|Home | About | Journals | Submit | Contact Us | Français|
Transgender women (i.e. male-to-female transgender persons) are individuals whose gender identities are discordant with the male sex they were assigned at birth. During the developmental period from early adolescence through young adulthood, many young transgender women struggle to develop a coherent sense of self while addressing feelings of guilt and shame about their identities, pressures to conform to familial, peer, and gender norms, and often the need for secrecy, either to “pass” in their chosen gender, or to hide their true feelings to avoid rejection and discrimination.1 Instead of support and understanding from family, friends and other adults, these women often experience social rejection and marginalization due to their gender identity and expression, as well as perceived sexual orientation.2-4 This rejection and marginalization is particularly salient during this period of developmental vulnerability and often results in severe consequences, as evidenced by high rates of homelessness, trading sex for food and other basic needs, and incarceration.5-9 A growing body of literature suggests that the marginalization experienced by these young women contributes to a wide range of negative health outcomes including psychological distress, substance abuse, and victimization, including verbal, physical, and sexual abuse. All of these outcomes are related to sexual risk behavior and HIV infection.5-8
The prevalence of HIV infection among transgender women is equal to or greater than among other traditionally high risk groups, such as men who have sex with men (MSM).10 In a review of 29 studies of HIV incidence, prevalence, and related risk behavior among transgender individuals completed between 1990 and 2003, the average laboratory confirmed HIV prevalence for transgender women across age groups was 27.7% (four studies) and the average self-reported HIV prevalence was 11.8% (18 studies, i.e., that included prevalence estimates).10 More recent data from local testing of over 500 transgender women with no known prior positive HIV test results in Miami Beach, Florida, San Francisco, and Los Angeles found 12% HIV infection, which suggests a high percentage of unrecognized HIV infection in this population.11 In an analysis by age, the highest number of new HIV infections, representing 45% of all cases, were detected among those ages 20-29.11
Estimates of HIV prevalence among young transgender women are scarce and based on very small, non-probability samples. A community-based study of ethnic minority young transgender women (ages 16-24; N=51) found 22% self-reported HIV-positive status.6 A prior analysis of study data from young transgender women, ages 15-24 (N=151, data further analyzed herein), found a comparable rate of 19% self-reported HIV infection.7 The higher rate of self-reported HIV-infection among young transgender women, in comparison to transgender women more generally, may be a result of relatively high rates of HIV testing. A total of 87% of young transgender women in this study had been tested for HIV infection at least once.7 However, self-reported prevalence of HIV-infection among these young women is still likely to be underestimated given evidence of unrecognized infection among those ages 20-29.11
High rates of unprotected receptive anal intercourse in this population,10 place them at risk of both acquiring and transmitting HIV infection. In the review of 29 studies referred to above, 31.7% of transgender women reported multiple, primarily male, sex partners and 48.3% reported having sex with casual partners. The average proportion of any unprotected receptive anal intercourse (URAI) was 44.1%, and the proportion of unprotected insertive anal intercourse (UIAI) was 27.4% (assessed across various recall periods).10 Estimates of sexual behavior among young transgender women in the community-based study referenced above (N=51) documented a rate of 59% self-reported unprotected anal intercourse (UAI, receptive or insertive) in the past 12 months.6
Multiple psychosocial health problems, including psychological distress, substance use, violence and victimization are common among transgender women. For example, community surveys suggest rates of depression and suicidality that are up to three times higher than the general population.3,12-14 Likewise, evidence indicates substance use is common (i.e., prior 30-day use of alcohol and marijuana of 50% and 38% respectively)15 with high prevalence of sex under the influence of drugs and alcohol.15-17 Studies of violence and victimization among transgender women estimate that between 21% to 68% have experienced forced sex9,18 and between 37% and 65% have experienced physical abuse9,15,18 either as a child or an adult.18
Similarly, evidence suggests that psychosocial health problems are also prevalent among young transgender women. Garofalo and colleagues, in a community-based study of 51 ethnic minority young transgender women cited above found that while self-esteem and depression were within the normal range on average, both were independently associated with UAI.5 With regard to substance abuse, past-year alcohol and marijuana use were reported by 65% and 71% of participants, respectively.6 Wilson and colleagues reported that over 90% of young transgender women in the study sample had used substances in their lifetime (alcohol 88%, marijuana 63%, cocaine 30%, ecstasy 32%, and methamphetamine 30%).7 Sex under the influence of alcohol or drugs was reported to be 50% and 53% in these separate studies 5,8 and was significantly associated with unprotected anal sex.5 With regard to violence and victimization, Garofalo and colleagues reported that over half of participants reported a history of forced sex, which was significantly associated with sexual risk behavior.5 Reported fear of partner anger and rejection were also noted as reasons young transgender women engaged in unsafe sex.6
We conclude that psychological distress and substance abuse, as well as frequent experiences of violence and other forms of victimization may contribute to HIV risk in this population, potentially fueling heightened rates of HIV infection. In light of the high HIV prevalence rates and the complexity of risk factors associated with risk behaviors and HIV acquisition, it may be appropriate to examine HIV risk among these young women within the framework of syndemic theory, as has been suggested by a leading group of experts in transgender health.19 Merrill Singer coined the term, “syndemic” in reference to the health crisis among poor and underserved inner-city women in the early 1990s, which included the co-occurrence of substance use, AIDS, and violence.20 As described by Singer, a syndemic involves “a set of enmeshed and mutually enhancing health problems that, working together in a context of deleterious social and physical conditions that increase vulnerability, significantly affect the overall disease status of a population” (p.15).21 Thus, a syndemic describes more than the interaction of diseases, but rather the mutually reinforcing interaction of disease and social conditions.21-23 Singer describes syndemics as occurring in “noxious social conditions” and are often produced by “structural violence of social inequality.”22
Stall and colleagues applied syndemic theory to the study of HIV-related sexual risk among urban men who have sex with men (MSM). They found that an increasing number of psychosocial health problems, including poly-substance use, depression, partner violence and childhood sexual abuse was significantly and positively associated with high-risk sexual behavior and HIV infection.24 Similarly, in an urban sample of ethnically diverse young MSM, ages 16-24, Mustanski and colleagues found that an increasing number of psychosocial health problems, including binge drinking, street drug use, regular marijuana use, psychological distress, intimate partner violence and sexual assault increased the odds of multiple anal sex partners, unprotected anal sex, and HIV-positive status.25
To our knowledge, syndemic theory has not previously been applied to the study of HIV risk among young transgender women; however given their marked social and economic marginalization and high prevalence of psychosocial health problems and HIV infection, the principles underlying this theory may well apply. The syndemic model, therefore, serves as a framework guiding the analysis described herein. We chose specific psychosocial health problems for inclusion in our syndemic model, as the data would allow, that are similar to factors examined among both urban poor and underserved women and urban MSM, and which reflect the life circumstances of young transgender women.
Based upon the evidence of the health disparities among young transgender women,5-8 we hypothesize that a syndemic of co-occurring health and psychosocial factors: low self-esteem, poly-substance use, victimization related to transgender identity (i.e., verbal threats and insults; being chased, followed or having property damaged; and being physically assaulted), and intimate partner violence (i.e., partner controlling activities; verbal harassment; threatening physical safety; sexual violence; and being pressured/forced to hide female gender identity) are additive and associated with HIV-infection and sexual risk for HIV infection. That is, the more psychosocial health problems reported, the greater the risk for both unsafe sexual behavior and HIV infection would be expected. In addition, Singer specified that a syndemic develops in a context of deleterious social conditions that increase vulnerability. Thus, we further test indicators of social marginalization as correlates of this clustering of psychosocial factors. Specifically, the objectives of this study were to: 1) assess whether or not multiple psychosocial factors are additive in their relationship to sexual risk behavior and self-reported HIV status among young transgender women (i.e., can be characterized as a syndemic) and 2) assess the relationship of indicators of social marginalization, such as a history of commercial sex work, homelessness, and incarceration to these psychosocial factors.
The Transgender Research Youth Project (TRYP), part of the NICHD-funded Adolescent Medicine Trials Network for HIV/AIDS Interventions, used a community-based participatory research model to conduct a two-phase cross-sectional study examining HIV-related risk behavior among young transgender women in Chicago and Los Angeles. Both cities have large communities of young ethnically-diverse transgender women. The first phase of the study included formation of a peer and community-based transgender advisory committee (TAC) at the lead site in Los Angeles (funding did not allow for the formation of advisory groups in each city) to guide questionnaire and study protocol development, refine recruitment strategies, and complete a pilot test of all study procedures. In the second phase of the study, an interviewer-administered survey was used to collect cross-sectional data. A sub-sample also completed in-depth qualitative interviews.
In 2005-2006, young transgender women were recruited for participation from medical clinic settings, social venues, including bars and clubs; and through street outreach and referrals from study participants. Eligibility criteria included: 1) self-identifying as a transgender woman and/or not identifying with assigned male birth sex; 2) residing in the Chicago or Los Angeles metropolitan area; 3) ability to speak and understand either English or Spanish; and 4) being 15-24 years of age.
Each site obtained Institutional Review Board (IRB) approval for the study protocol prior to implementation. A waiver of parental consent was obtained for participants aged 15-17. Written consent (or assent in case of minors) was obtained prior to study participation. Participants received a $25 incentive for completion of the study.
Data were collected through an interviewer-administered survey, which included questions regarding general demographic characteristics, HIV risk behaviors, indicators of social marginalization, as well as psychosocial health status. When possible, items from existing questionnaires were used or modified to be sensitive and appropriate for young transgender women. The entire survey instrument was reviewed and approved by the TAC.
Participants reported their age, race, ethnicity, and HIV status. Given the small sample size and in order to contrast the two predominant racial and ethnic groups in the sample (African American, Latina) with all others, we created two dummy variables for race and ethnicity to reflect 1) African American (1= African American, 0= non-African American) and 2) Latina (1= Latina, 0= non-Latina) race and ethnicity.
We created a composite syndemic index of four health and psychosocial factors: low self-esteem, poly-substance use, victimization, and intimate partner violence. Self-esteem was assessed using the total score of a modified version of the Rosenberg Self-Esteem Scale (RSES26; Rosenberg, 1965; i.e., modified by the TAC to refer to transgender status, specifically). Cronbach's alpha for the modified self-esteem scale was 0.72. For inclusion in the syndemic index, this variable was dichotomized at the mean (M=24, s.d. =4.5) to reflect low self esteem (1) vs. high self esteem (0). In addition to providing an advantageous distribution for the purposes of analysis, scores on this variable of 24 or higher correspond to agreement/strong agreement with indicators of a positive orientation towards the self. Participants reported any lifetime use of alcohol or other substances, including marijuana, hallucinogens (i.e. LSD, mushrooms), PCP, ecstasy/MDMA, ketamine, GHB, crack, cocaine, heroin and methamphetamines (1= yes, 0= no). These items were summed and a variable reflecting poly-substance use was created to reflect any (i.e., lifetime) self-reported use of three or more substances (1=yes, 0=no). Although both lifetime and 30-day recall periods were used to collect substance use information in this study, lifetime indicators were chosen for the poly-substance use indicator given the young age of participants and potential negative impact of even one-time use of substances. Victimization was assessed using a ten-item measure of victimization developed by D'Augelli and colleagues27 for lesbian, gay and bisexual youth, which was adapted by the TAC to reflect the language and experiences of young transgender women. The measure included three dimensions of victimization based on gender identity, including: (1) verbal threats and insults; (2) being chased, followed or having property damaged; and (3) being physically assaulted due to gender identity. For example, one item asked: “How many times have you been punched, kicked, or beaten because you are, or were thought to be transgender?” Responses were coded on a frequency scale (0=never, 1=once, 2=twice). Average scale scores were dichotomized to reflect endorsement of victimization due to gender identity (i.e., scores equaling 0 thru <1 were coded as 0 and scores ≥1 through 3 were coded as 1). Intimate partner violence was assessed with five items developed specifically for young transgender women by the TAC. The five items assessed experiences in five different domains, specifically: 1) partner controlling activities; 2) verbal harassment; 3) threatening physical safety; 4) sexual violence; and 5) pressured/forced to hide female gender identity. For example one item asked, “Has a partner made you do something that did not agree with your gender identification (e.g. make you hide your make-up)?” Responses were rated on a scale where 1=yes, 0=no. Summed responses were then recoded to reflect any experience of any intimate partner violence (1) versus no such experience (0). The Syndemic Index was created by summing these (four) dichotomous variables, as has been done in previous studies from a syndemic perspective.24,25 We censored the index at 3 to correct for non-normal distribution (i.e., only 4.6% of the sample scored 4; final range 0-3).
We created four dichotomous indicators of social marginalization using self-reports of ever having been homeless, incarcerated, or participating in commercial sex work/survival sex (1=ever, 0=never). Ever homeless was indicated by a “yes” response to the following item, “Have you ever spent one night or more in an emergency shelter, transitional housing facility, welfare hotel or a public or private place not designed for sleeping (e.g. car, park, etc.)? A prior history of incarceration was indicated by a “yes” response to the question, “Have you ever been in the correctional system? (For example, convicted of a crime and sent to juvenile corrections, jail, prison, probation, parole)? Participation in commercial sex work/survival sex was indicated by a “yes” response to the question, “Have you ever traded sexual activity or favors for food, money, a place to sleep, drugs or other material goods?”
Sexual risk behavior was a primary outcome variable and was defined as having any unprotected anal intercourse (i.e., insertive or receptive) with any type of sex partner within the past three months (1=Yes, 0=No; responses indicating no prior anal sex or always having protected anal sex in the past three months were coded as 0).
We created a dummy variable for self-reported HIV positive status (1=HIV positive, 0=HIV negative/untested/unknown).
Correlational analyses were conducted to measure the size and significance of relationships between components of the syndemic index (in their original un-dichotomized format), self-reported HIV status, and unprotected anal sex (i.e., point-biserial correlations). Multiple logistic regression models were used to assess the relationship between the syndemic index and primary outcomes (unprotected anal sex, self-reported HIV serostatus). Odds ratios were calculated for sexual risk behavior and self-reported HIV status for an increasing number of syndemic factors, with the syndemic index specified as categorical variable (with “0” as the reference group). The relationship of the syndemic index to social indicators of marginalization (i.e., ever homeless, ever incarcerated and ever participating in sex work) was then assessed using a multiple linear regression model. Age was used as a continuous variable and was controlled in all multiple regression analysis, given the well-established relationship between age, sexual risk behavior and HIV-infection. In the model regressing the syndemic index on indicators of social marginalization, African American race was also controlled in the analysis (in addition to age), given significant correlations with indicators of social marginalization.
A total of 151 young transgender women participated in the study (mean age=21, SD=2.5; Chicago n=75, Los Angeles n=76). A total of 39% of participants described their race and ethnicity as Black or African American, 38% as Latina, 13% as other (i.e. multiple races and/or ethnicities, Native American, other race and/or ethnicity), 5% White, and 5% Asian/Pacific Islander (see Table 1). Over half (61%) of the participants had at least a high school diploma or equivalent and a third (34%) were currently employed. Self-reported HIV-positive status of participants was 16% for the overall sample (19% among those who reported testing for HIV). Over a third (32%) of participants reported UAI in the past three months. In addition, almost half of participants had a prior history of homelessness (43%), more than half had a history of incarceration (52%), and a majority had engaged in sex work (67%).
With regard to correlations between syndemic factors, sexual risk behavior, and self-reported HIV status (as indicated in Table 2), we found that both poly-substance use and intimate partner violence were significantly and positively related to both self-reported HIV infection and sexual risk behavior, as well as with each other, and that victimization and intimate partner violence were significantly and positively related to each other. All other correlations (except for self-esteem with self-reported HIV infection and poly-substance use) were in the expected direction, although not significant.
In multiple logistic regression analysis, (as shown in Table 3), the effect of increasing numbers of syndemic factors was additive for self-reported HIV-status for one to two and two to three/four factors (i.e., OR=1.65 for one factor, OR=3.13 for two factors, and OR=6.61 for three/four factors). For sexual risk behavior the effect was additive from one to two factors (i.e., OR=2.87 for one factor, OR=5.46 for two factors, but not for three/four factors with OR=4.53).
In analysis of the relationship between indicators of social marginalization and the syndemic index, we found that a history of ever engaging in sex work was significantly associated with the syndemic index (B=.70, s.e. =.20, p<.01), as was a history of incarceration (B=.49, s.e.=.18, p<.01), however, a history of homelessness was not (Table 4).
These data provide preliminary evidence suggesting that multiple health-related psychosocial factors, including low self-esteem, lifetime poly-substance use, intimate partner violence and victimization may be additive in their association with sexual risk behavior and self-reported HIV infection among young transgender women. While there was no significant association between one syndemic factor and sexual risk and one/two syndemic factors and seropositivity, this may be a function of small sample size. The additive trend in these factors in their relationship to sexual risk behavior and HIV status among young transgender women are similar to prior findings among urban MSM and ethnically diverse young MSM.24,25 In addition, findings highlight the social situations or life experiences that are associated with these conditions; namely, a history of sex work and incarceration, which contribute to these co-occurring conditions and their resulting heightened HIV risk. While prior evidence of disproportionately high prevalence of psychosocial health conditions in young transgender women has been documented, this study provides evidence of a potential additive effect on sexual risk for HIV and self-reported HIV-positive status and the interrelationship of these conditions with key indicators of social marginalization.
With regard to the implications of these findings for intervention, given the young age of these transgender women, developmental considerations are of particular importance. The coalescence of low self-esteem, poly-substance use, victimization and intimate partner violence at such a young age may present a psychosocial health burden which is difficult to overcome. Future intervention studies are needed that could help to determine whether intervening on multiple health issues together or in sequence is effective in preventing sexual risk behavior. In addition, prevention efforts targeting this population of young people may require a multi-systems approach targeting both social and health-related factors.
Our findings regarding the relationship of sex work and incarceration with psychosocial health problems and the interrelationship of these psychosocial health problems with sexual risk and HIV infection are also important. In a prior analysis of these data, Wilson and colleagues found a direct relationship between sex work and HIV status; 26% of those who had engaged in sex work reported being HIV positive compared to 6% who had never engaged in sex work. This is consistent with prior studies suggesting disproportionate prevalence of HIV infection among transgender sex workers, particularly in comparison to other sex workers.28 It is important to note that not all sex work presents equal risk. New evidence is emerging that risk may differ by type of sex work (i.e., street work vs. internet-based work).29 Future studies should determine the varying risks by sex work type in order to more effectively target prevention interventions. Our findings suggest that in addition to the direct impact on self-reported HIV status, sex work may also have an indirect relationship on this status through psychosocial health problems. Thus, an exploration of the potential mediation of the effect of sex work on HIV status by psychosocial factors should be the subject of future study.
There are several additional implications for future research from these findings. From a theoretical perspective, it is important to note that this study was a secondary analysis of data collected for the general study of young transgender women. Therefore, syndemic theory did not directly influence the research design, identification of constructs, or data collection instrument. Consequently, measurement was a particular limitation in this study. In the absence of a measure of psychiatric distress, we used low self-esteem as a mental health indicator within our syndemic index. However, in correlational analysis, while we found self-esteem to be related to the other components of the syndemic index and outcomes of interest in the expected direction (in most cases), correlations were weak and not statistically significant. Similarly, our measure of victimization (with a limited frequency response scale, i.e., never, once, twice, more than twice) was weakly related to other syndemic components and outcomes of interest. Thus, more sensitive measures of psychological outcomes and victimization should be included in future studies of this population, particularly given recent evidence suggesting that early gender-related victimization is a key factor in psychiatric distress among young transgender women.13 A recent study of 571 transgender women (ages 19-59) in the New York City metropolitan area suggests a dose-response relationship between abuse and depression in adolescence. Prevalence of depression and suicidality during early adolescence (ages 10-14, based on retrospective recall) in their young cohort (ranging in age at time of the study from 19-39), was nearly two to three times higher than adolescents in general (which declined as they entered young adulthood). In addition, they found that gender-related abuse, both psychological and physical, was extremely high during early and late adolescence. Abuse and psychiatric distress were related in a dose-response manner in both early and late adolescence. While the impact of psychiatric distress and abuse on HIV-risk and HIV status were not assessed, these findings are important within the context of the early development of young transgender women and may have implications for other health conditions. Thus, research from a syndemic perspective with more sensitive measures is warranted.
These findings should be considered within the context of other limitations. Data were collected from a small convenience sample of young transgender women from two large cities, therefore results may not generalize to other populations of young transgender women. In particular, this sample was characterized by a high prevalence of prior homelessness (43%), incarceration (52%), and sex work (67%), which may not characterize other populations of transgender youth. This was a particularly high risk urban sample. In addition, given the small sample size, we were not able break down sexual risk by partner type or concordant/discordant HIV status. Furthermore, HIV testing was not included in the study protocol; therefore, HIV status was self-reported and thus, likely underreported. Larger samples with collection of biomarkers would allow for more rigorous analysis of risk. In addition, this was a cross-sectional study, which prevents an assessment of causal ordering. Future studies are needed to describe the onset and trajectory of psychosocial health problems and their impact on HIV-related risk. This requires longitudinal research focused specifically on syndemic components measured over time. As recently recommended by the Institute of Medicine in a groundbreaking report on the health of lesbian, gay, bisexual and transgender people, a more rigorous program of research, including longitudinal research on transgender health is needed, given that most studies of transgender individuals are cross-sectional with small non-probability samples.30 As well, data were collected using interviewer-administered surveys and thus, while under-reporting of sensitive behavior might be expected, respondents in this sample reported high levels of both substance use and sexual risk behavior.7,8 Finally, these data provide preliminary evidence of the additive relationship of multiple psychosocial health factors to HIV sexual risk behavior and HIV serostatus among young transgender women, but further validation studies are warranted, particularly given the small sample size.
This research was funded by the Adolescent Trials Network for HIV/AIDS Interventions (ATN) grant No. U01 HD052172 from the National Institutes of Health through the National Institute of Child Health and Human Development (R. Garofalo). We acknowledge the contribution of the investigators and staff at the following ATN sites that participated in this study: Children's Memorial Hospital, Chicago (R. Garofalo); and Children's Hospital of Los Angeles, Los Angeles, CA (M. Belzer). We thank the Adolescent Medicine Trials Network 039 study staff team, our advisory board members, and members of the transgender youth community who participated in the study.
Reprints can be ordered at http://www.ajph.org by clicking the “Reprints/Eprints” link.