Abstract
Inclusive measures of gender are critical for health equity research. This study compared the reliability and applications of two different approaches for measuring gender in response to emerging community concerns regarding the potential harms of asking about sex assigned at birth (SAAB) within transgender and gender diverse (TGD) populations. Using data from a 2021 survey of LGBTQ+ people in Washington state, we compared approaches for measuring gender via a two-step question that collected data on: (1) current gender and SAAB versus (2) current gender and transgender self-identification. Among 2,275 LGBTQ+ participants aged 9–81, 63% were cisgender, 35% TGD, and 2% were not categorized. There was near perfect agreement between the two methods in their ability to identify TGD participants (percent agreement=99.7%, unweighted Cohen’s Kappa=0.99). Among gender diverse participants, stratification by SAAB revealed differences in sexual health outcomes, while stratification by transgender self-identification revealed differences in access to gender-affirming care and lifetime experiences of discrimination. Ascertaining SAAB may be most useful for identifying sexual health disparities while transgender self-identification may better illuminate healthcare needs and social determinants of health among TGD people. Researchers and public health practitioners should critically consider the acceptability and relevance of SAAB questions to their research goals.
Keywords: transgender, nonbinary, gender diverse, two-step method, gender identity, gender modality, SOGI measurement
INTODUCTION
Transgender-inclusive approaches to asking about gender identity in research, clinical settings, and public health practice are critical for identifying health disparities among transgender and gender diverse populations as well as for developing effective interventions that promote health equity.1,2 Few national studies, clinical trials, and surveillance systems ask trans-inclusive questions about gender. Frequently, researchers use binary questions that conflate sex and gender (for example by asking “are you male or female?”) and as a result, misclassify transgender and gender diverse individuals as cisgender. This is a form of institutional erasure that significantly hinders the ability to conduct population health research and contributes to informational inequities for transgender and gender diverse populations.3–5
The two-step method for collecting gender identity data is considered a best practice and is recommended by numerous research institution and leading transgender organizations.6–9 Most commonly, the two-step method includes two questions that separately ask about a person’s gender identity and the sex they were assigned at birth. This method has been validated in both adult and adolescent populations,10–13 has high response rates when used in general population surveys,14 and provides more reliable data on gender compared to single-item questions.15 For example, using a two-step approach is associated with a nearly 5-fold increase in the number of transgender individuals identified relative to a one-step approach that conflates sex and gender.15
However, some researchers, advocates, and transgender and gender diverse people have expressed concern about sex assigned at birth questions.16,17 Asking about sex assigned at birth can be harmful or perceived as invasive for some transgender and gender diverse people, depending on the context.18–20 In a recent qualitative study by Puckett et al. participants recommended two ways to ask about gender identity: (1) the traditional two-step method that includes separate questions for gender identity and sex assigned at birth, and (2) an alternative two-step method that includes separate questions for gender identity and whether participants identify as transgender.18 However, there has been no quantitative assessment of how these two approaches to collecting gender identity data perform relative to one another.
These community-driven recommendations raise important questions about how two-step questions may perform overall, as well as specifically gender diverse populations (i.e., those who do not exclusively identify as cisgender or transgender men or women, such as nonbinary people). A recent report by the Williams Institute estimates that 1.2 million nonbinary people live in the United States, and that while 42% of nonbinary people identified as transgender, more than half (58%) did not.21 Overall, there is very limited research on gender diverse people, and even less investigation of the health and well-being of subgroups of gender diverse populations (for example, by specific gender identity, sex assigned at birth, transgender self-identification, race and ethnicity, etc.).
The present study used data collected from an online survey of LGBTQ+ adults and adolescents in Washington state to assess the reliability and applications of two approaches to measuring gender, with the overall goal of informing transgender health equity research. First, we compared two different approaches for assessing gender via a two-step question that collects data on (1) gender identity and sex assigned at birth, or (2) gender identity and transgender self-identification. For this aim, we assessed the reliability of these two measures in their ability to identify transgender and cisgender study participants. Second, we explored the ability of these different approaches to aid researchers in identify disparities within the gender diverse population. Gender diverse communities are heterogeneous and include individuals with many different gender identities. However, to date, most research on gender diverse communities has either aggregated them as a single group or stratified them based on their sex assigned at birth. Therefore, we specifically examined how data on gender identity, sex assigned at birth, and transgender self-identification may elucidate the health and healthcare needs of subgroups of the gender diverse community.
METHODS
Study Population and Recruitment
Data for the present analysis are from an anonymous online cross-sectional survey conducted by Public Health Seattle & King County (PHSKC) in 2021. The PHSKC “Pride Survey” is an annual survey that has been conducted for surveillance purposes through PHSKC’s HIV/STD Program since 2009. Individuals were eligible to participate if they self-identified as lesbian, gay, bisexual, transgender, queer, nonbinary, or otherwise not-heterosexual or not-cisgender (LGBTQ+) and currently lived in Washington state. The survey was conducted online over a 6-day period from June 25–30, 2021 using a RedCap computer assisted self-interview instrument that could be accessed on desktop or by mobile device. The survey was available in English and Spanish. The survey was advertised through PHSKC’s social media (Instagram, Facebook, Twitter), through LGBTQ+ social media groups, at virtual pride events, at in-person COVID-19 vaccine clinics, and fliers. Participants who completed the survey could optionally enter a drawing to win one of five iPad prizes.
Survey item development, data analyses, and interpretation of results were conducted in partnership with a community advisory board of transgender and nonbinary people from the Seattle area. Ethical approval for this study was received from the University of Washington Institutional Review Board.
Measures
The core PHSKC Pride Survey collects data on sociodemographic variables, sexual behavior, and HIV/STI-related topics that are used for local HIV surveillance purposes. An optional survey module included questions related to social determinants of health (e.g., food insecurity, health care access, and experiences of discrimination). Overall, 69% of participants responded to questions in the optional supplement.
Table 1 displays the survey items used to ascertain current gender identity, sex assigned at birth, and transgender self-identification. All three questions were asked of all participants. The response options for the gender identity question were developed with input from the community advisory board and mirror recommendations from the NIAID Division of AIDS Cross-Network Transgender and Gender Diverse Working Group.6 Importantly, the gender identity question includes several gender diverse response options, including gender nonconforming, genderqueer, nonbinary, and two-spirit. This question also included response options of man/male and woman/female both with and without cisgender and transgender qualifiers. Participants were able to select more than one current gender identity and were also provided with a write-in response option.
Table 1.
Survey Items used to ascertain gender identity, sex assigned at birth, and transgender self-identification, Public Health Seattle & King County’s 2021 Pride Survey, Washington State
| Sex Assigned at Birth | Current Gender | Transgender Self-Identification |
|---|---|---|
|
|
|
|
|
What was your sex assigned at birth? (select one) • Female • Intersex • Male • Don’t know |
Do you consider yourself to be…? (check all that apply) • Cisgender man • Cisgender women • Gender non-conforming • Genderqueer • Nonbinary • Man/male • Transgender man/trans man • Transgender woman/trans woman • Two-spirit • Woman/female • Not listed, specify (write-in) • Don’t know |
Are you transgender? (select one) • Yes • No • Don’t know |
|
|
||
| Two-Step Approach #1: Traditional Method Using Current Gender and Sex Assigned at Birth | ||
|
|
||
| Two-Step Approach #2: Alternative Method Using Current Gender and Transgender Self-Identification | ||
|
|
||
Analysis
We examined response patterns to the gender identity, sex assigned at birth, and transgender self-identification survey items using descriptive statistics. When participants selected more than one gender identity, their data appear in multiple columns and/or rows of descriptive tables.
We assessed the reliability of two different approaches for categorizing participants as cisgender or transgender and/or gender diverse (this binary construct is also called gender modality, and differs from self-identification as transgender).22,23 First, we used a two-step method based on gender identity and sex assigned at birth. Using this method, participants were considered to be transgender or gender diverse if their gender identity differed from their sex assigned at birth (for example, participants who identified as man/male and reported female sex assigned at birth were categorized as transgender men). Second, we used a two-step method based on gender identity and transgender self-identification. Using this method, participants were considered to be transgender if they self-identified as transgender (for example, participants who identified as woman/female and responded yes to the question “Are you transgender?”). We did not categorize participants based on their write-in responses. We were unable to categorize some participants into a gender identity group because they either only provided a write-in response to the gender identity question, responded “don’t know” to the gender identity questions, or were missing data. We calculated the percent agreement and unweighted Cohen’s Kappa to determine the reliability of these differing approaches to the two-step method for identifying participants who are transgender or gender diverse versus participants who only identified as cisgender men or cisgender women.
We also explored the ability of sex assigned at birth and transgender self-identification to identify disparities among gender diverse participants. For transgender men and transgender women who have binary identities, both sex assigned at birth and gender modality (i.e., transgender status) can be inferred. However, this is not the case for gender diverse people. Therefore, we conducted exploratory analyses to identify subgroups of the gender diverse population who may experience health inequities. For our analyses, we define gender diverse participants to be those who did not exclusively endorse a binary gender identity, and included individuals who were gender nonconforming, genderqueer, nonbinary, and/or two-spirit. Using counts, proportions, and chi-square statistics, we examined differences in sociodemographic factors (race/ethnicity, age), social determinant of health (income, unstable housing, food insecurity, experiences of discrimination), and healthcare related factors (insurance status, gender-affirming health care) among subpopulations of gender diverse participants stratified (1) sex assigned at birth, (2) transgender self-identification, and (3) gender identity. We use Poisson regression with robust standard errors to estimate age-adjusted prevalence ratios (PRs) and 95% confidence intervals (CI) comparing (1) gender diverse individuals assigned male at birth (AMAB) to those assigned female at birth (AFAB) and (2) gender diverse individuals who self-identity as transgender to those who do not. Participants could select more than one gender identity; therefore, we are unable to calculate PRs for these overlapping groups. All analyses were conducted in R statistical software (version 4.2.1).
RESULTS
Study Population
There were 2,275 participants who met eligibility criteria. The majority of participants lived in three neighboring counties in Washington state: King County (n=1328, 58%), Snohomish County (n=189, 8%), and Pierce County (m=190, 8%). Among participants from King County, 66% (879/1328) lived in the city of Seattle. The average age of participants was 36.7 (range 9 to 81 years). Most participants were adults aged 25–34 (34%), 35–44 (20%) and 45–64 (20%); only 6% of participants were adolescents <18, 13% were young adults aged 18–25, and 3% of participants were age 65 or older. The survey sample was predominantly White (73%). In addition, 10% of participants were Asian, 4% were Black/African American, 10% were Hispanic/Latinx, 4% were Native American/Alaska Native, and 2% were Native Hawaiian/Pacific Islander. Nearly all participants completed the survey in English (n=2258, 99%), and 17 (1%) participants completed in the survey in Spanish.
Survey Item Response Patterns
There was a high response rate to all three survey items of interest (Table 2). Nearly all participants (98%) selected at least one gender identity response option that was provided, <1% of participants only provided a write-in response, and <1% of participants only selected “don’t know.” Similar response rates were observed for sex assigned at birth (99.7%) and the transgender self-identification question (96%). The most common write-in responses were agender, genderfluid, and transmasculine (Table S1).
Table 2.
Response patterns to gender identity, sex assigned at birth, and transgender identity questions, Public Health Seattle & King County’s 2021 Pride Survey, Washington State
| Survey Item |
|||
|---|---|---|---|
| Response Type | Gender Identity n (%) | Sex Assigned at Birth n (%) | Transgender Self-Identification n (%) |
|
| |||
| Selected a response from the provided list | 2234 (98.2) | 2267 (99.7) | 2187 (96.1) |
| Selected “don’t know” | 51 (2.2) | 8 (0.4) | 88 (3.9) |
| Only selected “don’t know” | 21 (0.9) | ||
| Write-in responses | 57 (2.5) | ||
| Only provided write-in response | 20 (0.9) | ||
Overall, there were 192 unique combinations of responses to these three questions (gender identity, sex assigned at birth, and transgender self-identification). Many transgender and gender diverse participants reported having more than one gender identity (Table 3). This was most common among gender diverse participants, with 73% of gender non-conforming, 73% of genderqueer, 58% of nonbinary, and 67% of two-spirit participants reporting more than one gender. In addition, 43% of transgender men and 29% of transgender women reported additional gender identities, most commonly nonbinary (35% and 22%, respectively). A smaller proportion of cisgender men and cisgender women reported more than one gender identity (4% and 10%, respectively), most commonly nonbinary, genderqueer or gender non-conforming. Most participants that were categorized as cisgender selected the “cisgender man” and “cisgender women” options either alone (69%) or in combination (11%) with the “man/male” or “woman/female” options (Table S2). Similarly, most transgender participants self-selected the “transgender man” and “transgender woman” option either alone (73% and 61%, respectively) or in combination with the “man/male” or “woman/female” options (24%).
Table 3.
Multiple and overlapping gender identities from a select-all-that-apply question format, Public Health Seattle & King County’s 2021 Pride Survey, Washington State
| Cisgender Man | Cisgender Woman | Gender Non-conforming | Genderqueer | Nonbinary | Transgender Man | Transgender Woman | Two-Spirit | |
|---|---|---|---|---|---|---|---|---|
| n (%) | n (%) | n (%) | n (%) | n (%) | n (%) | n (%) | n (%) | |
|
|
||||||||
| N | 614 | 940 | 203 | 253 | 439 | 133 | 77 | 33 |
| Reported Single Gender | 589 (95.9) | 848 (90.2) | 54 (26.6) | 69 (27.3) | 185 (42.1) | 76 (57.1) | 55 (71.4) | 11 (33.3) |
| Reported Multiple Genders | 25 (4.1) | 92 (9.8) | 149 (73.4) | 184 (72.7) | 254 (57.9) | 57 (42.9) | 22 (28.6) | 22 (66.7) |
| Additional Genders Reported: | ||||||||
| Cisgender Man | NA | 0 (0.0) | 9 (4.4) | 8 (3.2) | 10 (2.3) | 0 (0.0) | 0 (0.0) | 3 (9.1) |
| Cisgender Woman | 0 (0.0) | NA | 37 (18.2) | 33 (13.0) | 40 (9.1) | 0 (0.0) | 0 (0.0) | 8 (24.2) |
| Gender Non-conforming | 9 (1.5) | 37 (3.9) | NA | 76 (30.0) | 99 (22.6) | 19 (14.3) | 10 (13.0) | 7 (21.2) |
| Genderqueer | 8 (1.3) | 33 (3.5) | 76 (37.4) | NA | 140 (31.9) | 17 (12.8) | 7 (9.1) | 7 (21.2) |
| Non-binary | 10 (1.6) | 40 (4.3) | 99 (48.8) | 140 (55.3) | NA | 46 (34.6) | 17 (22.1) | 10 (30.3) |
| Transgender Man | 0 (0.0) | 0 (0.0) | 19 (9.4) | 17 (6.7) | 46 (10.5) | NA | 0 (0.0) | 2 (6.1) |
| Transgender Woman | 0 (0.0) | 0 (0.0) | 10 (4.9) | 7 (2.8) | 17 (3.9) | 0 (0.0) | NA | 1 (3.0) |
| Two-Spirit | 3 (0.5) | 8 (0.9) | 7 (3.4) | 7 (2.8) | 10 (2.3) | 2 (1.5) | 1 (1.3) | NA |
| Write-in response | 0 (0.0) | 5 (0.5) | 13 (6.4) | 14 (5.5) | 28 (6.4) | 5 (3.8) | 3 (3.9) | 0 (0.0) |
| Don’t Know | 3 (0.5) | 11 (1.2) | 7 (3.4) | 10 (4.0) | 15 (3.4) | 4 (3.0) | 0 (0.0) | 1 (3.0) |
Participants were able to select more than one response option to the gender identity question, so participants may appear in more than one column.
Reliability Between Two-Step Approaches
Using either approach, 63% of participants were categorized as cisgender and 35% were transgender and gender diverse. There was near perfect agreement between the two-step methods in their ability to identify cisgender and transgender and gender diverse survey participants (Table 4). The percent agreement was 99.7% and the unweighted Cohen’s Kappa was 0.99 when we compared a two-step method based on sex assigned at birth to the two-step method based on transgender self-identification. Both methods were unable to categorize 2% of study participants who had missing data, responded “don’t know” to one or more questions, or only provided write-in gender identity responses. When we stratified by age category, we observe that these questions had high reliability (>99.6% agreement, Cohen’s Kappa >0.84) across all age groups (Table 5). When stratified by race and ethnicity, there was perfect agreement (100% agreement, Cohen’s Kappa = 1.0) for all racial and ethnic minority participants.
Table 4.
Reliability between the two-step methods for identifying cisgender versus transgender and gender diverse participants, Public Health Seattle & King County’s 2021 Pride Survey, Washington State
| Percent Agreement 99.7%; Cohens Kappa 0.990 | |||||
|---|---|---|---|---|---|
|
| |||||
| Two-Step Method with Transgender Self-Identification |
|||||
| Cisgender Only | Any Transgender and Gender Diverse | Not Categorized | Row Totals | ||
|
|
|||||
| Two Step Method with Sex Assigned at Birth | Cisgender Only | 1435 | 1 | 4 | 1440 |
|
|
|||||
| Any Transgender and Gender Diverse | 0 | 792 | 0 | 792 | |
|
|
|||||
| Not Categorized | 1 | 1 | 41 | 43 | |
|
|
|||||
| Column Totals | 1436 | 794 | 45 | 2275 | |
Table 5.
Reliability between the two-step methods methods for identifying cisgender versus transgender and gender diverse participants stratified by age, race, and ethnicity, Public Health Seattle & King County’s 2021 Pride Survey, Washington State
| N | Percent agreement | Cohens Kappa | |
|---|---|---|---|
|
|
|||
| Overall | 2,275 | 99.7% | 0.99 |
| Age Strata | |||
| <18 | 126 | 98.4% | 0.93 |
| 18–24 | 284 | 100% | 1.00 |
| 25–34 | 761 | 99.6% | 0.98 |
| 35–44 | 461 | 100% | 1.00 |
| 45–54 | 281 | 99.6% | 0.97 |
| 55–64 | 183 | 100% | 1.00 |
| 65+ | 69 | 99.6% | 0.84 |
| Race and ethnicity | |||
| Asian | 234 | 100.0% | 1.00 |
| Black | 80 | 100.0% | 1.00 |
| Latinx | 235 | 100.0% | 1.00 |
| Native American/Alaska Native | 89 | 100.0% | 1.00 |
| Native Hawaiian/Pacific Islander | 33 | 100.0% | 1.00 |
| White | 1829 | 99.7% | 0.99 |
Gender Diverse Participants
Given that these two methods performed nearly identically, we then explored different approaches to identifying disparities and heterogeneity among gender diverse participants. There were 660 gender diverse participants (29% of our sample) who reported at least one of their gender identity labels to be gender non-conforming, genderqueer, nonbinary, or two-sprit. Most gender diverse participants selected more than one gender identity, most commonly another gender diverse identity (Table 3). Overall, 116 (18%) were assigned male at birth (AMAB), 534 (81%) were assigned female at birth (AFAB), 3 were intersex, and 7 responded “don’t know” to the sex assigned at birth question. Among these participants, 300 (45%) identified as transgender, 289 (44%) did not identify as transgender, and 71 (11%) responded “don’t know.”
Table 6 presents demographic and health-related outcomes data collected from gender diverse participants stratified by sex assigned at birth and transgender self-identification. Table S3 presents the same data from gender diverse participants stratified by gender identity (e.g., gender non-conforming, genderqueer, nonbinary, or two-sprit). All approaches revealed differences in age among gender diverse participants, such that trans-identified gender diverse participants as well as those AFAB were younger and had lower incomes.
Table 6.
Characteristics of gender diverse participants who identity as gender non-conforming, genderqueer, non-binary or two-spirit, stratified by sex assigned at birth and transgender self-identification, Public Health Seattle & King County’s 2021 Pride Survey, Washington State
| Gender Diverse Participants Stratified by Sex Assigned at Birth |
Gender Diverse Participants Stratified by Transgender Self-Identification |
||||||
|---|---|---|---|---|---|---|---|
| Assigned Male at Birth | Assigned Female at Birth (reference) | aPR (95% CI) | Transgender | Not Transgender (reference) | aPR (95% CI) | Donť Know | |
| n (%) | n (%) | n (%) | n (%) | n (%) | |||
|
|
|
||||||
| N | 116 | 534 | 300 | 289 | 71 | ||
| Age | |||||||
| <18 | 4 (3.4) | 49 (9.2) | 1.08 (1.04, 1.12) | 22 (7.3) | 19 (6.6) | 0.85 (0.82, 0.87) | 12 (16.9) |
| 18–24 | 7 (6.0) | 88 (16.5) | 56 (18.7) | 33 (11.4) | 11 (15.5) | ||
| 25–34 | 57 (49.1) | 214 (40.1) | 147 (49.0) | 105 (36.3) | 23 (32.4) | ||
| 35–44 | 19 (16.4) | 96 (18.0) | 42 (14.0) | 59 (20.4) | 15 (21.1) | ||
| 45+ | 20 (17.2) | 66 (12.4) | 22 (7.3) | 57 (19.7) | 7 (9.9) | ||
| Race/ethnicity | |||||||
| Asian | 10 (8.6) | 55 (10.3) | 1.04 (0.53, 2.05) | 29 (9.7) | 29 (10.0) | 0.81 (0.48, 1.37) | 7 (9.9) |
| Black | 3 (2.6) | 17 (3.2) | 1.05 (0.30, 3.65) | 6 (2.0) | 9 (3.1) | 0.44 (0.15, 1.34) | 5 (7.0) |
| Latinx | 24 (20.7) | 40 (7.5) | 2.93 (1.73, 4.98) | 27 (9.0) | 32 (11.1) | 0.71 (0.41, 1.22) | 7 (9.9) |
| Native American/Alaska Native | 6 (5.2) | 39 (7.3) | 0.79 (0.33, 1.87) | 18 (6.0) | 25 (8.7) | 0.72 (0.38, 1.38) | 3 (4.2) |
| Native Hawaiian/Pacific Islander | 7 (6.0) | 9 (1.7) | 3.84 (1.42, 10.38) | 6 (2.0) | 10 (3.5) | 0.53 (0.19, 1.50) | 1 (1.4) |
| White | 92 (79.3) | 444 (83.1) | 0.96 (0.76, 1.21) | 259 (86.3) | 230 (79.6) | 1.10 (0.91, 1.33) | 55 (77.5) |
| Homelessness in the past year1 | 4 (3.4) | 25 (4.7) | 0.83 (0.29, 2.41) | 19 (6.4) | 9 (3.1) | 1.77 (0.79, 3.96) | 2 (2.8) |
| Food insecurity in the past year1 | 28 (32.9) | 129 (32.7) | 0.99 (0.65, 1.52) | 92 (37.9) | 57 (28.6) | 1.29 (0.91, 1.84) | 13 (30.2) |
| Has health insurance | 110 (95.7) | 499 (94.2) | 1.02 (0.82, 1.26) | 283 (95.0) | 267 (93.0) | 1.02 (0.86, 1.22) | 67 (95.7) |
| Gender affirming health care1 | |||||||
| Wants/needs gender affirming medical procedures | 38 (46.3) | 199 (51.2) | 0.98 (0.68, 1.40) | 188 (78.0) | 39 (20.1) | 3.54 (2.48, 5.07) | 15 (35.7) |
| Experienced barriers to accessing gender affirming care, past year2 | 25 (64.1) | 147 (73.9) | 0.92 (0.59, 1.43) | 145 (77.1) | 20 (50.0) | 1.47 (0.92, 2.35) | 11 (73.3) |
| Ever use hormones | 17 (20.7) | 89 (22.8) | 1.00 (0.59, 1.69) | 97 (39.9) | 11 (5.7) | 7.81 (4.01, 15.2) | 1 (2.4) |
| Prefers to receive care at a clinic that specializes in trans health | 44 (37.9) | 211 (39.5) | 1.02 (0.73, 1.43) | 190 (63.3) | 54 (18.7) | 3.40 (2.46, 4.69) | 15 (21.1) |
| Lifetime Experiences of Discrimination1 | |||||||
| Misgendered in a health care setting | 38 (32.8) | 205 (38.4) | 0.94 (0.66, 1.34) | 181 (60.3) | 50 (17.3) | 3.41 (2.46, 4.75) | 15 (21.1) |
| Dead named in a health care setting | 20 (17.2) | 133 (24.9) | 0.80 (0.50, 1.29) | 123 (41.0) | 26 (9.0) | 4.00 (2.59, 6.16) | 8 (11.3) |
| Denied health care | 20 (17.2) | 126 (23.6) | 0.78 (0.49, 1.25) | 87 (29.0) | 52 (18.0) | 1.57 (1.10, 2.25) | 11 (15.5) |
| Workplace discrimination3 | 36 (31.0) | 179 (33.5) | 0.92 (0.64, 1.33) | 109 (36.3) | 89 (30.8) | 1.38 (1.02, 1.87) | 20 (28.2) |
| Workplace microaggression4 | 58 (50.0) | 296 (55.4) | 0.94 (0.71, 1.26) | 186 (62.0) | 146 (50.5) | 1.25 (0.99, 1.57) | 28 (39.4) |
| Denied service in a place of public accommodation5 | 34 (29.3) | 147 (27.5) | 0.99 (0.67, 1.46) | 96 (32.0) | 76 (26.3) | 1.42 (1.03, 1.95) | 12 (16.9) |
| Verbal harassment or physical violence6 | 45 (38.8) | 269 (50.4) | 0.79 (0.57, 1.1) | 159 (53.0) | 132 (45.7) | 1.19 (0.93, 1.52) | 28 (39.4) |
| Sexual Health | |||||||
| Received HPV vaccine | 33 (28.4) | 211 (39.5) | 0.81 (0.55, 1.19) | 131 (43.7) | 100 (34.6) | 1.04 (0.80, 1.37) | 18 (25.4) |
| HIV testing, past year | 41 (46.6) | 69 (24.8) | 1.83 (1.23, 2.72) | 52 (31.5) | 45 (26.5) | 1.12 (0.74, 1.70) | 14 (40.0) |
| Currently using PrEP | 11 (13.9) | 3 (1.1) | 12.9 (3.59, 46.9) | 4 (2.3) | 8 (5.3) | 0.42 (0.12, 1.42) | 2 (5.9) |
| Diagnosed with STI, past year | 4 (15.4) | 1 (1.1) | 23.4 (2.12, 258.9) | 2 (3.6) | 2 (4.1) | 0.72 (0.10, 5.18) | 1 (8.3) |
| Living with HIV | 4 (4.3) | 1 (0.3) | 3.36 (2.83, 3.98) | 4 (2.2) | 0 (0.0) | 4.38 (3.48, 5.51) | 1 (2.6) |
Adjusted prevalence ratios (aPR) compare gender diverse participants assigned male at birth to those assigned female at birth, and compares gender diverse participants who self-identify as transgender compared to those who do not self-identify as transgender, all adjusted for age. There were 3 gender diverse participants who were intersex and 7 gender diverse participants who responded “don’t know” to the sex assigned at birth question.
The variables were ascertain in the optional supplement that followed the core 2021 Pride Survey. Overall, 484 (73%) of gender diverse participants completed these questions
Among participants who want/need gender affirming medical procedures
This question asked “Have you ever been red, denied a job or promotion, or received a negative evaluation for any reason?”
This question asks “Have you ever experienced microaggressions in the work place, for example, have you been overlooked, avoided, left out of conversations/emails/meetings, treated disrespectfully, been treated as inferior, or had your ideas or opinions minimized or ignored?”
This question asks “Have you ever been denied equal treatment or service in a place of business, government agency, or public place (e.g. a restaurant, hotel, public transportation) for any reason?”
This question asked “Have you ever been hit, physically attacked, sexually assaulted, threatened with violence, verbally insulted/abused?”
Stratification by sex assigned at birth revealed differences across all sexual health outcomes. Gender diverse participants AMAB were more likely to have tested for HIV in the past year (47% v. 25%; aPR 1.83, 95%CI 1.23, 2.72), currently use PrEP (14% v. 1%; aPR 12.9, 95%CI 3.59, 46.9) have been diagnosed with an STI in the past year (15% v. 1%; aPR 23.4, 95%CI 2.12, 258.9), or be currently living with HIV (4% v. 0.3%; aPR 3.36, 95%CI 2.83, 3.98). There were no differences in gender affirming care or other experiences of discrimination when stratifying by sex assigned at birth.
Stratification by transgender self-identification revealed significant differences in access to gender-affirming care and lifetime experiences of discrimination. A higher proportion of those who identified as transgender desired gender-affirming medical care (78% v. 20%; aPR 3.54, 95%CI 2.48, 5.07), preferred to receive care at a clinic that specializes in trans care (63% v. 19%; aPR 3.40, 95%CI 2.46, 4.69), or had ever used hormones (40% v. 6%; aPR 7.81, 95%CI 4.01, 15.2). Transgender-identified gender diverse participants were also significantly more likely to report being misgendered (60% v. 17%; aPR 3.41, 95%CI 2.46, 4.75) or deadnamed (41% v. 9%; aPR 4.00, 95%CI 2.59, 6.16) in a health care setting or ever being denied health care (29% v. 18%; aPR 1.57, 95%CI 1.10, 2.25). They were also more likely to report lifetime experiences of workplace discrimination (36% v. 31%; aPR 1.38, 95%CI 1.02, 1.87) and discrimination in a place of public accommodation (32% v. 26%; aPR 1.42, 95%CI 1.03, 1.95) compared to gender diverse participants who did not identify as transgender. Lastly, transgender-identified gender diverse individuals were more likely to be living with HIV (2% v. 0%; aPR 4.38, 95%CI 3.48, 5.51).
Stratification by gender identity revealed minimal differences between individual who were gender nonconforming, genderqueer, and nonbinary (Table S3). However, there were differences between two-spirit participants and other gender diverse individuals. Namely, two-spirit participants were more likely to be older (39% age 45+), and were most likely to experience homelessness (9%) and food insecurity in the past year (64%). Two-spirit participants were also less likely to have received HPV vaccination (12%) and most likely to be living with HIV (7%).
DISCUSSION
Given emerging concerns regarding the potential harms of asking about sex assigned at birth,16,18,19 this study examined the reliability and utility of an alternative approach to measuring gender identity in order to inform transgender health equity research. In our study of LGBTQ adults and adolescents, an alternative two-step method that asks about gender identity and transgender self-identification performed almost identically to the traditional two-step method that asks about gender identity and sex assigned at birth in its ability to identify cisgender and transgender and gender diverse study participants.
Our study also adds to the growing body of evidence that gender diverse people represent a large population, and that as many as 11–30% of all LGBTQ+ people identify as gender diverse.21,24 Our results highlight significant heterogeneity within this population. Most gender diverse participants endorsed multiple current gender identities, including overlap with binary cisgender and transgender identities. Consistent with prior research, we also observed that fewer than half (45%) of gender diverse participants identify as transgender.21 Although stratifying gender diverse participants by sex assigned at birth revealed large differences in sexual health outcomes, stratifying by transgender self-identification revealed significant differences in access to gender-affirming care and experiences of discrimination. Notably, gender diverse participants who self-identify as transgender were more likely to desire gender affirming medical interventions and prefer to receive care at clinics specialized in transgender health compared to gender diverse participants who did not identify as transgender. Transgender-identified gender diverse participants were also more likely to report lifetime experiences of discrimination and mistreatment in medical settings, at work, and in places of public accommodation. These findings suggest that transgender self-identification, not just current gender identity, may be an important factor for understanding how gender diverse people differ with respect to their identities, burdens of disease, barriers and facilitators of health care, and other socio-structural factors that impact their health and wellness.
In 2022, the National Academies of Sciences, Engineering, and Medicine (NASEM) conducted a review of existing measures of sex, gender, and sexual orientation and published recommendations and guiding principles for measuring these constucts.25 These recommendations have been critiqued26,27 as they do not fully reflect current best practices for measuring gender or the well-established preferences of LGBTQ+ communities.15,18,20,22,28 Our analysis offers empirical insight into the limitations of the current gender identity question recommended by NASEM. Most notable is that the report’s template question excludes most gender diverse response options (recommended response options: male, female, transgender, and two-spirit) and conflates concepts (namely, that transgender is a gender modality and not a gender identity). Importantly, our study found that the gender diverse population is sizable and that not all gender diverse participants identify as transgender.29 This highlights the critical need for research that assesses the reliability and validity of culturally-competent and more broadly inclusive approaches for measuring gender identity.
Strengths and Limitations
Our study was strengthened by the large number of number of participants who identified as gender non-conforming, genderqueer, and nonbinary, which allowed us to conduct detailed subgroup analyses. Another strength of our study was the use of a gender identity question with a check-all-that-apply format and many gender diverse response options. As a result, nearly all participants (98%) selected a response from our provided list, which increased our ability to accurately represent their identities. Most transgender and gender diverse participants selected more than one current gender identity: 42% of transgender men and 29% of transgender women also identifies as nonbinary or genderqueer, and >70% of gender diverse participants selected 2 or more nonbinary identities (e.g., both nonbinary and genderqueer).
Our findings should be interpreted in light of several limitations. Our study population is a convenience sample, predominantly White, and largely resided in the metropolitan areas of Washington State. Encouragingly, we observed that there was perfect agreement between the two measures among racial and ethnic minority participants. However, further validation of these measures in more geographically and racially diverse samples is warranted. The availability of large, diverse, population-based surveys that comprehensively assess gender identity is a known limitation to transgender health research.30 In the context of these systemic barriers to conducting trans health research, community-based non-probability samples are critical for developing trans-inclusive research methodologies that can be available for use and validation in future epidemiological and population health studies.
In addition, our question for sex assigned at birth included intersex as a response option. This does not reflect current recommendations or best practices, which recommend asking a separate question that ascertains if individuals identify as intersex or have intersex traits.25 Lastly, because our study only included people who self-identified as LGBTQ people, there remains a need to assess the performance of the alternative two-step approach within cisgender and heterosexual populations.
PUBLIC HEALTH IMPLICATIONS
This study has several important implications for research and public health practice. First, given that the two different two-step approaches for measuring gender performed nearly identically, researchers, clinicians, and public health practitioners should carefully consider the relevance and acceptability of sex assigned at birth and transgender self-identification questions for their specific contexts.20 Transgender self-identification was more useful than sex assigned at birth for identifying the healthcare needs and social determinants of health among transgender and gender diverse individuals. However, since most institutions recommend using a two-step question that asks about current gender and sex assigned at birth, using this approach may facilitate comparison across studies, including meta-analysis. Although sex assigned at birth is useful for identifying differences in sexual and reproductive health epidemiology among transgender and gender diverse populations, we caution against the inappropriately using sex assigned at birth as a proxy for current anatomy or biological processes. We recommend instead asking about specific anatomical structures (e.g., presence of a uterus) or biological processes (e.g., menstruation, ability to produce sperm, hormonal milieu) that are relevant to specific research questions. Given these complex considerations, it may be beneficial to use a multidimensional measure that asks about current gender, sex assigned at birth, and transgender self-identification. Community-engaged and transgender-led research are critical approaches to determining the acceptability of these different methodologies for specific research, clinical, and public health applications. In addition, gender diverse options should always be included when ascertaining gender, preferably allowing participants to select multiple gender identity options. This is critical for accurately measuring gender, representing study participants identities, and identifying and improving transgender health disparities.
Supplementary Material
Acknowledgements:
The Seattle Trans and Non-binary Sexual Health (STARS) Advisory Board includes the following individuals: Aleks Martin, Atlas Fernandez, Bennie Gross, Billy Caracciolo, Brian Minalga, Nicole Ó Catháin, Sayen Lentini, and William B. Heberling, as well as those who wish to remain anonymous. Data were presented at as oral abstracts at the 2023 Society for Epidemiologic Research annual meeting (Portland, OR) and the 2023 National Transgender Health Summit (San Francisco, CA).
Funding:
This research was funded by the NIH National Institute of Allergy and Infectious Diseases grant number F31AI152542 awarded to DMT. This work was also supported by the American Sexually Transmitted Diseases Association and the Northwest Center for Public Health Practice at the University of Washington’s School of Public Health.
Footnotes
Conflict of Interest: The authors have no conflicts of interest.
Disclaimer: The funders had no role in the conceptualization, analysis, or presentation of findings of this study. The contents of this paper are solely the responsibility of the authors and do not necessarily represent the official views of the funders.
Clinical trial registration number and website: N/A
Joint Authorship: N/A
Data Availability Statement:
Data access is provided and managed at the discretion of Public Health—Seattle & King County; visit https://kingcounty.gov/depts/health/data/data-request-service.aspx for more information regarding data inquiries. Computing code can be made available by request of the corresponding author.
REFERENCES
- 1.Reisner SL, Deutsch MB, Bhasin S, et al. Advancing methods for US transgender health research. Current opinion in endocrinology, diabetes, and obesity. 2016;23(2):198–207. doi: 10.1097/MED.0000000000000229 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Reisner SL, Radix A, Deutsch MB. Integrated and Gender-Affirming Transgender Clinical Care and Research. Journal of acquired immune deficiency syndromes (1999). 2016;72 Suppl 3:S235–42. doi: 10.1097/QAI.0000000000001088 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Bauer GR, Hammond R, Travers R, Kaay M, Hohenadel KM, Boyce M. “I Don’t Think This Is Theoretical; This Is Our Lives”: How Erasure Impacts Health Care for Transgender People. Journal of the Association of Nurses in AIDS Care. 2009;20(5):348–361. doi: 10.1016/j.jana.2009.07.004 [DOI] [PubMed] [Google Scholar]
- 4.Wesp LM, Malcoe LH, Elliott A, Poteat T. Intersectionality Research for Transgender Health Justice: A Theory-Driven Conceptual Framework for Structural Analysis of Transgender Health Inequities. Transgender Health. 2019;4(1):287–296. doi: 10.1089/trgh.2019.0039 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Tordoff DM, Minalga B, Gross BB, et al. Erasure and Health Equity Implications of Using Binary Male/Female Categories in Sexual Health Research and Human Immunodeficiency Virus/Sexually Transmitted Infection Surveillance: Recommendations for Transgender-Inclusive Data Collection and Reporting. Sexually Transmitted Diseases. 2022;49(2). doi: 10.1097/OLQ.0000000000001533 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Division of AIDS Cross-Network Transgender Working Group. Recommended Approach to Data Collection: The Two-Step Method. NIH National Institute of Allergy and Infectious Diseases; 2020. Accessed January 22, 2022. https://daidslearningportal.niaid.nih.gov/pluginfile.php/1/local_pages/pagecontent/Guidance on Use of Two-Step Method.docx [Google Scholar]
- 7.National Academies of Sciences Engineering and Medicine. Understanding the Well-Being of LGBTQI+ Populations. Published online 2020. doi: 10.17226/25877 [DOI] [Google Scholar]
- 8.National Institutes of Health Sexual and Gender Minority Research Office. Strategic Plan to Advance Research on the Health and Well-Being of Sexual & Gender Minorities.; 2020. [Google Scholar]
- 9.The GenIUSS Group. Best Practices for Asking Questions to Idenfity Transgender and Other Gender Minority Respondents on Population-Based Surveys; 2014. [Google Scholar]
- 10.Tate CC, Ledbetter JN, Youssef CP. A Two-Question Method for Assessing Gender Categories in the Social and Medical Sciences. Journal of Sex Research. 2013;50(8):767–776. doi: 10.1080/00224499.2012.690110 [DOI] [PubMed] [Google Scholar]
- 11.Bauer GR, Braimoh J, Scheim AI, Dharma C. Transgender-inclusive measures of sex/gender for population surveys: Mixed-methods evaluation and recommendations. Dalby AR, ed. PLOS ONE. 2017;12(5):e0178043. doi: 10.1371/journal.pone.0178043 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.KM K, GM S, SD R, et al. Operationalizing and analyzing 2-step gender identity questions: Methodological and ethical considerations. Journal of the American Medical Informatics Association : JAMIA. Published online September 2, 2021. doi: 10.1093/JAMIA/OCAB137 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Reisner SL, Conron KJ, Tardiff LA, Jarvi S, Gordon AR, Austin SB. Monitoring the health of transgender and other gender minority populations: Validity of natal sex and gender identity survey items in a U.S. National cohort of young adults. BMC Public Health. 2014;14(1):1–10. doi: 10.1186/1471-2458-14-1224/TABLES/4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.O’Brien RP, Blosnich JR. Refusal Rates to Sexual Orientation and Gender Identity Items in the Behavioral Risk Factor Surveillance System, 2014–2019. Am J Public Health. 2022;112(3):443–452. doi: 10.2105/AJPH.2021.306625 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Tordoff DM, Morgan J, Dombrowski JC, Golden MR, Barbee LA. Increased Ascertainment of Transgender and Non-binary Patients Using a 2-Step Versus 1-Step Gender Identity Intake Question in an STD Clinic Setting. Sexually Transmitted Diseases. 2019;46(4). doi: 10.1097/OLQ.0000000000000952 [DOI] [PubMed] [Google Scholar]
- 16.Alpert AB, Ruddick R, Manzano C. Rethinking sex-assigned-at-birth questions. BMJ. 2021;373:n1261. doi: 10.1136/bmj.n1261 [DOI] [PubMed] [Google Scholar]
- 17.Shteyler VM, Clarke JA, Adashi EY. Failed Assignments — Rethinking Sex Designations on Birth Certificates. New England Journal of Medicine. 2020;383(25):2399–2401. doi: 10.1056/nejmp2025974 [DOI] [PubMed] [Google Scholar]
- 18.Puckett JA, Brown NC, Dunn T, Mustanski B, Newcomb ME. Perspectives from Transgender and Gender Diverse People on How to Ask About Gender. LGBT Health. Published online June 26, 2020:lgbt.2019.0295. doi: 10.1089/lgbt.2019.0295 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Cahill S, Singal R, Grasso C, et al. Do Ask, Do Tell: High Levels of Acceptability by Patients of Routine Collection of Sexual Orientation and Gender Identity Data in Four Diverse American Community Health Centers. PLoS ONE. 2014;9(9). doi: 10.1371/JOURNAL.PONE.0107104 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Suen LW, Lunn MR, Sevelius JM, et al. Do Ask, Tell, and Show: Contextual Factors Affecting Sexual Orientation and Gender Identity Disclosure for Sexual and Gender Minority People. https://home.liebertpub.com/lgbt. 2022;9(2):73–80. doi: 10.1089/LGBT.2021.0159 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Wilson BD, Meyer IH. Nonbinary LGBTQ Adults in the United States. The Williams Institute; 2021. doi: 10.1037/sgd0000384 [DOI] [Google Scholar]
- 22.Kronk CA, Everhart AR, Ashley F, et al. Transgender data collection in the electronic health record: Current concepts and issues. Journal of the American Medical Informatics Association. 2022;29(2):271–284. doi: 10.1093/JAMIA/OCAB136 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Ashley F ‘Trans’ Is My Gender Modality: A Modest Terminological Proposal. In: Trans Bodies, Trans Selves. 2nd ed. Oxford University Press; 2022. [Google Scholar]
- 24.Clark KD, Luong S, Lunn MR, et al. Healthcare Mistreatment, State-Level Policy Protections, and Healthcare Avoidance Among Gender Minority People. Sexuality Research and Social Policy. Published online July 15, 2022:1–14. doi: 10.1007/S13178-022-00748-1/FIGURES/3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.National Academies of Sciences Engineering and Medicine. Measuring Sex, Gender Identity, and Sexual Orientation. The National Academies Press; 2022. 10.17226/26424 [DOI] [PubMed] [Google Scholar]
- 26.Logie CH, van der Merwe LLA, Scheim A. Measuring sex, gender, and sexual orientation: one step to health equity. The Lancet. 2022;400(10354):715–717. doi: 10.1016/S0140-6736(22)01425-8 [DOI] [PubMed] [Google Scholar]
- 27.National Academies of Sciences, Engineering, and Medicine. Consensus Study Report, Issue Brief: Measuring Sex, Gender Identity, and Sexual Orientation Research Recommendations.; 2022. https://nap.nationalacademies.org/resource/26424/Issue_Brief_Research_Recs.pdf [PubMed] [Google Scholar]
- 28.Broussard KA, Warner RH, Pope ARD. Too Many Boxes, or Not Enough? Preferences for How We Ask About Gender in Cisgender, LGB, and Gender-Diverse Samples. Sex Roles. 2018;78(9–10):606–624. doi: 10.1007/S11199-017-0823-2/TABLES/4 [DOI] [Google Scholar]
- 29.James SE, Herman JL, Rankin S, Keisling M, Mottet L, Anafi M. The Report of the 2015 U.S. Transgender Survey. National Center for Transgender Equality; 2016. [Google Scholar]
- 30.Lett E, Everhart A. Considerations for transgender population health research based on US national surveys. Annals of Epidemiology. 2022;65:65–71. doi: 10.1016/J.ANNEPIDEM.2021.10.009 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data access is provided and managed at the discretion of Public Health—Seattle & King County; visit https://kingcounty.gov/depts/health/data/data-request-service.aspx for more information regarding data inquiries. Computing code can be made available by request of the corresponding author.