Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Oct 29.
Published before final editing as: Psychol Sex Orientat Gend Divers. 2026 Oct 22:10.1037/sgd0000965. doi: 10.1037/sgd0000965

Gender Minority Stress and Resilience Factors Associated With Persistent Depressive Symptoms Among Transgender and Gender Diverse Adolescents and Young Adults Accessing Gender-Affirming Medical Care

Bridgid Mariko Conn 1,2,5, Ramón Durazo-Arvizu 3,4, Carolyn F Wong 1,2, Amy E West 1,5, Diane Chen 6,7,8,9, Yee-Ming Chan 10,11, Diane Ehrensaft 12,13, Robert Garofalo 6,8,14, Marco A Hidalgo 15, Stephen M Rosenthal 12,13, Johanna Olson-Kennedy 1,2
PMCID: PMC13714780  NIHMSID: NIHMS2208730  PMID: 42953113

Abstract

While many transgender and gender diverse (TGD) youth experience improved mental health after initiating gender-affirming hormone therapy (GAHT), some continue to experience persistent depressive (PD) symptoms. We examined associations of risk factors (e.g., negative expectations for the future, internalized transphobia) and protective factors (e.g. self-efficacy) for PD among TGD youth accessing GAHT. We used data from the Trans Youth Care U.S. Study, a national longitudinal prospective cohort study of TGD youth ages 12-20 years (N=315 at baseline). Experiencing PD, gender minority stress, resilience, and other factors were assessed at four timepoints (baseline, 6, 12, 18, and 24-months). After identifying two groups based on reported PD, we estimated differences at baseline and change over time of risk and protective factors using linear mixed-effects models. We also examined the association between these rates and experiencing PD using logistic regression. Two groups of PD youth (n=62) and non-PD youth (non-PD; n=253) were identified. Greater negative expectation at baseline was associated with increased odds of PD, while the opposite was observed for greater self-efficacy. Increased rate of change in body esteem was associated with decreased odds of PD. There was an interaction effect between negative expectation and self-efficacy. While self-efficacy was a protective factor, this effect lost its strength as perceived negative expectation increased. These findings identify important areas of assessment and intervention for TGD youth who experience PD despite accessing GAHT. Addressing these factors in treatment appears vital to ensuring adequate treatment for PD.

Keywords: transgender, adolescent and young adults, minority stress, self-efficacy, persistent depression

Introduction

Transgender and gender diverse (TGD) adolescents and young adults report significantly greater depression and associated suicidality and self-harm behavior compared to their cisgender peers (Aboussouan et al., 2019; Gibson et al., 2021; Lam et al., 2021; Lefevor et al., 2019; Pellicane & Ciesla, 2022). While access to gender-affirming medical care, including gender-affirming hormone therapy (GAHT; i.e., testosterone and estradiol used to medically support the development of gender-congruent secondary sex characteristics) has been found to lead to improvement in mental health overall, some TGD youth continue to report significant mental health challenges that are chronic and severe and negatively impact their functioning (Chen et al., 2023). Understanding the predictors of persistent depression among these youth is imperative as prior research with cisgender adolescents has found that ongoing depression has significant implications later in adulthood including, functional impairment, suicidality and self-harm, ongoing mental health difficulties, and not being engaged in education/training or employment (Weavers et al., 2021). Given the important implication for later well-being, it is of utmost importance to identify intervenable mechanisms, or factors that may alter mental health outcomes if targeted for treatment or structural change and reduce the likelihood of persistent depression (Hankin & Griffith, 2023; Morales-Muñoz et al., 2023).

The Gender Minority Stress (GMS) Model (Hendricks & Testa, 2012) provides a crucial socio-ecological framework for how chronic stress and invalidation contributes to significant mental health disparities for TGD individuals. Specifically, the GMS model posits that both proximal (i.e., factors internal or closer to the individual) and distal stressors (e.g., factors external to the individual, including in one’s environment or cultural context) contribute to mental health challenges for TGD youth, including increased depression, anxiety, suicidality, and trauma symptoms (Ching et al., 2024; Chodzen et al., 2019; Dolezal et al., 2023; Hammack et al., 2024; Lefevor et al., 2019; Parchem et al., 2025; Price-Feeney et al., 2020; Trujillo et al., 2017). Specifically, chronic experiences of discrimination, gender-based invalidation, interpersonal rejection, stigma, and associated internalized transphobia are seen as exacerbating mental health issues specifically for TGD youth and may have a greater impact for youth with a vulnerability to developing depressive symptoms (Turnamian & Liu, 2023; Wiglesworth, 2024).

The GMS framework also includes intrapersonal and social resilience factors, which can buffer against the substantial toll of minority stress on mental health outcomes in TGD youth (Hendricks & Testa, 2012). Self-efficacy, the self-perception of one’s ability to cope with and overcome adversity, has been frequently cited in the literature as a significant source of resiliency (Salifu Yendork & Somhlaba, 2015). This factor represents a potential intervenable mechanism by which mental health providers may improve outcomes through increasing problem-solving skills and sense of self-confidence and self-reliance.

The GMS model has continued to be expanded to reflect additional potential factors and mechanisms associated with positive mental and physical health outcomes for TGD youth, such as parental acceptance and gender embodiment (i.e., appearance congruence and body esteem for TGD individuals; see Figure 1 for conceptual model adapted from (Conn et al., 2023; Coyne et al., 2020). Body image, or body esteem, is also a crucial influence on adolescents’ self-esteem, with prior literature highlighting that this factor tends to increase over time for many adolescents, though this research has been conducted predominantly with cisgender youth (Bornioli et al., 2021; Morin et al., 2011). However, research has suggested that body dissatisfaction among cisgender adolescents is associated with increased depression over time, particularly among cisgender girls (Rawana & Morgan, 2014). A recent study found that TGD youth reported higher rates of body dissatisfaction compared to cisgender youth, with such dissatisfaction being positively associated with gender dysphoria, poor peer relations, and hostility related to their TGD identity (Brecht et al., 2024). On the other hand, Kuper and colleagues observed that for youth accessing GAHT, body esteem increased over time within the first year of initiating GAHT (Kuper et al., 2020). A recent study also found that reductions in body dissatisfaction were associated with fewer symptoms of depression and better psychosocial functioning one-year post-initiating GAHT (Chelliah et al., 2024).

Figure 1.

Figure 1.

Open in a new tab

Adapted gender minority stress and resilience model (Conn et al, 2023; Coyne et al., 2020; Hendricks & Testa, 2012).

*Bolded factors considered within current analysis.

Despite growing awareness of the significant mental health needs of TGD youth, few studies have examined specific factors that might contribute to or ameliorate chronic depression, which has been linked to greater risk for suicidality and other deleterious health outcomes (Ribeiro et al., 2018). The current study addresses this gap in our knowledge by using data from a prospective cohort of 315 TGD youth over two years before and after initiating GAHT to examine: (1) the characteristics of the subset of TGD youth who report clinically significant persistent depressive symptoms, and, (2) associations between experiencing persistent depression and minority stressors as well as protective factors. We hypothesized that greater levels of minority stressors and lower levels of protective factors for TGD youth would be associated with greater odds of experiencing persistent depression.

Methods

Study design.

The Trans Youth Care U.S. study (TYCUS) is a longitudinal, multisite, observational study of TGD youth who planned to initiate medical treatment with gender-affirming hormones (i.e., testosterone or estradiol) for phenotypic gender transition (Olson-Kennedy et al., 2019). The aims of the TYCUS study include the longitudinal tracking of gender-affirming treatment outcomes of TGD young people. Participants were recruited from July 2016 through June 2019 at four pediatric gender care clinics: Pacific Southwest Clinic, Midwest Clinic, Pacific West Clinic, and Northeast Clinic (clinic names have been de-identified for anonymity and are labeled based on general geographic descriptors). Researchers received institutional review board approval from all study sites, and written informed consent/assent was obtained for all participants. Eligibility criteria included: (1) presence of gender dysphoria as determined by a clinician; (2) appropriate and ready to initiate GAH (i.e., estrogen or testosterone) as determined by the clinical team; (3) age 20 years or younger; (4) English proficiency; and (5) psychiatrically and cognitively able to give informed consent/assent for participation. At baseline, participants were adolescents and young adults (12 to 20-years-old) who were initiating GAH through their respective clinics. Surveys were scheduled to occur within close temporal proximity to the first medical visit to start GAH. Parental consent was obtained for minors (i.e., those under age 18) to participate in the study (n=240). Study visits were conducted at baseline, prior to starting GAHT, and at 6, 12, 18, and 24-months after GAHT initiation. Details on study procedures have been published, and the protocol is available online Chen et al., 2021; Olson-Kennedy et al., 2019). While prior analyses removed youth with inconsistent use of GAHT (Chen et al., 2023), the present analysis retains all participants who initiated GAHT as a review for reasons for pausing GAH were not related to changes in gender identity. The majority of participants were observed to resume GAH at later time point. Given prior analysis indicating the potential role of designated sex at birth in mental health for TGD youth, this variable was examined in our analyses as a potential covariate.

Measures

Demographic characteristics and covariates.

Participants self-reported demographic characteristics including designated sex at birth, age, experience of endogenous puberty, and race/ethnicity. Youth were described as either not experiencing endogenous puberty based on not experiencing extensive gender-incongruent puberty due to puberty-delaying medication in early puberty (Tanner stage 2 or 3) or having a later age at onset of endogenous puberty and receiving GAH in Tanner stage 3 (at 13 to 15 years of age). This aligns with the definition utilized in prior publications (Chen et al., 2023). Post-traumatic stress disorder (PTSD) was also included as a covariate given the significant association of prior trauma experiences and depression among young people (Vibhakar et al., 2019). PTSD symptoms were assessed using the child and adult Mini International Neuropsychiatric Interview (MINI), a brief psychodiagnostic interview for major psychiatric disorders (Sheehan et al., 1998). Participants who met diagnostic criteria for PTSD were coded 1 while those who did not were coded 0. See Table 1.

Table 1.

Characteristics of transgender and nonbinary youth experiencing non-persistent vs. persistent depressive symptoms in the 2 years after starting gender-affirming hormone treatment.

Characteristics Non-Persistent Depression (N=253 at baseline, 158 at 24 months) Persistent Depression (N=62 at baseline, 59 at 24 months)
Designated female at birth 168 (66.4%) 36 (58.1%)
Gender identity
 Transmasculine/male 155 (61.2%) 35 (56.5%)
 Transfeminine/female 81 (32.0%) 25 (40.3%)
 Non-binary/genderqueer/Other 17 (6.7%) 2 (3.2%)
Age at baseline 15.9 (1.8) 16.3 (2.0)
Had not experienced endogenous pubertyǂ 23 (9.1%) 1 (1.6%)**
Race/Ethnicity
  Asian/Pacific Islander 9 (3.6%) 1 (1.6%)
  Black or African American 7 (2.8%) 4 (6.5%)
  Non-White Hispanic/Latino 42 (16.6%) 8 (12.9%)
  White 149 (58.9%) 35 (56.5%)
  White Hispanic/Latino 18 (7.1%) 7 (11.3%)
  Other 1 (0.4%) 0 (0.0%)
  Multiracial/Multiethnic 26 (10.3%) 6 (9.7%)
  Not Reported 1 (0.4%) 0 (0.0%)
Study site
 Pacific Southwest Clinic 86 (34.0%) 31 (50.0%)**
 Midwest Clinic 72 (28.5%) 22 (35.5%)**
 Pacific West Clinic 60 (23.7%) 7 (2.8%)**
 Northeast Clinic 35 (13.8%) 2 (3.2%)**
Open in a new tab
*

p<.05

**

p<.01

ǂ

Defined as youth who did not experience extensive gender-incongruent puberty due to receiving puberty-delaying medication in early puberty (Tanner stage 2 or 3) or having a later age at onset of endogenous puberty and receiving GAH in Tanner stage 3 (at 13 to 15 years of age).

Persistent depressive symptoms.

The Depressive Symptoms Scale from the Youth Self Report (YSR), for those under 19, and its counterpart for participants 19 and older, the Adult Self Report (ASR), were used to assess depressive symptoms over the past 6 months. The Youth Self Report (Achenbach & Rescorla, 2004) is a 112-item measure for youth ages 11 to 18 years and the Adult Self Report (ASR) is a 121-item measure for adults ages 18 to 59 years assessing internalizing and externalizing symptoms. The DSM-oriented scales were developed to identify items consistent with specific DSM diagnostic categories (Achenbach, 2014), which were then used to construct DSM-oriented scales for those categories – first for the DSM-IV (American Psychiatric Association, 2013) and then for the DSM-5 (Achenbach, 2015; American Psychiatric Association, 2013). The DSM-oriented scales have been found to be more strongly associated with actual DSM-5 diagnoses, including Major Depressive Disorder (MDD) among youth compared to the anxious/depressed syndrome scales of the YSR (Jiang et al., 2023; van Lang et al., 2005). The YSR/ASR were administered annually at baseline, 12-months, and 24-months to assess depressive symptoms over the past 6-months. Of note, the YSR was administered to all participants for the first two waves of data collection and scored using the program software. The ASR was introduced for those 19 and older at 24-months. There was good internal reliability across all time-points for the YSR Depressive Symptoms Scale (McDonald’s ω=.84[95%CI:0.81; 0.86], ω=.85[95%CI:0.82; 0.87], ω=.85[95%CI:0.82; 0.88], respectively) and the ASR Depressive Symptoms Scale at 24-months (ω=.95[95%CI:0.78; 0.99]).

We defined persistent depression (PD) as having T-scores on the Depressive Symptoms scale above 64, which defines scores in the ‘clinically concerning range’, at baseline, 12-months, and 24-months, indicating a consistent and persistent report of depression symptoms in the clinically significant range over the two years of the study. Based on the research team’s clinical experiences and prior findings (Chen et al., 2023), youth in our clinic and within this study generally report reduced mental health symptoms after two years of GAHT associated with greater appearance congruence; thus, this subcohort was identified based on this two-year timeframe. Participants with scores in the clinically significant range at all 3 time-points were categorized as part of the persistent depression (PD) group (n=62 [18.7% of the baseline sample; 26.1% of the sample at 2 years]). All other youth were categorized in the non-PD group (n=256), i.e., youth who did not report depressive symptoms in the clinically significant range at any time as well as youth who reported such depressive symptoms at one or two of the time points but not all three. Thus, the outcome variable for PD was operationalized as a dichotomous variable (1=PD; 0=no PD). Those in the non-PD group who had two of the three timepoints in the clinical range were identified as those having baseline and 24-month data (n=80) and those with 12- and 24-month data (n=74). While examining different configurations of groupings based on various timepoints was explored, due to small sample size and our research question, these analyses were deemed outside the scope for our current study; however, future planned analyses will further explore the experiences of these participants. Figure 2 depicts a flow chart illustrating participant selection and inclusion in our PD/non-PD groups. Further, the use of the YSR/ASR was supported by examination of concurrent assessment of depression through the MINI and the Beck Depression Inventory-Second Edition (BDI-II;), a widely used screener of depression (Beck et al., 1996), which identified significant differences between PD and non-PD youth in depressive symptoms (see Table 2).

Figure 2.

Figure 2.

Open in a new tab

Flow diagram for participant inclusion in the present analyses.

Table 2.

Differences between key predictor variables at baseline, 6-months, 12-months, 18-months, and 24-months for those with and without persistent depression.

Variable Non-Persistent Depression (N=253 at baseline, 158 at 24 months)
M (SD)/N (%)		 Persistent Depression (N = 62 at baseline, 59 at 24 months)
M (SD)/N (%)
Gender Minority Stress Factor
Negative expectations for the future
Baseline 51.8 (23.0) 58.8 (23.9)*
6-months 52.6 (24.8) 57.8 (20.3)
12-months 52.7 (23.7) 59.8 (23.9)*
18-months 52.5 (22.6) 63.8 (26.5)**
24-months 50.4 (25.1) 62.7 (23.6)**
Resilience Factor
Self-Efficacy
Baseline 66.0 (16.5) 58.9 (19.0)**
6-months 68.9 (15.8) 59.3 (15.1)**
12-months 69.1 (17.4) 58.0 (18.6)**
18-months 71.1 (16.6) 61.1 (17.8)**
24-months 72.1 (16.7) 60.1 (18.4)**
Body Esteem
Baseline 34.15 (6.86) 27.70 (7.94)**
6-months 30.61 (7.26) 25.92 (8.36)**
12-months 32.24 (7.37) 27.21 (8.18)**
18-months 32.91 (6.62) 26.62 (9.02)**
24-months 34.15 (6.86) 27.70 (7.94)**
Mental Health Characteristics
YSR Depressive Problems Scale ⸸⸸
Baseline 63.3 (9.9) 74.7 (6.9)**
6-months 61.0 (8.9) 75.1 (7.4)**
12-months 61.5 (23.5) 53.1 (26.4)*
18-months 68.5 (20.8) 59.8 (25.4)*
24-months 72.9 (18.7) 61.9 (23.5)**
Beck Depression Inventory-II
Baseline 11.1 (8.7) 25.1 (11.5)**
6-months 10.4 (8.7) 22.1 (12.7)**
12-months 9.0 (8.8) 24.2 (11.4)**
18-months 9.3 (9.0) 24.8 (13.5)**
24-months 8.9 (8.4) 24.4 (14.0)**
Suicidal ideation past 6-months
Baseline 75 (29.6%) 33 (53.2%)*
12-months 39 (44.3%) 31 (62%)*
24-months 35 (22.2%) 26 (44.1%)*
Met PTSD criteria ⸸⸸⸸
Baselineǂ 4 (1.6%) 5 (8.1%)*
12-monthsǂ 6 (3.8%) 7 (12.7%)**
24-monthsǂ 4 (2.6%) 7 (13.5%)**
Open in a new tab
⸸

Past 6-months

⸸⸸

YSR - Youth Self-Report T-scores

⸸⸸⸸

based on the MINI

*

p<.05

**

p<.01

ǂ

could not determine for one participant

Gender Minority Stress and Resilience.

The Gender Minority Stress and Resilience Scale for Adolescents (GMSR-A) is a self-report measure assessing social stigma and psychosocial resilience associated with gender minority identity, including internalized transphobia and gender identity pride (Testa et al., 2015), which was adapted for adolescents (Hidalgo et al., 2019). For the first two years of the study, participants completed 6 of the 9 total scales of the GMSR-A. The other three were included after 36 months and will be examined in future analysis. In total, there were 55 items. The present study used four of six subscales: negative expectations for the future (e.g., “If I express my gender identity, I could be a victim of crime or violence”, “If I express my gender identity, most people would look down on me”, range=0-36, ω=0.91 [95%CI:0.89; 0.92]), internalized transphobia (e.g., “When I think about my gender identity or expression, I feel unhappy,” “I feel that my gender identity or expression is embarrassing”, range=0-32, ω=0.90 [95%CI:0.88; 0.92]), gender identity pride (“I am proud to be a person whose gender identity is different from my sex assigned at birth”, “I have no problem talking about my gender identity and gender history to almost anyone”, range=0-32, ω=0.88 [95% CI:0.85; 0.90]), and community connectedness (“I feel part of a community of people who share my gender identity”, “When interacting with members of the community that shares my gender identity, I feel like I belong”; range=0-20, ω=0.79 [95%CI:0.75; 0.83]). All items within each subscale are rated on a 5-point Likert scale (0=strongly disagree, 4=strongly agree). Items are summed with higher scores indicating greater minority stress or resilience, respectively.

Parental Support Scale – Youth Version.

Participants responded to four questions about general parental support, which were adapted from the Multidimensional Scale for Perceived Social Support – parent subscale (Zimet et al., 1988). The statements reflected general parental support (i.e., “I get the emotional help and support I need from my parent or parents”, “I can talk about my problems with my parent or parents”). Items are rated on a 5-point Likert scale (1=Strongly Disagree, 5=Strongly Agree) that were averaged with greater scores indicating greater perceived parental support (ω=0.89 [95%CI:0.87; 0.92]).

Self-efficacy.

Participants completed the 10-item Self-Efficacy subscale from the NIH Toolbox-Emotion Battery (NIHTB-EB; Hodes et al., 2013) as an indicator of internal resiliency and self-perceived capacity to overcome challenges (e.g., “I can manage to solve difficult problems if I try hard enough” and “If someone tries to keep me from getting what I want, I can find a way to get what I want”). Items are rated on a 5-point scale (1=Never, 5=Very Often) that are summed with greater scores indicating greater self-efficacy (ω=0.91 [95%CI:0.90; 0.93]).

Appearance congruence.

To assess appearance congruence, participants completed the 9-item appearance congruence subscale (“My outward appearance represents my gender identity”) from the Transgender Congruence Scale (TCS; Kozee et al., 2012). Each item was rated on a 5-point scale (1= Strongly Disagree, 5= Strongly Agree) and scores for all items averaged, with higher scores reflecting greater appearance congruence. This scale demonstrated good reliability at all timepoints (ω=0.86 [95%CI:0.84; 0.88]).

Body Esteem.

To assess body esteem, or the self-evaluation of one’s body or appearance, we used one of the subscales from the Body Esteem Scale for Adolescents and Adults (BESAA). Overall, this measure has been found to have good internal consistency, test-retest reliability, and construct validity in a sample of cisgender youth and adolescents (Cragun et al., 2013; Mendelson et al., 2001). The BESAA has three subscales: Body Esteem-Weight, Body Esteem-Appearance, and Body Esteem-Attribution. Our study utilized two of the three subscales--Attribution (4 items) and Appearance (13 items) during data collection. In the current analysis, the Appearance subscale was used to reflect general esteem or liking one’s body appearance. The Appearance subscale is comprised of 13 items asking youth to respond on a 4-point scale to indicate how frequently they agree (0=never; 1=strongly agree; 2=agree; 3= disagree; and 4= strongly disagree) with statements about their bodies (e.g., “I’m pretty happy about the way I look”). In an examination of the longitudinal association between self-esteem and body esteem, Mendelson and colleagues (1996) found that body esteem about one’s appearance was positively associated with self-esteem. This scale demonstrated generally good consistency at all time points (ω=0.93 [95%CI:0.91; 0.94]).

Statistical Analysis Plan.

All statistical analyses were conducted with SPSS version 29 (IBM), STATA 18.0 software (College Station, TX), and R version 4.4.1 (2024-06-14). Missing data were addressed using mean imputation for the GMSR variables with at least 75% data and to minimize missingness. Missingness was minimal across all variables. Models were run for participants with complete data (n=217 at 24 months). Several participants were considered lost to follow-up, and their data were not included in the longitudinal analysis (n=31) due to the method used to identify group membership as the outcome. Categorical variables were examined across groups using Fisher’s exact test and continuous variables using t-tests (See Tables 1 and 2). Significant differences in the descriptive characteristics (e.g., age, designated sex at birth, gender identity, race/ethnicity) between those who were lost to follow-up and those who remained in the study were not identified; thus, our final sample is considered representative of participants who enrolled in the study. Correlations for key variables were examined and are illustrated using a heat map (see Figure 3). In all models, designated sex at birth (female as the reference group) and appearance congruence were included as additional potential predictors based on prior analysis (e.g., Chen et al., 2023).

Figure 3.

Figure 3.

Open in a new tab

Heat map correlation table of variables for the entire sample at baseline.

Note: For the heat map in color: grey boxes indicate negative correlations, while the darker the pink/red color boxes, the stronger the positive correlation.

Eight potential factors were hypothesized as potentially predictive of PD: Negative expectations for the future, internalized transphobia, gender identity pride, appearance congruence, parental acceptance, body esteem, community connectedness, and self-efficacy. Designated sex at birth and meeting criteria for PTSD at any time during the study period were included initially as covariates. Descriptive statistics, including graphical and numerical statistics, were generated and presented by the PD status. The modeling strategy was divided into the following steps: (a) Graph mean levels and confidence intervals of each factor over time (baseline, 6, 12, 18, and 24 months); (b) For each study participant and factor, calculate the annual change of the factor as the conditional mean of the random slope given the vector of responses after regressing the factor on the time from baseline in years using a linear mixed-effects model (Fitzmaurice et al., 2011). Stepwise multivariable logistic regression was used to assess the association between persistent depression and rate of change adjusting for baseline levels of the factors.

We applied a stepwise multivariable logistic regression model adapted from the literature (Hosmer, Jr et al., 2013) to select the factors associated with persistent depression through 24 months using a backward selection algorithm. Initial stepwise regressions indicated the influence of a confounding relationship between variables, resulting in a posthoc exploratory examination of the interaction terms for all key predictor variables. Step 1: For each potential factor of interest, fit a logistic regression with PD as the outcome and the annual rate of change of the factor as the independent variable, adjusting for designated sex at birth, PTSD, and the baseline level of the factor. Select for further analyses only those variables associated with PD (p-value < 0.15; Hosmer, Jr et al., 2013) or those that did not reach this level of significance but were considered relevant to the analysis (i.e., designated sex at birth). Step 2: A backward variable selection algorithm was implemented with clinical input and a threshold p-value of 10%. The model included baseline levels of each covariate selected in step 1, designated sex at birth, and PTSD at any time during the study. The p-value was calculated using the likelihood ratio test statistic since two variables were tested simultaneously (the annual rate of change and the corresponding baseline level of the factor). Step 3: Once variables (annual rates of change) were identified, two-way interaction terms were entered into a multivariable logistic regression model to improve the model and address effect modification between two of the variables. All 2-way significant interactions were investigated (p<.10). Significant interaction terms were then retained in the model. Given the lack of statistically significant role of appearance congruence, this variable was not retained as a factor in the next step. Step 4: The model’s goodness-of-fit was tested using the Hosmer-Lemeshow chi-square statistic, and the model’s level of discrimination was assessed using the area under the ROC statistic. P-values were set higher than the traditional level (0.05) to reduce the likelihood of missing variables of importance (Hosmer, Jr et al., 2013). Some authors have suggested threshold values as large as 20% or higher (Dales & Ury, 1978; Mickey & Greenland, 1989). An additional criterion for the inclusion/removal of a variable was its impact on the estimated effect size (e.g., odds ratio) of another variable. Furthermore, clinical expertise and common sense were exercised to decide on the variables entering the final model. See the Supplemental Material for the technical data notation and formulas for the final model.

Positionality Statement.

The research team identities encompass a range of racial, ethnic, generational status, sexual orientation, and gender identity statuses. While our team is broad and includes several team members across study sites, for the purposes of this positionality statement, we are sharing the positionalities of the main author team. The first author (B.M.C.) is a second-generation Japanese-American and Jewish, queer, cisgender woman, daughter of immigrants, raised in the colonized islands of Hawai’i. She is a Board-Certified Child and Adolescent Psychologist and an Associate Professor of Clinical Pediatrics and Psychology within an academic medical center. The second author (R. D.-A.) is a Latino cisgender man, son of a farmer, born and raised in the middle of the Sonoran Desert and has higher education privilege. He is the Director of the Biostatistics Core at Children’s Hospital Los Angeles and a Professor of Research Pediatrics and Neurology. The third author (C.F.W.) is a Chinese-American immigrant from Hong Kong who came to the US at age 10. She is a heterosexual cisgender woman who is trained as a health psychologist who currently holds the position of Associate Professor of Research. The fourth author (A.E.W.) is a White and Indigenous (enrolled member of Cheynne & Arapaho Tribes) heterosexual, cisgender woman raised in Washington D.C. in a family with economic and educational privilege. She is a Professor of Clinical Pediatrics, Psychiatry & Behavioral Sciences, and Psychology and a Board-Certified Child and Adolescent Psychologist. The senior author (J.O.-K.) is a White, cisgender woman with higher educational privilege and a Professor of Clinical Pediatrics and a gender health physician. Several of our research team members have expertise in quantitative analysis, including longitudinal methods and most have been well-published in the area of transgender health within peer-reviewed journals. The motivation and impetus for carrying out this investigation was drawn from the clinical experiences and observation of the first and senior author and a sincere desire to understand the drivers of persisting mental health needs for youth accessing GAH. The research team members described above were the core team involved in developing and carrying out this quantitative analysis plan, including conceptualization, selecting variables, and interpreting the results.

Results

Analytic sample characteristics.

The sample identified as White non-Hispanic/Latiné (59%), White Hispanic/Latiné (8%), non-White Hispanic/Latiné (16%), multiracial (10%), Black/African American (4%), Asian/Pacific Islander (3%), and other (1%); see Table 1. Participants were transmasculine/male (60.3%), transfeminine/female (33.7%), non-binary (3.8%), genderqueer/gender fluid (1.3%), and other (1%). These responses were recoded into three categories: transmasculine/male (60.3%), transfeminine/female (33.7%), and non-binary (6.0%).

There were no significant differences between the PD and non-PD groups for age, self-reported race/ethnicity, and experience of endogenous puberty. We did observe differences between geographic location (i.e., the site at which they enrolled) across PD and non-PD groups, with a greater proportion of PD youth enrolled at Children’s Hospital Los Angeles and Lurie Children’s Hospital and a lower proportion of PD youth enrolled at Boston Children’s Hospital and UCSF Benioff Children’s Hospital, which may be associated with greater number of participants being enrolled at these sites. Individuals who endorsed PTSD diagnosis at any time during the 24-month study period had 3.23 times higher odds of PD compared to individuals who did not (OR=3.23, 95% CI:1.37,7.54, p=0.008). Designated sex at birth was not associated with PD group membership (OR=0.78, 95% CI:0.40,1.52, p=0.460). The PD group also had a higher proportion of participants reporting suicidal ideation in the past 6 months at baseline and at 24 months compared to the non-PD group. We observed some longitudinal differences between these groups. For example, for the PD group, negative expectations appeared to increase over time, while this remained relatively stable for the non-PD group. The PD group reported generally lower levels of depression at 24 months compared to baseline, while the proportion of those reporting suicidal ideation also decreased over time. For the non-PD group, we observed a general reduction also of recent suicidal ideation, as well as recent depression based on the BDI-II. However, for this group, there was a noted increase in depressive symptoms based on the YSR/ASR by 24-months. In terms of resilience, the PD group had smaller increases in self-efficacy over time, while there was a larger increase for the non-PD group. Table 2 illustrates the means of key variables at all time points within the PD and non-PD groups and differences between PD and non-PD groups on significant predictor variables.

The Hosmer-Lemeshow goodness-of-fit statistic with 10 groups was not statistically significant (p=0.229), indicating the model fit the data well, and the area under the ROC curve demonstrated good discrimination (Area=79.4%, 95% CI: 72.8%-86.1%). Results focus on the three key variables that were identified as significant predictors in the modeling and variable selection analysis (i.e., negative expectations for the future, self-efficacy, and body esteem; see Table 4).

Table 4.

Step 2. Statistically significant variables (p<0.10) and covariates in the multivariate model examining the impact of all factors on odds of persistent depression as determined by backward elimination.

Variables OR1 95% CI1 p
Negative Expectations for the Future – Baseline 1.07 1.01, 1.12 0.010
Negative Expectations for the Future – Annual Change 1.17 0.97, 1.42 0.094
Self-Efficacy – Baseline 0.93 0.88, 0.98 0.004
Self-Efficacy – Annual Change 0.78 0.60, 1.01 0.055
Internalized Transphobia – Baseline 1.00 0.95, 1.05 0.916
Internalized Transphobia – Annual Change 0.96 0.79, 1.17 0.708
Community Connectedness – Baseline 1.05 0.96, 1.15 0.273
Community Connectedness – Annual Change 1.02 0.71, 1.47 0.928
Parental Acceptance – Baseline 0.97 0.90, 1.06 0.517
Parental Acceptance – Annual Change 0.84 0.53, 1.34 0.467
Body Esteem – Baseline 0.96 0.90, 1.02 0.198
Body Esteem – Annual Change 0.70 0.55, 0.90 0.005
Appearance Congruence – Baseline 0.93 0.52, 1.69 0.821
Appearance Congruence – Annual Change 1.62 0.19, 14.5 0.661
Sex Designated at Birth 0.79 0.40, 1.57 0.492
Post-traumatic Stress Disorder Diagnosis 2.38 0.99, 5.56 0.052
Open in a new tab
1

OR = Odds Ratio, CI = Confidence Interval.

Note: Bolded variables highlight those factors retained due to significance or clinical importance.

Change in negative expectations, self-efficacy, and body esteem over 2 years.

Trend analyses examined the rate of change of our key predictor variables (i.e., negative expectations for the future, self-efficacy, body esteem) at baseline, 6 months, 12 months, 18 months, and 24 months for both the PD and non-PD groups. Figure 4a-c illustrates the mean change over time for body esteem, negative expectations, and self-efficacy between the PD and non-PD groups. Within the bivariate analyses (Step 1 of our analysis), the rates of change adjusted for baseline total scores were significantly associated with the odds of PD for negative expectations for the future (OR=1.25, 95% CI: 1.08-1.45, p=0.002), appearance congruence (OR=0.09, 95% CI:0.02-0.38,p=0.003), self-efficacy (OR=0.71, 95% CI:0.57-0.87,p=0.001), internalized transphobia (OR=1.19, 95% CI:1.03-1.38, p=0.032), parental acceptance (OR=0.63, 95% CI:0.44-0.91, p=0.013), and body esteem (OR=0.66, 95% CI:0.55-0.78, p=0.001) (See Table 3).

Figure 4.

Figure 4.

Open in a new tab

Body esteem (a), Negative expectations for the future (b), and self-efficacy (c) between persistent and non-persistent depression groups at baseline, 6, 12, 18, and 24-month visits.

Table 3.

Step 1. Bivariate models examining associations of all proposed factors (baseline values and annual change) on odds of persistent depression.

Variables OR1 95% CI1 p
Negative Expectations for the Future – Annual Change 1.25 1.08, 1.45 0.002
Negative Expectations for the Future – Baseline 1.07 1.03, 1.12 0.001
Self-Efficacy – Annual Change 0.71 0.57, 0.87 0.001
Self-Efficacy – Baseline 0.91 0.87, 0.95 0.001
Internalized Transphobia – Annual Change 1.19 1.03, 1.38 0.016
Internalized Transphobia – Baseline 1.04 1.00, 1.08 0.032
Gender Identity Pride – Annual Change 1.05 0.91, 1.21 0.483
Gender Identity Pride – Baseline 1.00 0.96, 1.03 0.831
Community Connectedness – Annual Change 0.76 0.57, 1.01 0.056
Community Connectedness – Baseline 1.01 0.94, 1.09 0.731
Parental Acceptance – Annual Change 0.63 0.44, 0.91 0.013
Parental Acceptance – Baseline 0.92 0.86, 0.98 0.006
Body Esteem – Annual Change 0.66 0.55, 0.78 0.001
Body Esteem – Baseline 0.92 0.88, 0.96 0.001
Appearance Congruence – Annual Change 0.09 0.02, 0.38 0.001
Appearance Congruence – Baseline 0.55 0.36, 0.81 0.003
Open in a new tab
1

OR = Odds Ratio, CI = Confidence Interval.

Note: All variables with p <0.15 were included in the multivariate model. Note: Bolded variables highlight those factors retained due to significance or clinical importance.

Multivariable Adjusted Analysis

All factors listed in Table 3 with p-value < 0.10 were selected to enter a multivariable logistic regression model (Step 2; Table 4) followed by an automatic backward elimination stepwise regression procedure (Step 3; Table 5). Risk factors with p-value < 0.10 were selected to confirm the “main effects” model. This resulted in the following three candidate factors, Negative Expectations for the Future – Annual Change, Self-Efficacy – Annual Change, and Body Esteem – Annual Change. In Step 4, we built a model with these variables, the corresponding baseline total scores, sex designated at birth, post-traumatic stress disorder diagnosis and the two-way interactions of the three main effects. Odds ratios, 95% confidence intervals and p-values for each of the variables included in this model are listed in Table 6. The Negative Expectations for the Future – Annual Change by Self-Efficacy – Annual Change interaction was statistically significant (p=0.009). Table 7 lists the OR estimates of PD compared to non-PD group membership and corresponding 95% confidence intervals for all factors in the final regression model, including annual change (and baseline levels) in negative expectations for the future, self-efficacy, body esteem, designated sex at birth, and post-traumatic stress disorder. Although designated sex at birth was not statistically associated with the odds of PD, it was retained in the model due to its clinical importance. In Tables 6–7, main effects for the key predictors were retained in models examining the interaction effects of those variables. Of note, a model examining a three-way interaction between the significant predictors was run and was not significant.

Table 5.

Step 3. The multivariate model of variables selected by stepwise regression with backward elimination.

Variables OR1 95% CI1 p
Negative Expectations for the Future – Baseline 1.06 1.01, 1.11 0.009
Negative Expectations for the Future – Annual Change 1.15 0.97, 1.37 0.105
Self-Efficacy – Baseline 0.92 0.88, 0.97 0.002
Self-Efficacy – Annual Change 0.78 0.62, 0.97 0.028
Body Esteem – Baseline 0.95 0.91, 1.00 0.061
Body Esteem – Annual Change 0.73 0.59, 0.89 0.002
Sex Designated at Birth 0.75 0.39, 1.45 0.394
Post-traumatic Stress Disorder Diagnosis 2.56 1.12, 5.76 0.027
Open in a new tab
1

OR = Odds Ratio, CI = Confidence Interval.

Note: Bolded variables highlight those factors retained due to significance or clinical importance.

Table 6.

Step 4. The multivariate model with variables selected by stepwise regression with change-over-time two-way interactions.

Characteristic OR1 95% CI1 p-value
Baseline Negative Expectations for the Future 1.06 1.01, 1.11 0.010
Annual Change Negative Expectations for the Future 1.23 1.01, 1.50 0.729
Baseline Self-Efficacy 0.91 0.87, 0.96 0.001
Annual Change Self-Efficacy 0.72 0.55, 0.92 0.011
Baseline Body Esteem 0.96 0.91, 1.01 0.086
Annual Change Body Esteem 0.67 0.53, 0.83 0.001
Sex Designated at Birth 0.77 0.39, 1.50 0.434
Post-traumatic Stress Disorder Diagnosis 3.29 1.39, 7.72 0.007
Negative Expectations for the Future – Annual Change × Body Esteem – Annual Change 1.04 0.97, 1.13 0.339
Self-Efficacy – Annual Change × Body Esteem – Annual Change 1.03 0.93, 1.14 0.608
Negative Expectations for the Future – Annual Change × Self-Efficacy – Annual Change 1.15 1.04, 1.28 0.009
Open in a new tab
1

OR = Odds Ratio, CI = Confidence Interval.

Note: Bolded variables highlight those factors retained due to significance or clinical importance. Main effects were retained when examining interaction effects.

Table 7.

Step 5. The final multivariate model with variables and two-way interactions selected by stepwise regression.

Variables OR1 95% CI1 p-value
Baseline Negative Expectations for the Future 1.06 1.02, 1.11 0.008
Negative Expectations for the Future – Annual Change 1.21 1.01, 1.46 0.581
Baseline Self-Efficacy 0.91 0.87, 0.96 0.001
Self-Efficacy – Annual Change 0.72 0.56, 0.91 0.002
Baseline Body Esteem 0.96 0.91, 1.01 0.098
Body Esteem – Annual Change 0.69 0.55, 0.84 0.001
Sex Designated at Birth 0.78 0.40, 1.52 0.460
Post-traumatic Stress Disorder Diagnosis 3.23 1.37, 7.54 0.008
Negative Expectations for the Future – Annual Change × Self-Efficacy – Annual Change 1.16 1.05, 1.28 0.003
Open in a new tab
1

OR = Odds Ratio, CI = Confidence Interval

Note: Bolded variables highlight those factors retained due to significance or clinical importance. Main effects were retained when examining interaction effects.

Negative expectations, self-efficacy, and body esteem at baseline and covariates.

Multivariate logistic regression analysis showed that a one-point increase in the baseline score for negative expectations for the future was associated with 6% increased odds of PD (adjusted odds ratio [AOR]= 1.06, 95% CI:1.02-1.11), p=.008). In contrast, the likelihood of PD decreased by 9% with each one-point increase in the score for greater self-efficacy at baseline (OR=0.91, 95% CI: 0.87-0.96, p= 0.001). Body esteem at baseline was not significantly associated with odds of PD (OR=0.96, 95% CI:0.91, 1.01, p = 0.098). See Table 7.

Associations between annual change in negative expectations and self-efficacy by PD/non-PD group.

There was a 31% decrease in the odds of PD for each one point/year increase in the annual change in body esteem score (AOR=0.69, 95%CI:0.55-0.84, p=0.001). The effect of the change in self-efficacy on the odds of PD was modified by the change in negative expectations for the future, as indicated by a significant interaction between these two factors (OR for interaction=1.16, 95%CI:1.05-1.28, p=0.003) as shown in Figure 5. Specifically, the protective effect of increasing self-efficacy on the likelihood of PD status was diminished for those with more rapid increases in negative expectations. If the rate of change in negative expectations was one or more point increase in the score per year, the protective effect of self-efficacy was erased, contributing to greater odds of PD. In other words, faster increase in negative expectations resulted in the loss of the protective effect of self-efficacy on odds of PD. As an illustration: for a hypothetical group of individuals with a −1-point annual rate of negative expectations, the odds of PD per one unit increase in the annual change of self-efficacy decreased by 40% (OR=0.60, 95% CI:0.44-0.81; Note: 0.60=0.69/1.16). Similarly, individuals with no change in negative expectations over 24 months, the effect of increasing self-efficacy was still protective (OR=0.69, 95% CI:0.54-0.88). For individuals whose negative expectation worsened by more than 1 unit per year, the protective effect of improving self-efficacy “disappeared”. For example, for annual changes in negative expectation or 1, 2, and 3 units per year, the OR of PD per unit increase in self-efficacy are 0.8 (95%CI:0.63-1.00), 0.92 (95%CI:0.92-1.18), and 1.07 (95% CI:0.79-1.44), respectively (see Figure 4). Thus, fast and positive increases in negative expectations negate the positive effect of improvements in self-efficacy.

Figure 5.

Figure 5.

Open in a new tab

Interaction between self-efficacy and negative expectations on the odds of PD. At a ~2-points/year rate of increase in negative expectation scores, the effects of negative expectations negated the protective effect of improvements in self-efficacy.

In other words, it appears that the negative effect of negative expectations may be a strong predictor of PD despite the positive effect of self-efficacy. It may also be that a significant positive change each year in negative expectations may negate the effect of increases in self-efficacy. Likewise, decreases in negative expectations may increase the positive effect of self-efficacy on reducing odds of PD.

Discussion

While prior research suggests that GAHT results in mental health improvement for many TGD youth, a subset report PD. This study identifies characteristics and factors that contribute to PD symptoms among TGD youth, building upon prior research examining mental health outcomes for TGD youth after two-years of accessing GAHT (e.g., (Chen et al., 2023; Olson-Kennedy et al., 2025). This is particularly crucial given the mental health disparities observed among TGD youth compared to their cisgender peers (Reisner et al., 2015), especially in the current socio-historical context of gender-affirming care, including increased anti-transgender legislation in the U.S. (American Civil Liberties Union, 2021; Lee et al., 2024). This analysis represents the crucial first step in developing a tool to identify TGD youth at highest risk for developing PD.

We identified a cohort of TGD youth who reported PD, that is, clinically significant depressive symptoms throughout the first two years of GAHT. We found that one key factor associated with an increased likelihood of PD was negative expectations for the future. This aspect of gender minority stress encompasses specific expectations of discrimination, anti-trans violence, and social rejection (Horwitz et al., 2017; Kosnes et al., 2013)—that is, negative expectations directly tied to gender identity rather than part of the general experience of clinical depression. Prior cross-sectional analysis found that negative expectations were significantly linked to depression for TGD youth; however, this link was no longer statistically significant when internalized transphobia and community connectedness were added as additional variables, suggesting that negative expectations were important but less so than other factors for that sample (Chodzen et al., 2019). In a recent study, Helsen and colleagues (2022) also found support for expectations of future rejection (a component of negative expectations for the future) as a significant predictor of mental health difficulties among TGD adults. Another study examining identity-specific intimate partner violence and mental health outcomes for TGD young adults found that negative expectations for the future emerged as a significant mediator in this relationship. The authors posit that TGD young adults who have experienced abuse associated with their gender identity may develop an expectation for future disrespect or maltreatment from others due to their identity; these negative expectations may then contribute to the development of worse mental health outcomes (Taber et al., 2023). Thus, negative expectations for the future continue to be identified as a potential mechanism for more direct treatment, i.e., an intervenable factor, to improve mental health outcomes (Coyne et al., 2020).

We further found that greater self-efficacy was a significant protective factor against PD. Self-efficacy has been identified as a key factor in mental well-being that is both impacted by depression (Li et al., 2024) and an important target for intervention (Blakeslee et al., 2023). This finding provides further support that this is a potential mechanism or target for treatment among youth who are at-risk for PD. We also found that higher body esteem at baseline was a protective factor for PD, though this result was marginally significant.

Finally, we observed that the rates of change in both negative expectations for the future and self-efficacy were significantly associated with PD. Finally, an exploration of interaction effects found a significant interaction between expectations and self-efficacy, suggesting a potential attenuating effect of negative expectations on the influence of self-efficacy or potentially a stronger influence of negative expectations despite self-efficacy, though this was beyond the scope of the current analysis. Finally, we observed an association between PTSD diagnosis at any time during the first two years of the study and increased likelihood of PD, which is aligned with current research highlighting the importance of addressing trauma within models of gender-affirming mental health care (Salomaa et al., 2024).

Limitations.

The current study has several limitations. We defined PD using a single indicator of depressive symptoms, the ASR/YSR. We chose this measure because it is well-validated in the literature and has been found to be highly associated with MDD diagnoses (Jiang et al., 2023; van Lang et al., 2005). Indeed, we observed convergent validity for this categorization based on other measures of depression within this study, including a brief screener of depressive symptoms (BDI-21) and a semi-structured clinical interview of DSM-5 criteria for Major Depressive Disorder (see Table 1). We also chose the current measure because it reflects symptoms over the past six months and was assessed annually in the TYCUS study; however, this measure may not capture other fluctuations or periods without depressive symptoms outside of these time points. In addition, we defined PD as depression across all three timepoints given our analytic approach; future analyses will explore factors associated with trajectories of depression accounting for changes in depressive symptoms at different timepoints. As observed in Table 2, there are some participants experiencing significant depressive symptoms at 24-months that were not captured in the PD group; thus, additional analytical exploration is warranted to further examine subgroups of youth who may have developed worsening depressive symptoms by later time points. We observed overrepresentation of transmasculine, Latiné and non-Latiné White, while multiracial youth and Black and transfeminine youth were underrepresented in our study. Still, the study sample is considered similar to many pediatric gender care programs in the U.S. and thus, may be considered generalizable to these populations (Chen et al., 2021). To date, there has been mixed evidence of differences in gender minority stress and resilience across gender identity between transmasculine, transfeminine, and non-binary youth (Chavanduka et al., 2021; Poquiz et al., 2021), highlighting one of the limitations of the GMS model regarding the role of intersectionality. While the present study accounted for designated sex at birth, future studies should include an examination of the differences in these experiences based on gender identity as well. The current analysis did not include external minority stressors, such as gender-based victimization, rejection, and discrimination that are in the GMSR measure, which were captured in later survey waves but not in the first two waves.

Finally, while our research team as a whole was comprised of individuals with decades of experience within the transgender health space, we acknowledge the limitations of our perspectives in considering our findings given that our current research team has limited representation from the TGD community. At various points in our analysis, we have sought input and feedback from team members with significant expertise in TGD mental health, including thoughtful and careful conversations regarding implications of these findings and potential for our results to be taken out of context and mis-represented in our current climate of anti-transgender misinformation. Thus, additional care was taken in the development of the analytic plan, interpretation of the results, and framing of these findings to center the lived experiences of TGD individuals, encourage policies and efforts to support their mental health, and identify intervenable factors to help us with these efforts.

Summary and future directions.

The finding of PD in a subset of youth treated with GAHT highlights the fact that GAHT is not a panacea for the mental health issues faced by TGD youth, but rather, there may be other factors that contribute to PD even when access to GAHT is beneficial. These findings do not convey that gender-affirming medical care does not provide benefit – to the contrary, the majority of our participants were in the non-PD group and even within the PD group, certain positive changes were noted. For those who did experience PD, the framework provided by the GMS model emphasizes multiple factors that contribute to the disparity in mental health faced by TGD youth compared to their peers, and our work pinpoints critical and potentially intervenable factors associated with depressive symptoms that persist despite the initiation of GAHT. Still, while inclusive of myriad stressors, the GMS model is limited in its consideration of longitudinal impacts of such factors on TGD mental health, as well as the role of diverse resilience and coping factors. In the present study, we found support for the unique role of negative expectations related to discrimination and hardship associated with gender identity in the persistence of depressive symptoms for TGD youth. Our sample lived in areas of the U.S. that had policies that protect gender care and TGD youth at the time of data collection. There is a burgeoning need to examine the impact of minority stress and resilience factors on the risk of PD for TGD youth living in socio-historical environments with greater anti-transgender stigma and legislation, lack of support and acceptance, concerns about future opportunities for education, leisure, or employment, and worries about future risk of victimization and discrimination, particularly in our current socio-political context (Kaltiala et al., 2020). Our study also identified several resilience or protective factors that are not included in the GMS model that should be examined further in future studies. Recent research indicates that increased knowledge of gender minority youth, mental health services, and community resources (i.e., peer support groups, student organizations, online resources) are key areas for continued development to augment resilience for TGD youth (Hammack et al., 2024; Rogowska & and Cisek, 2024), and our work points to the need for such inventions at the community level as well as at the clinical and policy levels to optimize outcomes for TGD youth.

Supplementary Material

SupMat

Public Significance Statement:

The present study suggests that while gender-affirming hormone therapy improves mental health for many transgender and gender diverse youth, addressing cognitive risk factors like negative expectations for the future and boosting self-efficacy are crucial for those experiencing persistent depressive symptoms. Tailoring psychological interventions to these specific risk and protective factors can help maximize treatment efficacy and support overall mental health outcomes in this population.

Funding Source/Acknowledgements:

This work was supported by funding from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD; R01HD082554). The protocol for this study is available online through JMIR research protocols at: https://www.researchprotocols.org/2019/7/e14434/

Footnotes

Disclosure/Conflicts of Interest: The authors have no conflicts of interest or financial disclosures to report relevant to the present analysis/manuscript.

Data availability:

Data used for analyses and analytic code are available from the authors upon request.

References

  1. Aboussouan A, Snow A, Cerel J, & Tucker RP. (2019). Non-suicidal self-injury, suicide ideation, and past suicide attempts: Comparison between transgender and gender diverse veterans and non-veterans. J Affect Disord, 259, 186–194. 10.1016/j.jad.2019.08.046 [DOI] [PubMed] [Google Scholar]
  2. Achenbach TM. (2014). DSM-Oriented Guide for the Achenbach System of Empirically Based Assessment (ASEBA®). University of Vermont Research Center for Children, Youth, and Families. [Google Scholar]
  3. Achenbach TM. (2015). Achenbach System of Empirically Based Assessment (ASEBA). In The Encyclopedia of Clinical Psychology (pp. 1–8). John Wiley & Sons, Ltd. 10.1002/9781118625392.wbecp150 [DOI] [Google Scholar]
  4. Achenbach TM, & Rescorla LA. (2004). The Achenbach System of Empirically Based Assessment (ASEBA) for Ages 1.5 to 18 Years. In The Use of Psychological Testing for Treatment Planning and Outcomes Assessment (3rd ed.). Routledge. [Google Scholar]
  5. American Civil Liberties Union. (2021). Legislation Affecting LGBTQ Rights Across the Country 2021. American Civil Liberties Union. https://www.aclu.org/documents/legislation-affecting-lgbtq-rights-across-country-2021 [Google Scholar]
  6. American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders: DSM-5™, 5th ed. American Psychiatric Publishing, Inc. [Google Scholar]
  7. Beck AT, Steer RA, & Brown G. (1996). Beck Depression Inventory–II. 10.1037/t00742-000 [DOI] [Google Scholar]
  8. Blakeslee JE, Kothari BH, & Miller RA. (2023). Intervention development to improve foster youth mental health by targeting coping self-efficacy and help-seeking. Children and Youth Services Review, 144, 106753. 10.1016/j.childyouth.2022.106753 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Bornioli A, Lewis-Smith H, Slater A, & Bray I. (2021). Body dissatisfaction predicts the onset of depression among adolescent females and males: A prospective study. J Epidemiol Community Health, 75(4), 343–348. 10.1136/jech-2019-213033 [DOI] [PubMed] [Google Scholar]
  10. Brecht A, Bos S, Ries L, Hübner K, Widenka P-M, Winter SM, & Calvano C. (2024). Analyzing body dissatisfaction and gender dysphoria in the context of minority stress among transgender adolescents. Child and Adolescent Psychiatry and Mental Health, 18(1), 30. 10.1186/s13034-024-00718-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Chavanduka TMD, Gamarel KE, Todd KP, & Stephenson R. (2021). Responses to the gender minority stress and resilience scales among transgender and nonbinary youth. Journal of LGBT Youth, 18(2), 135–154. 10.1080/19361653.2020.1719257 [DOI] [Google Scholar]
  12. Chelliah P, Lau M, & Kuper LE. (2024). Changes in Gender Dysphoria, Interpersonal Minority Stress, and Mental Health Among Transgender Youth After One Year of Hormone Therapy. The Journal of Adolescent Health: Official Publication of the Society for Adolescent Medicine, S1054-139X(24)00005-3. 10.1016/j.jadohealth.2023.12.024 [DOI] [PubMed] [Google Scholar]
  13. Chen D, Berona J, Chan Y-M, Ehrensaft D, Garofalo R, Hidalgo MA, Rosenthal SM, Tishelman AC, & Olson-Kennedy J. (2023). Psychosocial Functioning in Transgender Youth after 2 Years of Hormones. New England Journal of Medicine, 388(3), 240–250. 10.1056/NEJMoa2206297 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Chen D, Lash B, Kim E, Hidalgo MA, Muldoon AL, Liu E, Jensen J, Grabert R, Chan Y-M, Garofalo R, & Tishelman A. (2021). A Comparison of Demographic and Psychosocial Characteristics Between Transgender Youth Enrolling Versus Not Enrolling in a Multisite Study. Transgender Health, 6(4), 229–234. 10.1089/trgh.2020.0060 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Ching THW, Finkelstein-Fox L, Lee SY, & Watson RJ. (2024). Effects of sexual and gender minority stress on depressive symptoms among adolescents of color in the United States. Cultural Diversity & Ethnic Minority Psychology, 30(2), 309–318. 10.1037/cdp0000562 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Chodzen G, Hidalgo MA, Chen D, & Garofalo R. (2019). Minority Stress Factors Associated With Depression and Anxiety Among Transgender and Gender-Nonconforming Youth. The Journal of Adolescent Health : Official Publication of the Society for Adolescent Medicine, 64(4), 467–471. 10.1016/j.jadohealth.2018.07.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Conn BM, Chen D, Olson-Kennedy J, Chan Y-M, Ehrensaft D, Garofalo R, Rosenthal SM, Tishelman A, & Hidalgo MA. (2023). High Internalized Transphobia and Low Gender Identity Pride Are Associated With Depression Symptoms Among Transgender and Gender-Diverse Youth. Journal of Adolescent Health, 72(6), 877–884. 10.1016/j.jadohealth.2023.02.036 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Coyne CA, Poquiz JL, Janssen A, & Chen D. (2020). Evidence-Based Psychological Practice for Transgender and Non-Binary Youth: Defining the Need, Framework for Treatment Adaptation, and Future Directions. Evidence-Based Practice in Child and Adolescent Mental Health, 5(3), 340–353. 10.1080/23794925.2020.1765433 [DOI] [Google Scholar]
  19. Cragun D, DeBate RD, Ata RN, & Kevin TJ. (2013). Psychometric properties of the Body Esteem Scale for Adolescents and Adults in an early adolescent sample. Eating and Weight Disorders, 18(3), 275–282. 10.1007/s40519-013-0031-1 [DOI] [PubMed] [Google Scholar]
  20. Dales LG, & Ury HK. (1978). An improper use of statistical significance testing in studying covariables. International Journal of Epidemiology, 7(4), 373–375. 10.1093/ije/7.4.373 [DOI] [PubMed] [Google Scholar]
  21. Dolezal ML, Decker M, Higgins M, & Littleton H. (2023). Evaluation of the gender minority stress model in transgender and gender diverse college students. Psychology of Sexual Orientation and Gender Diversity. 10.1037/sgd0000669 [DOI] [Google Scholar]
  22. Fitzmaurice GM, Laird NM, & Ware JH. (2011). Linear Mixed Effect Models. In Applied Longitudinal Analysis (2nd ed., pp. 209–210). John Wiley & Sons, Ltd. https://www.wiley.com/en-us/Applied+Longitudinal+Analysis%2C+2nd+Edition-p-9780470380277#tableofcontents-section [Google Scholar]
  23. Gibson DJ, Glazier JJ, & Olson KR. (2021). Evaluation of Anxiety and Depression in a Community Sample of Transgender Youth. JAMA Network Open, 4(4), e214739. 10.1001/jamanetworkopen.2021.4739 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Hammack PL, Pletta DR, Hughes SD, Atwood JM, Cohen EM, & Clark RC. (2024). Community support for sexual and gender diversity, minority stress, and mental health: A mixed-methods study of adolescents with minoritized sexual and gender identities. Psychology of Sexual Orientation and Gender Diversity, 11(2), 250–268. 10.1037/sgd0000591 [DOI] [Google Scholar]
  25. Hankin BL, & Griffith JM. (2023). What Do We Know About Depression Among Youth and How Can We Make Progress Toward Improved Understanding and Reducing Distress? A New Hope. Clinical Child and Family Psychology Review, 26(4), 919–942. 10.1007/s10567-023-00437-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Helsen V, Enzlin P, & Gijs L. (2022). Mental health in transgender adults: The role of proximal minority stress, community connectedness, and gender nonconformity. Psychology of Sexual Orientation and Gender Diversity, 9(4), 466–477. 10.1037/sgd0000530 [DOI] [Google Scholar]
  27. Hendricks ML, & Testa RJ. (2012). A conceptual framework for clinical work with transgender and gender nonconforming clients: An adaptation of the Minority Stress Model. Professional Psychology: Research and Practice, 43(5), 460–467. 10.1037/a0029597 [DOI] [Google Scholar]
  28. Hidalgo MA, Petras H, Chen D, & Chodzen G. (2019). The Gender Minority Stress and Resilience Measure: Psychometric Validity of an Adolescent Extension. Clinical Practice in Pediatric Psychology, 7(3), 278–290. 10.1037/cpp0000297 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Hodes RJ, Insel TR, Landis SC, & On behalf of the NIH Blueprint for Neuroscience Research. (2013). The NIH Toolbox. Neurology, 80(11_supplement_3), S1–S1. 10.1212/WNL.0b013e3182872e90 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Horwitz AG, Berona J, Czyz EK, Yeguez CE, & King CA. (2017). Positive and Negative Expectations of Hopelessness as Longitudinal Predictors of Depression, Suicidal Ideation, and Suicidal Behavior in High-Risk Adolescents. Suicide & Life-Threatening Behavior, 47(2), 168–176. 10.1111/sltb.12273 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Hosmer DW Jr, Lemeshow S, & Sturdivant RX. (2013). Applied Logistic Regression. John Wiley & Sons. [Google Scholar]
  32. Jiang Z, Wang D, Xu H, Zhang A, Zhao Q, Yan J, Wang X, Zhang W, Li Y, Yang K, Hu S, Cui Y, & Li Y. (2023). Diagnostic efficiency and psychometric properties of CBCL DSM-oriented scales in a large sample of Chinese school-attending students aged 5–16. Asian Journal of Psychiatry, 88, 103724. 10.1016/j.ajp.2023.103724 [DOI] [PubMed] [Google Scholar]
  33. Kaltiala R, Heino E, Työläjärvi M, & Suomalainen L. (2020). Adolescent development and psychosocial functioning after starting cross-sex hormones for gender dysphoria. Nordic Journal of Psychiatry, 74(3), 213–219. 10.1080/08039488.2019.1691260 [DOI] [PubMed] [Google Scholar]
  34. Kosnes L, Whelan R, O’Donovan A, & McHugh LA. (2013). Implicit measurement of positive and negative future thinking as a predictor of depressive symptoms and hopelessness. Consciousness and Cognition, 22(3), 898–912. 10.1016/j.concog.2013.06.001 [DOI] [PubMed] [Google Scholar]
  35. Kuper LE, Stewart S, Preston S, Lau M, & Lopez X. (2020). Body Dissatisfaction and Mental Health Outcomes of Youth on Gender-Affirming Hormone Therapy. Pediatrics, 145(4), e20193006. 10.1542/peds.2019-3006 [DOI] [PubMed] [Google Scholar]
  36. Lam JSH, Abramovich A, Victor JC, Zaheer J, & Kurdyak P. (2021). Characteristics of Transgender Individuals With Emergency Department Visits and Hospitalizations for Mental Health. Psychiatric Services, appi.ps.202100306. 10.1176/appi.ps.202100306 [DOI] [PubMed] [Google Scholar]
  37. Lee WY, Hobbs JN, Hobaica S, DeChants JP, Price MN, & Nath R. (2024). State-level anti-transgender laws increase past-year suicide attempts among transgender and non-binary young people in the USA. Nature Human Behaviour, 8(11), 2096–2106. 10.1038/s41562-024-01979-5 [DOI] [PubMed] [Google Scholar]
  38. Lefevor GT, Boyd-Rogers CC, Sprague BM, & Janis RA. (2019). Health disparities between genderqueer, transgender, and cisgender individuals: An extension of minority stress theory. J Couns Psychol, 66(4), 385–395. 10.1037/cou0000339 [DOI] [PubMed] [Google Scholar]
  39. Li X, Xia B, Shen G, Dong R, Xu S, & Yang L. (2024). The interplay of depressive symptoms and self-efficacy in adolescents: A network analysis approach. Frontiers in Psychology, 15. 10.3389/fpsyg.2024.1419920 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Mendelson BK, Mendelson MJ, & White DR. (2001). Body-esteem scale for adolescents and adults. Journal of Personality Assessment, 76(1), 90–106. 10.1207/S15327752JPA7601_6 [DOI] [PubMed] [Google Scholar]
  41. Mendelson BK, White DR, & Mendelson MJ. (1996). Self-esteem and body esteem: Effects of gender, age, and weight. Journal of Applied Developmental Psychology, 17(3), 321–346. 10.1016/S0193-3973(96)90030-1 [DOI] [Google Scholar]
  42. Mickey RM, & Greenland S. (1989). The impact of confounder selection criteria on effect estimation. American Journal of Epidemiology, 129(1), 125–137. 10.1093/oxfordjournals.aje.a115101 [DOI] [PubMed] [Google Scholar]
  43. Morales-Muñoz I, Mallikarjun PK, Chandan JS, Thayakaran R, Upthegrove R, & Marwaha S. (2023). Impact of anxiety and depression across childhood and adolescence on adverse outcomes in young adulthood: A UK birth cohort study. The British Journal of Psychiatry, 222(5), 212–220. 10.1192/bjp.2023.23 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Morin AJS, Maïano C, Marsh HW, Janosz M, & Nagengast B. (2011). The Longitudinal Interplay of Adolescents’ Self-Esteem and Body Image: A Conditional Autoregressive Latent Trajectory Analysis. Multivariate Behavioral Research, 46(2), 157–201. 10.1080/00273171.2010.546731 [DOI] [PubMed] [Google Scholar]
  45. Olson-Kennedy J, Chan Y-M, Garofalo R, Spack N, Chen D, Clark L, Ehrensaft D, Hidalgo M, Tishelman A, & Rosenthal S. (2019). Impact of Early Medical Treatment for Transgender Youth: Protocol for the Longitudinal, Observational Trans Youth Care Study. JMIR Research Protocols, 8(7), e14434. 10.2196/14434 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Olson-Kennedy J, Wang L, Wong CF, Chen D, Ehrensaft D, Hidalgo MA, Tishelman AC, Chan Y-M, Garofalo R, Radix AE, & Rosenthal SM. (2025). Emotional Health of Transgender Youth 24 Months After Initiating Gender-Affirming Hormone Therapy. The Journal of Adolescent Health: Official Publication of the Society for Adolescent Medicine, 77(1), 41–50. 10.1016/j.jadohealth.2024.11.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Parchem B, Berona J, Coyne CA, Moskowitz JT, Chan Y-M, Ehrensaft D, Garofalo R, Hidalgo MA, Olson-Kennedy J, Rosenthal SM, Tishelman AC, & Chen D. (2025). Minority stress, mental health, and substance use in transgender youth: The moderating role of positive affect. Journal of Pediatric Psychology, 50(5), 412–419. 10.1093/jpepsy/jsaf018 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Pellicane MJ, & Ciesla JA. (2022). Associations between minority stress, depression, and suicidal ideation and attempts in transgender and gender diverse (TGD) individuals: Systematic review and meta-analysis. Clinical Psychology Review, 91, 102113. 10.1016/j.cpr.2021.102113 [DOI] [PubMed] [Google Scholar]
  49. Poquiz JL, Coyne CA, Garofalo R, & Chen D. (2021). Comparison of Gender Minority Stress and Resilience Among Transmasculine, Transfeminine, and Nonbinary Adolescents and Young Adults. Journal of Adolescent Health, 68(3), 615–618. 10.1016/j.jadohealth.2020.06.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Price-Feeney M, Green AE, & Dorison S. (2020). Understanding the Mental Health of Transgender and Nonbinary Youth. The Journal of Adolescent Health : Official Publication of the Society for Adolescent Medicine, 66(6), 684–690. 10.1016/j.jadohealth.2019.11.314 [DOI] [PubMed] [Google Scholar]
  51. Rawana JS, & Morgan AS. (2014). Trajectories of Depressive Symptoms from Adolescence to Young Adulthood: The Role of Self-esteem and Body-Related Predictors. Journal of Youth and Adolescence, 43(4), 597–611. 10.1007/s10964-013-9995-4 [DOI] [PubMed] [Google Scholar]
  52. Reisner SL, Greytak EA, Parsons JT, & Ybarra ML. (2015). Gender minority social stress in adolescence: Disparities in adolescent bullying and substance use by gender identity. Journal of Sex Research, 52(3), 243–256. 10.1080/00224499.2014.886321 [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Ribeiro JD, Huang X, Fox KR, & Franklin JC. (2018). Depression and hopelessness as risk factors for suicide ideation, attempts and death: meta-analysis of longitudinal studies. The British Journal of Psychiatry, 212(5), 279–286. 10.1192/bjp.2018.27 [DOI] [PubMed] [Google Scholar]
  54. Rogowska AM, & and Cisek A. (2024). Minority stress, perceived social support, and depression in people diverse in sexual and gender minority status. Psychology & Sexuality, 15(4), 679–693. 10.1080/19419899.2024.2333858 [DOI] [Google Scholar]
  55. Salifu Yendork J, & Somhlaba NZ. (2015). Do Social Support, Self-efficacy and Resilience Influence the Experience of Stress in Ghanaian Orphans? An Exploratory Study. Child Care in Practice, 21(2), 140–159. 10.1080/13575279.2014.985286 [DOI] [Google Scholar]
  56. Salomaa AC, Berke D, Harper K, Valentine SE, Sloan CA, Hinds Z, Gyuro L, Herbitter C, Bryant WT, Shipherd JC, & Livingston NA. (2024). A patient-centered model of mental health care for trauma and minority stress in transgender and gender diverse people: A bottom-up network analysis. Psychology of Sexual Orientation and Gender Diversity. 10.1037/sgd0000705 [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Sheehan DV, Lecrubier Y, Sheehan KH, Amorim P, Janavs J, Weiller E, Hergueta T, Baker R, & Dunbar GC. (1998). The Mini-International Neuropsychiatric Interview (M.I.N.I.): The development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. The Journal of Clinical Psychiatry, 59 Suppl 20, 22–33;quiz 34-57. [PubMed] [Google Scholar]
  58. Taber JL, Stults CB, Song H, & Kaczetow W. (2023). The role of internalized transphobia and negative expectations in the relationship between identity-specific intimate partner violence and mental health outcomes in transgender and gender nonconforming young adults. Psychology of Sexual Orientation and Gender Diversity, No Pagination Specified-No Pagination Specified. 10.1037/sgd0000641 [DOI] [Google Scholar]
  59. Testa RJ, Habarth J, Peta J, Balsam K, & Bockting W. (2015). Development of the Gender Minority Stress and Resilience Measure. Psychology of Sexual Orientation and Gender Diversity, 2(1), 65–77. 10.1037/sgd0000081 [DOI] [Google Scholar]
  60. Trujillo MA, Perrin PB, Sutter M, Tabaac A, & Benotsch EG. (2017). The buffering role of social support on the associations among discrimination, mental health, and suicidality in a transgender sample. The International Journal of Transgenderism, 18(1), 39–52. 10.1080/15532739.2016.1247405 [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Turnamian MR, & Liu RT. (2023). Gender identity and expression in relation to depression and anxiety in racial and ethnic minority youth: Evaluations of intersectionality in a population-based study. Journal of Affective Disorders, 339, 219–226. 10.1016/j.jad.2023.07.023 [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. van Lang NDJ, Ferdinand RF, Oldehinkel AJ, Ormel J, & Verhulst FC. (2005). Concurrent validity of the DSM-IV scales Affective Problems and Anxiety Problems of the Youth Self-Report. Behaviour Research and Therapy, 43(11), 1485–1494. 10.1016/j.brat.2004.11.005 [DOI] [PubMed] [Google Scholar]
  63. Vibhakar V, Allen LR, Gee B, & Meiser-Stedman R. (2019). A systematic review and meta-analysis on the prevalence of depression in children and adolescents after exposure to trauma. Journal of Affective Disorders, 255(1), 77–89. 10.1016/j.jad.2019.05.005 [DOI] [PubMed] [Google Scholar]
  64. Weavers B, Heron J, Thapar AK, Stephens A, Lennon J, Bevan Jones R, Eyre O, Anney RJ, Collishaw S, Thapar A, & Rice F. (2021). The antecedents and outcomes of persistent and remitting adolescent depressive symptom trajectories: A longitudinal, population-based English study. The Lancet. Psychiatry, 8(12), 1053–1061. 10.1016/S2215-0366(21)00281-9 [DOI] [PubMed] [Google Scholar]
  65. Wiglesworth A. (2024). Risk for suicide in adolescence: How identity, development, and minority stress may shape neurobiological responses to social rejection. Biological Psychiatry, 95(12). https://www.biologicalpsychiatryjournal.com/article/S0006-3223(24)01255-1/abstract [DOI] [PubMed] [Google Scholar]
  66. Zimet GD, Dahlem NW, Zimet SG, & Farley GK. (1988). The Multidimensional Scale of Perceived Social Support. Journal of Personality Assessment, 52(1), 30–41. 10.1207/s15327752jpa5201_2 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

SupMat
NIHMS2208730-supplement-SupMat.docx (18.3KB, docx)

Data Availability Statement

Data used for analyses and analytic code are available from the authors upon request.

RESOURCES