partner; our primary hypothesis was that main partnerships would be the most risky.5,13 Previous research on partner types among transgender women and young men who have sex with men identified associations between HIV risk behavior and age, ethnicity, education, monthly income, and participant's reported HIV status.13,18 We therefore adjusted our model for these characteristics. We defined statistical significance as P≤.05.
Of the 151 transgender female youths surveyed, 120 (80%) provided information on at least 1 of the 3 different partner types and were included in the statistical analysis. Respondents' median age was 21 years; 25% were aged 15 to 18 years; 30% were aged 19 to 21 years; and 45% were aged 22 to 24 years. The sample primarily consisted of ethnic minority youths. Forty percent identified as Hispanic, Latino/Latina and 40% as African American or Black (Table 1).
More than 40% of respondents reported incomes of less than $500 in the past 30 days. Most youths reported that they were not currently in school (69%); 28% reported currently being in school or a general equivalency diploma program. Approximately 60% of youths reported graduating from high school, having some college or technical training, or graduating from college or a technical school. Forty-seven percent of participants reported ever being homeless for 1 night or more; the median number of nights of homelessness was 30. Twenty-six percent reported ever having been a ward of the court or state (i.e., in the foster care system), and 61% reported ever being in the correctional system.
Table 2 presents the distribution of HIV risk behaviors according to sexual partner type. Fifty-eight youths currently had a main sexual partner, 67 had intercourse with a casual partner during the past 3 months, and 65 had intercourse with a commercial partner during the past 3 months.
Forty-five percent of respondents who reported having a main partner had insertive anal intercourse with this partner; 39% of respondents had insertive anal intercourse with a casual partner and 49% with a commercial partner. Insertive anal intercourse was not significantly associated with partner type (P>.4). Participants were more likely to always use a condom while having insertive anal intercourse with a commercial partner than they were with a main partner (81% versus 50%; P=.01). The odds of always using a condom during insertive anal intercourse with a commercial partner were more than 4 times as high as were the odds of always using a condom during insertive anal intercourse with a main partner (odds ratio [OR]=4.3; 95% confidence interval [CI]=1.4, 13.3; P = .01). Condom use did not differ significantly during insertive anal intercourse with casual versus main partners.
Seventy-two percent of respondents with main partner had receptive anal intercourse with this partner; 78% of respondents reported having receptive anal intercourse with a casual partner and 72% with a commercial partner. Receptive anal intercourse was not significantly associated with partner type (P > .5), but condom use during receptive anal intercourse was (P < .02). Condom use was significantly more likely to occur during receptive anal intercourse with casual and commercial partners than it was with a main partner (P = .01); the odds of always using a condom during receptive anal intercourse with a casual partner were nearly 3 times as high as were the odds of always using a condom while having receptive anal intercourse with a main partner (OR = 2.8; 95% CI =1.2, 6.3). The odds of always using a condom during receptive anal intercourse with a commercial partner were more than 3 times as high as were the odds of always taking this precaution with a main partner (OR = 3.1; 95% CI =1.3, 7.4; P =.01).
The adjusted odds ratios for always using a condom during receptive anal intercourse were somewhat higher than were the unadjusted odds ratios (casual versus main partners, OR = 3.3; 95% CI =1.3, 8.3; P =.01; commercial versus main partners, OR = 4.6; 95% CI =1.6, 13.8; P < .01), although the un-adjusted values were well within the confidence intervals reported in the adjusted model. The covariates (age, ethnicity, education, monthly income, and participant's HIV status) did not have a significant independent association with consistent condom use during receptive anal intercourse. We also fit an adjusted model to the data with race (Black versus other) substituted for ethnicity (Hispanic versus other), with the results differing only modestly.
Youth participants were significantly less likely to talk to a casual partner about that person's HIV status than they were to talk with their main partner (OR = 0.2; 95% CI = 0.1, 0.4; P <.01).
Having intercourse while under the influence of drugs or alcohol did not differ according to partner type (P > .1). Participants were more likely to use condoms with commercial partners than they were with main partners while under the influence (OR = 0.3; 95% CI = 0.1, 0.9); this difference was statistically significant (P = .03).
Our findings, from the only large study of transgender female youths to date, corroborate well-established findings from numerous studies of adult transgender women that show their relationships with main partners to be the riskiest. For example, Nemoto et al.13 found that transgender women of color were significantly more likely to have recently engaged in unprotected receptive anal intercourse with primary partners than they were with casual or commercial sex partners.
Our data may also help explain why unsafe sex with main partners was more prevalent than with other partner types. One important correlate was substance use. Transgender female youths in our study were significantly less likely to use condoms with main partners than they were with commercial partners while under the influence of drugs or alcohol. Substance use during intercourse may have reduced concerns about safe sex in the face of other, more important desires and concerns. Youths may have feared losing intimacy with their main partner, or perhaps they perceived negative attitudes toward condoms from their main partners, as has been found among other groups of adolescents.19 Prevention messages and interventions may need to focus on the unique barriers to safe sex for transgender female youths with main partners, especially those youths who use substances.
It is also possible that transgender female youths made informed decisions to have unprotected sex with main partners after discussing their partners' HIV status. Our respondents were significantly more likely to talk to a main partner than to a casual partner about that partner's HIV status. Therefore, it appears that youths were more interested in knowing their main than their casual partners' HIV status. The parallel between lower condom use with main partners and higher likelihood of discussing main partners' HIV status may be evidence that our respondents were serosorting.
Our findings support the idea that participants may have been more interested in knowing their main partners' HIV status than knowing the status of casual partners because they intended to have unprotected sex with main partners of the same serostatus. Participants who were HIV negative may have made informed decisions to have unsafe sex with their main partners because they determined their own HIV-infection risk to be low after their main partners disclosed an HIV-negative status. Similarly, HIV-positive youths may have determined that there was no risk of infecting main partners who reported the same serostatus. Our findings indicate that interventions to reduce HIV transmission among transgender female youths should address risk reduction behaviors these youths are already implementing via partner communication and serosorting.
Nemoto et al.'s study of risk behaviors by partner type found that demographic and social factors such as ethnic minority status, participant's HIV status, low income, and intercourse while under the influence were associated with unprotected receptive anal intercourse with different partner types.13 Our findings differ somewhat from those of Nemoto et al. in that these demographic and social factors did not contribute significantly to predicting condom use during receptive anal intercourse. This difference may stem from our specific cross section of the transgender female youth population: our respondents may have had similar social networks, neighborhoods, or lifestyles because of how they were recruited, making it difficult to detect significant differences in demographic and social factors that might be apparent in a more diverse sample.20 Our results more closely resemble those from research with people living with HIV/ AIDS showing that partnership covariates, not demographic or social factors, were significantly associated with unprotected sex.11
The primary limitation of our study was that the data were drawn from a specific cross section representing only 1 segment of a large and diverse population of transgender female youths. Our participants were recruited from HIV clinics and street or club venues in 2 large urban areas, limiting the generalizability of our findings to other segments of the transgender female youth population. Another limitation of our data collection was that the partner type variable would not support modeling the association for some HIV risk behaviors, such as the HIV status of commercial partners, because we did not ask for that information.
Our questions about main, casual, and commercial partner types also differed. For main partners, we asked about the current main partner, if applicable. Respondents were asked about only the most recent casual partner but about any commercial partner. Therefore, risk behaviors that occurred with a casual partner who was not the most recent were not captured, even though they may have represented the predominant pattern of behavior with all such partners. By contrast, a behavior engaged in with any commercial partner would potentially be identified, whether or not it represented the respondent's typical behavior with such partners.
Our analysis was also limited by the discrete list of demographic variables we used. If we had included more demographic variables, they might have explained the variance in risk behavior better than did partner type. Finally, in-depth data on how decisions were made by partner type and how decisions were made after discussions about the HIV status of partners, which were not collected, might have contributed to a better understanding of the findings.
Research and interventions that focus on understanding and mitigating risk behaviors by partner type may make important contributions to risk reduction in the transgender female youth community. A longitudinal study of risk behaviors by partner type, coupled with periodic HIV testing, would contribute to a better understanding of which risk behaviors and partner types account for increased HIV risk among transgender female youths. Programs and interventions are also needed that focus on working with transgender youths and their main partners to reduce risk within steady relationships.
Future research and programs focusing on main partners should integrate effective communication strategies for discussing HIV status and should formulate ways that youths can negotiate being tested for HIV with their partners to ensure that condom-use decisions are well-informed. Finally, skills building that enables transgender female youths to evaluate the fidelity of their partnerships may also be an important part of an intervention with this group of youths and their partners.■
The Adolescent Trials Network for HIV/AIDS Interventions (ATN) is funded by the National Institute of Child Health and Human Development (grant U01 HD40533), with supplemental funding from the National Institute on Drug Abuse, the National Institute of Mental Health, and the National Institute on Alcohol Abuse and Alcoholism.
We acknowledge the contributions of the staff members who contributed to collection, management, analysis, and review of these data: Max Madrigal and Gilberto Soberanis. We acknowledge the contribution of the investigators and staff at participating ATN sites: Children's Memorial Hospital, Chicago, IL, and Children's Hospital of Los Angeles, CA. The study was scientifically reviewed by the ATN's Behavioral Leadership Group. Network scientific and logistical support was provided by the ATN Coordinating Center, University of Alabama at Birmingham (Craig Wilson, Cindy Partlow). Network operations and analytic support were provided by the ATN Data and Operations Center, Westat, Inc (Jim Korelitz, Barbara Driver). We are grateful to the members of the Transgender Advisory Committee for their insight and counsel and are particularly indebted to the youths who participated in this study. We also acknowledge Gary Harper, who served as a consultant on this project.
Reprints can be ordered at http://www.ajph.org by clicking the “Reprints/ Eprints” link.
Note. The content of this article is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Human Participant Protection
Protocol approval was obtained for this study from the institutional review boards of Children's Hospital Los Angeles, Howard Brown Health Center, and Children's Memorial Hospital Chicago. They granted a waiver of parental consent for minors; therefore written assent was obtained for youths aged younger than 18 years.