Structural Needs, Substance Use, and Mental Health Among Transgender and Nonbinary Young Adults in the San Francisco Bay Area: Findings from the Phoenix Study

Abstract

Transgender and nonbinary (trans) young adults report high rates of substance use and adverse mental health outcomes; however, few studies have examined how social, economic, and legal factors may contribute to health inequities in this population. Guided by the structural vulnerability framework, this study sought to explore structural needs and whether these needs were associated with substance use and mental health outcomes among trans young adults. Between 2019 and 2021, 215 trans young adults aged 18–29 from San Francisco Bay Area were recruited into a longitudinal study. Baseline data were used to examine bivariate and multivariable associations between structural needs and substance use and mental health outcomes. There were bivariate differences in the number of structural needs by education, income source, incarceration history, and ethnicity, and the number of unmet structural needs was associated with education and income source. After adjusting for sociodemographics, the number of structural needs was associated with daily marijuana use (AOR 1.29, 95% CI: 1.10–1.49) and suicidal ideation (AOR 1.24, 95% CI: 1.06–1.45), and the number of unmet structural needs was associated with daily marijuana use (AOR 1.30, 95% CI: 1. 10–1.55) and depressive symptoms (β 2.00, 95% CI: 1.00–3.00). Additionally, both numbers of structural needs and unmet structural needs mediated the relationship between income source (traditional employment vs. other income only) and depressive symptoms (TIE β 2.51, 95% CI: 0.99–4.04; β 1.37, 95% CI: 0.23–2.52, respectively). Findings highlight a need for multisector efforts to address structural vulnerabilities among trans young adults.

Introduction

Transgender and nonbinary (trans) individuals have a high prevalence of mental health and substance use problems that persist across developmental periods [1–12].  For example, a recent study found inequities between trans and cisgender adults regarding alcohol, polysubstance, and other substance use disorders with the sharpest group inequities observed between 18 and 25 years old [13]. Studies that have found differences in mental health and substance use problems between subgroups of trans adolescents also suggest that trans youth tend to have worse outcomes compared to cisgender youth [14–17].

Inequities impacting trans young adults have generally been attributed to cisgenderism—systemic bias against trans individuals that manifests across socioecological levels—and its impacts on access to gender affirmation, healthcare experiences, social support, and resources for coping with stigma among trans people [4, 15, 18–21]. Studies grounded in the Gender Minority Stress Model suggest that many trans individuals engage in substance use to cope with internalized and interpersonal stigma in the absence of other coping resources [8, 22–25]. Moreover, stigmatizing experiences contribute to increased levels of depression, anxiety, and other mental health symptoms among trans people [26–28].

While understanding how gender minority stress impacts substance use and mental health is important, cisgenderism extends beyond internalized and interpersonal stigma [18, 29]. Attending to how cisgenderism shapes the social, economic, and legal landscapes in which trans young adults live is crucial to contextualizing health inequities [30–34]. Furthermore, for trans young adults of color, considering how ethno-racism intersects with cisgenderism provides critical context to understanding health inequities [35–38]. The structural vulnerability framework explains how marginalized groups are subjected to physical and emotional suffering as a product of economic exploitation and social stigma such as cisgenderism and ethno-racism [39]. Structural vulnerabilities are experiences that threaten an individuals’ survival, such as limited food access, housing deprivation, and interpersonal violence [40]. Applying the structural vulnerability framework to substance use and mental health problems among trans young adults reframes these inequities as products of systemic oppression rather than as consequences of individual behaviors.

On a population level, trans young adults can be considered structurally vulnerable due to the extent to which they experience family rejection, homelessness, violence, poverty, and discrimination in schools, employment, the criminal-legal system, social services, and healthcare [35, 41–46]. Structural vulnerabilities including financial hardship, incarceration, and housing deprivation have been linked to adverse mental health and substance use outcomes among trans adults [8, 47–51]. However, many studies focused on understanding mental health and substance use among trans young adults have statistically adjusted for structural vulnerability indicators (e.g., income) without attending to how they may produce health inequities [8, 11, 52, 53].

A better understanding of structural needs of trans young adults and their relationship to mental health and substance use may have important implications for intervention [54, 55]. Rather than presume participants are structurally vulnerable because they report, for example, having a low income or not having a high school diploma, examining self-reported structural needs reflects participants’ lived experiences of structural vulnerability. Therefore, this study is aimed at describing sociodemographic patterns in self-reported structural needs and unmet structural needs and examining the relationship between structural needs and mental health and substance use outcomes among trans young adults. This study is intended to identify upstream factors that can be intervened upon to prevent occurrence of mental health and substance use problems among trans young adults.

First, we expect to find socioeconomic and racial/ethnic variation in structural needs such that participants with more socioeconomic resources (e.g., employment, higher education) and who benefit from white supremacy report fewer overall structural needs and unmet structural needs than those with fewer socioeconomic resources and participants of color, respectively [37, 38, 56, 57]. Given associations between structural vulnerability markers and substance use and mental health outcomes in previous studies with trans adults, we hypothesize that the number of structural needs and unmet structural needs will be associated with mental health and substance use outcomes [8, 47–51]. Finally, we expect to find that the number of structural needs and unmet structural needs mediates any associations found between sociodemographic variables and substance use or mental health outcomes as the structural vulnerability framework suggests that societal power dynamics like cisgenderism and ethno-racism determine which sociodemographic groups will be able to meet their basic survival needs [39].

Methods

Study Design

From 2019 to 2021, the Phoenix Study recruited 215 trans and nonbinary young adults living in the San Francisco (SF) Bay Area to participate in a two-year, mixed-methods longitudinal investigation of substance use behaviors. This study presents a cross-sectional analysis of the quantitative baseline data collected via an interviewer-administered survey. Participants were recruited from community venues (e.g., nonprofit organizations and bars) and online social networks using purposive and snowball sampling techniques. The eligibility criteria included being 18–29 years old, self-identifying as trans or a gender identity that did not match their sex assigned on their original birth certificate, living in the SF Bay Area, and being able to complete study procedures in English. Those who completed the baseline survey received a $30 gift card. All study procedures were approved by an Institutional Review Board at the Public Health Institute (Oakland, CA).

Measures

Structural Needs

Participants reported whether they needed medical or social services in the last 6 months and, if so, whether they received these services. Of the 17 included services, we considered the following to reflect structural needs: housing (needing either permanent housing or temporary shelter), food assistance, job training, disability benefits, health insurance, assistance with utility payments, unemployment benefits, legal assistance (needing either legal aid or immigration services), and crisis intervention. We created dummy variables for each need and each unmet need (i.e., if they had not received services for their structural needs in the last 6 months) and conducted exploratory factor analyses of all the need and unmet need variables separately. Both factor analyses indicated that a one-factor solution best represented the data. Cronbach’s alpha was 0.77 for structural needs and 0.74 for unmet structural needs, indicating acceptable internal reliability. We totaled participants’ number of structural needs and number of unmet structural needs, resulting in scales ranging from 0 to 9.

Sociodemographic Characteristics

Participants reported their age in years, whether they were born in the US (yes/no), and whether they had ever been incarcerated (yes/no). Participants also reported their educational attainment, source of income, race and ethnicity, and gender identity. We categorized participants’ highest level of education into bachelor’s degree or higher, postsecondary (no degree), and high school, GED, or lower. Participants were grouped into the postsecondary (no degree) category if they reported completing a vocational or technical program or some college. Participants completed a multiselect item asking for their income sources in the past 6 months. We categorized participants’ income sources into traditional employment, other sources only, and no income. We considered participants to be traditionally employed if they reported having a full or part-time job regardless of other sources of income. Participants who reported having income but not from a full or part-time job were categorized as having other income sources only. These sources of income included gig or contract work, public benefits, income from family members or partners, sex work, and write-in responses.

Participants self-reported their race/ethnicity from the following options: African American, Latinx or Hispanic, Asian or Pacific Islander, Native American, white, multiracial/multiethnic, or other. All participants were additionally given the option of providing open-response data. From this data, we categorized 5 of the 7 participants who reported an “other” race/ethnicity into one of the other categories; the remaining were coded as missing ethnicity.

Participants chose which gender identity they considered themselves to have from the following options: cis man, cis woman, trans man, trans woman, genderqueer/gender nonconforming/nonbinary, do not know, and other. Participants who selected “other” (n = 40) were given an open-response item asking for greater detail. We created three gender categories from the multiple choice and free response data. Participants were categorized as nonbinary + if they selected nonbinary/genderqueer/gender nonconforming, do not know, or wrote in any term other than trans feminine, trans masculine, or their close variations. For example, participants in the nonbinary + category identified as “nonbinary trans femme,” and “agender.” Participants were categorized as men/masculine if they chose trans man or cis man or wrote in any variation of trans masculine on its own. For instance, participants who wrote in “nonbinary trans masculine” were categorized as nonbinary + while participants who wrote in “transmasculine” only were classified as men/masculine. Similarly, participants were categorized as women/feminine if they chose trans woman or cis woman or wrote in any variation of trans feminine on its own.

Substance Use and Mental Health Outcomes

Participants reported their substance use over the past 6 months using the ASSIST 3.0 [58]. Based on the high prevalence of marijuana and tobacco use in this sample (79.7% and 83.8% respectively), we modeled both any use and daily use. Club drug use was defines as any cocaine, amphetamine, or hallucinogen use. Participants were classified as heavy episodic drinkers if they reported ever having 5 or more alcoholic drinks on one occasion in the past 6 months.

Mental health outcomes included suicidal ideation, depression, and gender-related posttrauma stress symptoms disorder (PSTD) symptoms. Suicidal ideation was assessed with a single item asking whether participants had thought about trying to complete suicide in the past 6 months. Depressive symptoms were assessed with the 20-item Center for Epidemiological Studies Depression scale in which participants reported the frequency of common symptoms of depression in the past week (α = 0.91) [59]. Gender-related PTSD symptoms were measured with a modified version of the Primary Care PTSD Screen [60]. Participants completed four dichotomous items regarding their PTSD symptoms over the past 6 months stemming from frightening, horrible, or upsetting experiences related to their gender identity (α = 0.89).

Analysis

We first generated frequencies and means to characterize the sample by structural needs, sociodemographic characteristics, and health outcomes. We then analyzed sociodemographic differences in the distribution of structural needs and unmet structural needs using ANOVA tests for categorical variables and correlations for continuous variables.

To examine the relationship between structural needs and mental health and substance use outcomes, we fit bivariate regression models predicting each outcome by number of structural needs and by number of unmet structural needs. We then fit adjusted models which included gender and the sociodemographic characteristics that were significantly associated with structural needs and unmet structural needs in bivariate analyses. We used multivariable logistic regression for dichotomous outcomes and linear regression for continuous outcomes.

Finally, we explored whether structural needs potentially mediated the relationship between sociodemographic characteristics and mental health and substance use outcomes using structural equation modeling. All analyses were conducted in STATA version 16.1 (College Station, TX), and mediation analyses used the medsem package with bootstrapped standard errors for the total indirect effects (1000 replications) [61].

Results

Characteristics of the sample are reported in Table 1. Participants ranged in age from 18 to 29 (M = 24.4, SD = 3.2). Most participants were categorized as nonbinary + (52.1%); 29.3% were categorized as women/feminine and 18.6% as men/masculine. The majority of participants (87.4%) were born in the US. Regarding education, 41.9% of the sample had a bachelor’s degree or higher and 35.8% completed some postsecondary education. Most participants (53.5%) earned income from outside of traditional employment, and a small proportion (3.3%) reported having no income. Finally, 14.6% reported ever having been incarcerated.

Table 1.

Characteristics of the sample (n = 215)

Demographic characteristics	%/M	n/SD
Race/ethnicity
White	37.4	80
Black/African American	8.9	19
Latino/a/x	19.2	41
Asian/Pacific Islander	17.3	37
Native American	1.9	4
Multiracial/multiethnic	15.4	33
Gender
Man/masculine	18.6	40
Woman/feminine	29.3	63
Nonbinary +	52.1	112
Born in the US	87.4	188
Age	24.4	3.2
Highest level of education
Bachelor’s or higher	41.9	90
Postsecondary, no degree	35.8	77
High school, GED, or lower	22.3	48
Income sources
Traditional employment	43.3	93
Other sources only	53.5	115
No income	3.3	7
Ever incarcerated	14.6	31
Substance use, past 6 months	%	n
Heavy episodic drinking	52.8	112
Club drug use	50.5	108
Tobacco use
Never	46.7	100
Once or twice	16.8	36
Monthly	8.9	19
Weekly	8.4	18
Daily or almost daily	19.2	41
Marijuana use
Never	23.0	49
Once or twice	10.8	23
Monthly	12.7	27
Weekly	17.8	38
Daily or almost daily	35.7	76
Mental health	%/M	n/SD
Suicidal ideation, past 6 months	32.5	68
Depressive symptoms, current, possible range: 0–60	26.2	12.9
Gender-related PTSD symptoms, past 6 months, possible range: 0–4	1.4	1.5
Structural needs, past 6 months
Number of structural needs, possible range: 0–9	2.4	2.3
Number of unmet structural needs, possible range: 0–9	1.2	1.7

Approximately half of participants reported heavy episodic drinking (52.8%) and club drug use (50.5%) in the past six months. Tobacco use was common; only 46.7% reported no use in the past 6 months, and 19.2% reported daily tobacco use. Similarly, 23.0% of participants reported no use of marijuana over the past 6 months, and 35.7% reported using marijuana daily. Nearly a third of participants reported suicidal ideation in the past 6 months (32.5%). The mean depressive symptom score was 26.2 (SD = 12.9), and the mean gender-related PTSD symptom score was 1.4 (SD = 1.5). To ensure that conclusions about health outcomes drawn from this study apply across gender subgroups, we conducted chi-square and ANOVA tests between gender and each outcome. The only significant difference was in suicidal ideation; 41.3% of nonbinary + participants reported suicidal ideation compared to 25.8% of women/feminine participants and 18.4% of men/masculine participants (χ² = 8.53, p = 0.014).

The number of structural needs ranged from 0 to 9 (M = 2.4, SD = 2.3) and unmet structural needs ranged from 0 to 8 (M = 1.2, SD = 1.7) over the past 6 months (Table 2). The most common structural need was health insurance (45.1%) followed by food assistance (43.7%) and housing (34.0%). The most common unmet structural need was legal assistance (19.1%), with 82.0% of participants who reported needing legal assistance also reporting that they did not receive legal assistance. Additionally, over a third of participants who reported needing housing (37.0%), job training (78.0%), disability benefits (66.7%), utility payment assistance (55.8%), unemployment benefits (80.0%), and crisis support (37.9%) did not receive these services.

Table 2.

Structural needs and unmet structural needs among trans young adult in the San Francisco Bay Area (n = 215)

Structural need	Had need		Need unmet
	n	%	n	% of sample	% of those with need
Housing	73	34.0	27	12.6	56.3
Food assistance	94	43.7	20	9.3	28.7
Job training	41	19.1	32	14.9	48.8
Disability benefits	51	23.7	34	15.8	62.8
Health insurance	97	45.1	28	13.0	35.1
Utility payments	43	20.0	24	11.2	65.1
Unemployment benefits	45	20.9	36	16.7	53.3
Legal assistance	50	23.3	41	19.1	72.0
Crisis support	29	13.5	11	5.1	37.9

Demographic Patterns of Structural Needs

Table 3 presents bivariate comparisons in the number of structural needs and unmet structural needs. The number of structural needs and number of unmet structural needs significantly differed by participants’ highest level of education and income source. Participants who had some postsecondary education had an average of 2.9 (SD = 2.5) structural needs compared to 1.8 (SD = 1.9) for those with a college degree (p = 0.003), and an average of 1.6 (SD = 2.1) unmet structural needs compared to 0.9 (SD = 1.3) (p = 0.041). Participants who only had nontraditional sources of income had an average of 3.1 (SD = 2.5) structural needs compared to 1.6 (SD = 1.7) for those with a traditional employment (p < 0.001) and an average of 1.5 (SD = 2.0) unmet structural needs compared to 0.8 (SD = 1.3) (p = 0.014).

Table 3.

Number of structural needs and number of unmet structural needs by demographic characteristics (n = 215)

Demographic characteristics	Number of structural needs			Number of unmet structural needs
	M	SD	ANOVA p value	M	SD	ANOVA p value
Education			0.003*			0.041*
College +	1.8	1.9		0.9	1.0.3
Postsecondary, no degree	2.9	2.5		1.6	2.1
HS/GED or less	2.8	2.4		1.1	1.6
Income sources			< 0.001*			0.014*
Traditional employment	1.6	1.7		0.8	1.3
Other income source only	3.1	2.5		1.5	2.0
No income	2.6	2.4		1.4	1.8
Ever incarcerated			< 0.001*			0.056
No	4.0	2.7		1.7	2.0
Yes	2.1	2.1		1.1	1.7
Race/ethnicity			0.006*			0.342
White	2.2	2.1		1.0	1.6
Black/African American	3.2	2.5		1.3	1.6
Latino/a/x	2.5	2.6		1.5	2.1
Asian/Pacific Islander	2.4	2.5		1.1	1.7
Native American	6.5	1.7		2.8	2.2
Multiracial/multiethnic	2.2	1.6		1.2	1.4
Gender			0.956			0.968
Man/masculine	2.4	2.2		1.2	2.0
Woman/feminine	2.4	2.5		1.1	1.7
Nonbinary +	2.5	2.2		1.2	1.6
Born in the US			0.404			0.145
No	2.8	2.6		1.6	2.1
Yes	2.4	2.3		1.1	1.7
	r		p value	r		p value
Age	0.03		0.681	0.06		0.380

^*Significant at α = 0.05

Additionally, incarceration history and race/ethnicity were associated with number of structural needs. Participants who had ever been incarcerated had an average of 4.0 (SD = 2.7) structural needs compared to 2.1 (SD = 2.1) for those who had not (p < 0.001). Native American participants had the highest average number of structural needs at 6.5 (SD = 1.7) among all racial/ethnic groups, followed by Black/African American participants at 3.2 (SD = 2.5). White and multiracial/multiethnic participants reported the lowest number of structural needs at 2.2 (SD = 2.1). Incarceration history and race/ethnicity were not associated with unmet structural needs, and there was no relationship between gender, being born in the US, or age and either the number of structural needs or the number of unmet structural needs.

Structural Needs and Health Outcomes

In bivariate logistic regression models, the number of structural needs was associated with higher odds of daily tobacco use (OR 1.18, 95% CI: 1.02–1.36), daily marijuana use (OR 1.22, 95% CI: 1.07–1.38), and suicidal ideation (OR 1.20, 95% CI: 1.05–1.36) (Table 4). Furthermore, the number of structural needs was associated with depressive symptoms (β 1.70, 95% CI: 0.98–2.42). The association between the number of structural needs and daily marijuana use, suicidal ideation, and depressive symptoms persisted when adjusting for gender, education, employment, race/ethnicity, and incarceration history. In the adjusted regression models, each additional structural need was associated with a 29% increase in the odds of daily marijuana use (95% CI: 1.10–1.49), an 24% increase in the odds of suicidal ideation (95% CI: 1.06–1.45), and a 1.80-point increase in depressive symptoms (95% CI: 0.96–2.63).

Table 4.

Odds of substance use and mental health outcomes by number of structural needs and number of unmet structural needs (n = 215)

Health outcome	Number of structural needs				Number of unmet structural needs
	OR	95% CI	aORa	95% CI	OR	95% CI	aORb	95% CI
Tobacco	1.08	0.96–1.22	1.04	0.91–1.20	1.09	0.93–1.28	1.08	0.91–1.27
Daily tobacco	1.18	1.02–1.36*	1.11	0.92–1.33	1.06	0.88–1.28	1.04	0.85–1.26
Heavy drinking	0.98	0.87–1.11	0.99	0.86–1.14	1.01	0.86–1.18	1.02	0.89–1.20
Marijuana	1.04	0.90–1.20	1.10	0.93–1.31	1.07	0.88–1.30	1.10	0.90–1.34
Daily marijuana	1.22	1.07–1.38*	1.29	1.10–1.49*	1.28	1.08–1.51*	1.30	1.10–1.55*
Club drugs	1.02	0.91–1.15	1.01	0.88–1.16	0.99	0.85–1.16	1.00	0.85–1.17
Suicidal ideation	1.20	1.05–1.36*	1.24	1.06–1.45*	1.18	0.99–1.39	1.14	0.96–1.37
	β	95% CI	aβa	95% CI	β	95% CI	aβb	95% CI
Depressive symptoms	1.70	0.98–2.42*	1.80	0.96–2.63*	2.07	1.10–3.05*	2.00	1.00–3.00*
Gender-related PTSD symptoms	0.09	0.00–0.18	0.09	-0.02–0.19	0.10	–0.02–0.22	0.10	-0.02–0.22

^aModel controls for education, employment, ethnicity, incarceration history, and gender

^bModel controls for education and employment

^*Significant at α = 0.05

In bivariate regression models, the number of unmet structural need was associated with higher odds of daily marijuana use (OR 1.28, 95% CI: 1.08–1.51) and depressive symptoms (β 2.07, 95% CI: 1.10–3.05). In adjusted models, each additional unmet structural need was associated with 30% higher odds of daily marijuana use (95% CI: 1.10–1.55) and 2.00-point increase in depressive symptoms (95% CI: 1.00–3.00).

Mediation Analysis

Given the bivariate associations between (a) the number of structural needs and daily tobacco use, daily marijuana use, suicidal ideation, depressive symptoms, and gender-related PTSD symptoms and (b) the number of unmet structural needs and daily marijuana use and depressive symptoms above, we first fit unadjusted regression models to determine which sociodemographic characteristics were also associated with these outcomes. We found associations between race/ethnicity and daily marijuana use and gender-related PTSD symptoms; however, because of the small sample sizes within particular racial/ethnic groups, we were unable to test whether structural needs mediated these relationships. However, we did find that incarceration history was positively associated with daily tobacco use (OR 2.90, 95% CI: 1.25–6.69) and having nontraditional sources of income only (vs. traditional employment) was associated with depressive symptoms (β 4.36, 95% CI: 0.84–7.88) (Table 5).

Table 5.

Total, direct, and total indirect effects of select demographic characteristics on health outcomes through structural needs

	n	Total effect		Direct effect		Total indirect effect
		OR	95% CI	OR	95% CI	OR	95% CI
Effect of incarceration history on daily tobacco use through number of structural needs	212	2.88	1.25–6.65*	2.31	0.95–5.60	0.23	-0.04–0.64
		β	95% CI	β	95% CI	β	95% CI
Effect of income sourcea on depressive systems through number of structural needs	208	4.41	0.87–7.94*	1.89	-1.66–5.45	2.51	0.99–4.04*
Effect of income sourcea on depressive systems through number of unmet structural needs	208	4.41	0.87–7.94*	3.03	-0.41–6.48	1.37	0.23–2.52*

^aOther sources of income only vs. traditional employment

^*Significant at α = 0.05

In examining the effect of incarceration history on daily tobacco use through number of structural needs, we did not find evidence for mediation. The total indirect effect was not significant (OR 1.26, 95% CI: 0.97–1.81), and neither was the direct effect of incarceration history on daily tobacco use when controlling for number of structural needs (OR 2.31, 95% CI: 0.95–5.60). However, both the number of structural needs and the number of unmet structural needs completely mediated the relationship between having nontraditional sources of income only (vs. a traditional job) and depressive symptoms. These models did not include the 7 participants who reported having no income sources. For the number of structural needs, the coefficient for the total indirect effect was 2.51 (95% CI: 0.99–4.04), and for the number of unmet structural needs, the coefficient was 1.37 (95% CI: 0.23–2.52).

Discussion

Findings highlight the salience of structural vulnerability in the lives of trans young adults. The degree to which participants reported unmet needs for housing, job training, disability benefits, utility payment assistance, unemployment benefits, legal assistance, and crisis support signals that the systems designed to meet these needs in the SF Bay Area are inadequately reaching this population. The SF Bay Area has historically been considered a safe and desirable setting for trans people due to relatively inclusive normative social attitudes and public health policies towards trans people [62]. Our findings reinforce prior research suggesting that despite this history, trans people in the SF Bay Area experience structural vulnerabilities that contribute to and widen health inequities [37]. In the absence of interventions that are directly aimed at reducing structural vulnerability by redistributing resources such as housing, money, and legal support, these inequities are likely to become more entrenched with rising gentrification and cost of living [62, 63].

As expected, participants with lower levels of education, less stable income sources, and a history of incarceration reported higher numbers of structural needs and unmet structural needs [40]. While our finding that these sociodemographic characteristics are associated with structural needs is unsurprising, it is reflective of the diminishing avenues for economic mobility and financial viability among young adults [64–69]. It is within the context of an economic era characterized by decreasing worker protections, dismantling of social welfare programs, and increasing surveillance and criminalization that trans young adults experience high rates of discrimination, exclusion, and maltreatment in schools, employment, public benefits, and healthcare [35, 70–75]. This discrimination, exclusion, and maltreatment further constrain economic opportunity and motivates engagement in criminalized labor (e.g., sex work and drug sales) [31, 75, 76]. Combined with policing practices and other state surveillance systems that target communities of color and poor communities, this results in the overrepresentation of trans people overall and trans women of color specifically in the criminal-legal system [77–80]. The relationship between education, income source, and incarceration history and structural needs found in this study therefore reflects the interconnectedness of these structural vulnerabilities [39].

It is particularly noteworthy that participants who reported working outside of the formal economy only—i.e., only working in contract/gig work, receiving government benefits, receiving financial support from partners or family, engaging in sex work, or other nontraditional roles—had higher numbers of reported structural needs and unmet structural needs than those with traditional income sources. These participants may experience intersecting forms of stigma that render them more structurally vulnerable than those in full- or part-time work (e.g., ableism and racism). Alternatively, this finding may reflect how access to resources such as rental housing, credit, and private health insurance is often contingent upon demonstrating proof of income from traditional employment. Over the past decade, participation in the gig economy has rapidly increased, particularly among young adults [81]. Initially, this shift was perceived as promising for workers who experienced discrimination in traditional employment including people of color, immigrants, women, and disabled people as it was by nature self-directed, flexible, and easy to enter [82–84]. Trans people living in places with strong legal protections for trans individuals such as the SF Bay Area have described how nondiscrimination policies do not eliminate experiences of cisgenderist harassment and discrimination in the workplace and how these experiences contribute to high rates of underemployment and unemployment for trans people [37, 85]. Our results suggest that structural needs may be more likely to go unmet for trans young adults who rely on informal or precarious forms of employment, contributing to health inequities [83].

Of note, there was no relationship between gender and number of structural needs or unmet structural needs. Research with trans adults typically centers trans women’s experiences [86], yet our findings suggest that young trans men and nonbinary + people in the SF Bay Area are comparably structurally vulnerable. Research and programming may therefore benefit from using a gender-inclusive approach in which trans women, trans men, and nonbinary + young adults have equal access to efforts to understand and address trans young adult’s structural needs [87]. We did find significant variation in the number of structural needs by race/ethnicity, though this finding should be interpreted cautiously due to the small sample size of Native American and Black/African American participants, the groups with the highest number of reported structural needs.

The number of structural needs was associated with daily marijuana use and suicidal ideation, and the number of unmet structural needs was associated with daily marijuana use when controlling for gender and other relevant sociodemographic covariates. Both the number of structural needs and the number of unmet structural needs mediated the relationship between having other sources of income only and depressive symptoms. Given the high prevalence of daily marijuana use, depression, and suicidality in this sample and in other studies of trans youth and young adults [4, 86, 88], these findings suggest that these inequities may be driven by structural vulnerability and therefore may be partially addressed by interventions that meet structural needs.

A variety of economic initiatives have emerged to address young adult’s structural needs relevant to our study population. California is currently piloting a Guaranteed Income program that will fund organizations who provide unconditional cash transfers to individuals who meet pre-determined requirements with priority to organizations serving former foster youth [89]. This program builds upon the successes of localized unconditional cash transfer programs, some of which are universally available to residents in particular communities and some of which are open to members of vulnerable populations [90, 91]. California lawmakers are also currently debating the development of unconditional cash transfer programs for low-income college students [92], and a coalition of Historically Black Colleges and Universities are embarking on an initiative to provide basic income support to their students [93]. Additionally, many trans-led organizations have used philanthropy to allocate financial assistance and other forms of support to trans people of color during the pandemic [94, 95]. For example, Trans* Activists for Justice and Accountability (TAJA’s Coalition) provides workshops, training, and service linkage to ensure community members have access to employment, education, and housing resources. Our findings suggest that evaluators should examine the extent to which these government, educational, and community-led programs reach the most vulnerable trans young adults, address their structural needs, and impact mental health and substance use. Furthermore, organizations providing mental health and substance use programming to trans communities should competently address employment and other structural vulnerabilities as part of comprehensive treatment or provide linkage to existing services.

Limitations

Our findings are limited by the cross-sectional study design, use of self-report measures, and convenience sampling. Although we found evidence of the mediating effect of structural needs in the relationship between income source and depressive symptoms, the cross-sectional study design prevents conclusions regarding causation. Self-report measures for substance use and mental health outcomes are subject to social desirability bias as participants may be less likely to report behaviors or symptoms because of the stigma attached to mental health and substance use, especially in an interviewer-administered interview. However, using self-report measures for structural needs is a strength of this study as it allows for inclusion of a greater variety of needs and may be more reflective of participants’ experiences of structural need than what could be extrapolated from typical survey measures. Our models relied on counting the number of structural needs and unmet structural needs, and we acknowledge that further research will benefit from a more refined computational approach. Finally, convenience sampling limits the generalizability of our findings. In particular, the participants in this study reported high levels of education that are not commensurate with those of the general trans adult population [96].

Conclusion

Structural needs and unmet structural needs are highly prevalent among trans young adults in the SF Bay Area. Findings suggest that addressing structural vulnerabilities among this population may be critical to reducing mental health and substance use inequities. Community- and government-led efforts to meet structural needs, including basic income schemes, may have an impact on mental health and substance use among trans young adults.