To inform the development of the survey instrument, we enrolled 13 TGW to participate in qualitative in-depth interviews where they described lived experiences of trauma, both individually and among members of their community (results reported elsewhere) (4). These participants were recruited through flyers posted in community spaces (e.g., bars, restaurants, nail salons, clubs) in the Birmingham, Alabama metropolitan area as well as through partnerships with local LGBTQ+ servicing organizations and clinics. Recruitment methods included word-of-mouth referral, snowball sampling, and social media messaging. Participants met with experienced qualitative researchers and were asked to share their lived experiences of trauma as TGW in the US Deep South as well as the role that SRH plays in that experienced trauma. The research team then developed measures relevant to the common themes of trauma elucidated from these interviews.

All trauma-related items in the survey were adapted from themes elucidated from in-depth interviews (4) as well as three existing, publicly available validated trauma instruments developed for the general population: the Life Events Checklist for DSM-5 (LEC-5) (14), the Trauma History Questionnaire (THQ) (12), and the Stressful Life Events Screening Questionnaire (SLESQ) (13). Some items required a response on a frequency scale of an experienced trauma (i.e., “never,” “once,” “2–3 times,” “4 or more times”) while others included “yes” or “no” responses, with some items requiring a follow up question (usually free text) if the answer was “yes.” The survey instrument contained some additional items that allowed participants to further characterize an experience, typically by providing free text responses. These components were not considered in the psychometric analyses of the survey instrument. Similar items were grouped together in five different subdomains (healthcare related experiences, sexual and relationship experiences, crime related and general trauma experiences, gender dysphoria experiences, and discrimination experiences) to capture information about similar types of traumatic experiences. These domains are summarized in Table 1, and the survey instrument development process is shown visually in Figure 1. The full survey that was administered to participants can be found in Appendix 1.

After initial development, the survey instrument was then tested with 14 TGW using cognitive interviews (20, 21). Cognitive interviewing is a qualitative research method which involves asking participants to review survey materials and provide insights and feedback into how they interpret questions and is a commonly used validation tool in survey development (22, 23). The research team reviewed feedback obtained during the cognitive interviews and further refined the items of the instrument before creating a web-based version of the instrument in Qualtrics that could be administered to study participants.

TGW were recruited from both community and clinical sites in Birmingham, AL to complete the survey instrument. A variety of community-engaged strategies were implemented to engage the local TGW community including administering the survey on tablets during clinic visits at the University of Alabama at Birmingham (UAB) Gender Health Clinic, an in-service presentation and enrollment event at a local advocacy organization run by and serving Black TGW, and a text messaging campaign inviting patients of a local LGBTQ+ community health center to participate. While recruitment efforts were based in Birmingham, the clinics and orgnizations involved in this study have wide catchment areas that serve the state of Alabama broadly, including rural, urban, and suburban communities. Recruitment took place from April 2024 to September 2024. Individuals were eligible for the study if they were 18 years or older, English-speaking, and self-identifying as a TGW or any other transfeminine gender identity.

This study protocol was approved by the UAB Institutional Review Board (Protocol #300009382). This approval was granted with the sensitive nature of data and vulnerability of participants in mind. In order to protect confidentiality and anonymity, potential participants were assigned a unique code preventing linkage of survey data to any individual, and they could either complete the survey in-person or on a tablet or be sent an anonymous link to the survey to complete on their own mobile device. Further, signed consent was waived to avoid participants being required to affix their name to any study documents; instead, participants reviewed an information sheet about the study and consented to participate by clicking an electronic check box. Participants were informed that they could decline to answer any survey item without explanation, and they could decide to end their study participation at any time without consequence. In the event that survey questions caused undue psychological or emotional stress, immediate referral to licensed mental health providers was available at study sites to intervene with participants. The survey instrument and additional socio-demographic questions were subsequently administered to participants who agreed to participate in the study. Participants were compensated with $20 for completing the survey.

Data were collected and stored in a secure, HIPAA-compliant Qualtrics database (Qualtrics, Provo, UT). Descriptive statistics for the sample were calculated. This included the overall sociodemographic characteristics of the sample as well as the results of the five trauma subdomains. Sociodemographic questions included details about race, ethnicity, age, insurance status, substance use history, sexual orientation, genders of sexual partners, and experience with gender-affirming care (e.g., gender-affirming hormone therapy, gender-affirming surgical procedures). For the healthcare related experiences and discrimination experiences subdomains, response options range from “never” (scored as 0) to “4 or more times” (scored as 3). Items are summed for a score range of 0 to 9 for the healthcare related experiences subdomain and 0 to 27 for the discrimination experiences subdomain. For the subdomains of sexual and relationship experiences, crime-related and general trauma experiences, and gender dysphoria experiences, response options were dichotomous with ‘No’ (scored as 0) and ‘Yes’ (scored as 1). For the sexual and relationship experiences subdomain, the items summed for a score range of 0 to 3. For the crime-related and general trauma experiences subdomain, the items summed for a score range of 0 to 12 and for the gender dysphoria experiences subdomain, the items summed for a score range of 0 to 2. Higher scores indicated more negative experiences for each subdomain. Cronbach’s alpha coefficients were calculated for each subdomain to determine internal consistency. The means and standard deviation were determined by dividing the sum of each subdomain by the number of items. Race was dichotomized into White versus non-White (Black/African-American, Native American, and Multiple races) for stratified analyses. To compare the means of the subdomains by race, the Mann–Whitney U test was employed due to the non-normal distribution of the data. To account for multiple subdomains within the survey instrument (n = 5), we input raw p-values and used proc. multtest to adjust p-values for multiple comparisons using the stepdown Bonferroni correction. Additionally, items under the healthcare related experiences and discrimination experiences subdomains were dichotomized to reflect whether participants had experienced a given situation at least once (scored as 1) versus none (scored as 0). Items under sexual and relationship experiences, crime-related and general trauma experiences, and gender dysphoria experiences were already dichotomized. The number and frequency for each item under each subdomain was calculated for descriptive purposes only. All analyses were performed using SAS 9.4 (Cary, NC, United States).

Between April and September 2024, 105 TGW enrolled and completed the instrument. The median age was 30 years (range 19–73 years). Most participants identified as White (n = 55, 52.9%) or Black/African American (n = 40, 38.1%). Most participants had health insurance (n = 94, 89.5%). The most frequently reported illicit substances used (lifetime) included marijuana (n = 63, 60.0%), crack or powder cocaine (n = 28, 26.6%), and hallucinogenic drugs (n = 20, 19.0%); 29 participants (27.6%) reported no history of illicit substance use. Mental health conditions were commonly reported, with 85 participants (81.0%) designating that they had at least one diagnosis. Depression (n = 64, 61.0%) and anxiety (n = 59, 56.2%) were reported as the most common mental health diagnoses. Participants reported a variety of sexual orientations and genders of sexual partners. Among our sample, most participants reported being on gender affirming hormone therapy at the time of survey completion (n = 79, 75.2%), but most reported no history of gender affirming surgeries or procedures (n = 82, 85.4%). Table 2 summarizes the demographic characteristics of this sample.

The subdomains of healthcare-related experiences, crime-related/general trauma experiences, and discrimination experiences all showed acceptable internal consistency (Cronbach’s α = 0.787, 0.870, and 0.870, respectively) after dichotomization. However, subdomains measuring sexual/relationship experiences and gender dysphoria had lower reliability (Cronbach’s α = 0.597 and 0.499, respectively). These results are displayed in Table 3.

The means of the subdomains of the instrument, stratified by race, for all participants are shown in Table 4. For descriptive purposes, the individual items of each subdomain, stratified by race, are presented in Supplementary Table 1.

Healthcare-Related Experiences. There were no statistically significant differences between White and non-White participants for the mean scores for this subdomain (p = 0.163). Among participants, over half reported having experienced mistreatment while receiving healthcare (n = 53, 51.4%) and having feared for their mental or emotional wellbeing in a healthcare setting (n = 52, 49.5%) because of their gender identity at least once (Supplementary Table 1).

Sexual and Relationship Experiences. There was a statistically significant difference in subdomain mean scores between White (mean score = 0.35 ± 0.36) versus non-White (mean score = 0.55 ± 0.34) participants (p = 0.011). Among participants, a history of forced sexual intercourse (n = 35, 33.3%), lifetime participation in transactional sex (n = 44, 41.9%), and having been in a relationship characterized by unequal power dynamics (n = 65, 58.0%) were common experiences. In our sample, non-White participants reported a greater frequency of participation in transactional sex (non-White n = 28 [56.0%] versus White n = 16 [29.1%]) (Supplementary Table 1).

Crime-Related and General Trauma Experiences. There was a statistically significant difference in subdomain mean scores between White (mean score = 0.37 ± 0.25) versus non-White (mean score = 0.59 ± 0.34) participants (p = 0.002). All items within this subdomain were frequently reported among participants, but the three most prevalent were experiencing unemployment or financial struggles (n = 70, 66.7%), having a serious or life-threatening illness (n = 63, 60.0%), or having been abandoned, disowned, or estranged by members of their biological family (n = 60, 57.1%). Non-White participants more frequently reported the following compared to White participants: experiencing robbery or attempted robbery (non-White n = 25 [50.0%] versus White n = 7 [12.7%]), having been arrested or incarcerated (non-White n = 24 [48.0%] versus White n = 12 [21.8%]), having been attacked with a gun/knife/other weapon (non-White n = 27 [54.0%] versus White n = 7 [12.7%]), and experiencing homelessness (non-White n = 36 [72.0%] versus White n = 10 [18.2%]; Supplementary Table 1).

Gender Dysphoria Experiences. There was a statistically significant difference in subdomain mean scores between White (mean score = 0.93 ± 0.2) versus non-White (mean score = 0.64 ± 0.38) participants (p = 0.001). Experiences with gender dysphoria associated with sex assigned at birth (n = 87, 83.7%) and being misgendered (n = 79, 76.7%) were common among study participants. White participants more commonly reported experiencing gender dysphoria associated with their sex assigned at birth compared to non-White participants (White n = 54 [98.2%] versus non-White n = 33 [67.4%]; Supplementary Table 1).

Discrimination Experiences. There were no statistically significant differences between White and non-White participants for the mean scores for this subdomain (p = 0.163). The three most commonly endorsed items were experiencing discrimination on the street or in a public setting (n = 70, 66.7%), at work (n = 58, 55.8%), or when getting hired for a job (n = 49, 46.7%) on the basis of gender identity. Non-White participants more frequently reported experiencing discrimination related to getting housing at least once compared to White participants (White n = 5 [9.1%] versus non-White participants (n = 20 [40%]; Supplementary Table 1).