Minority Stress, Community Involvement, and Cognitive Decline in Rural SGM Older Adults: A Longitudinal SEM and Growth Curve Approach

Abstract

Sexual and gender minority (SGM) older adults in rural areas face elevated cognitive health risks shaped by structural stigma, yet the pathways underlying these inequities are poorly understood. This study investigated how lifetime rural exposure, minority stress exposure, and community involvement influenced cognitive outcomes among SGM older adults in the U.S. South. Using data from a three-wave longitudinal panel of 1,259 participants, we applied structural equation modeling (SEM) to examine direct, indirect, and total effects of these factors on cognitive change scores. We then specified a generalized structural equation model (GSEM) with a Poisson latent growth curve to model repeated cognitive problem counts over time, addressing the count distribution of the outcome. SEM results indicated that lifetime rural exposure, gender minority status, and higher educational attainment were associated with increased minority stress exposure, whereas racial minorities and older participants reported lower minority stress. Minority stress was positively associated with community involvement, reflecting adaptive coping response, and mediated the relationships between gender minority status, education, and community involvement. However, neither minority stress nor community involvement predicted cognitive change when modeled as a difference score. In contrast, the GSEM growth model demonstrated that higher minority stress exposure significantly accelerated cognitive problem growth over time, while chronic conditions exacerbated and higher education slowed decline. Findings highlight that minority stress exerts cumulative, longitudinal effects on cognitive trajectories among SGM older adults. Community participation may foster engagement but cannot offset the neurocognitive consequences of structural stigma.

Introduction

Cognitive aging is a critical public health challenge with profound consequences for individuals and societies worldwide. Globally, an estimated 10–20% of older adults experience mild cognitive impairment or early-stage dementia, conditions that significantly compromise quality of life, autonomy, and social participation.^1–4 Cognitive decline has been strongly linked to social determinants of health, including socioeconomic disadvantage, structural discrimination, and restricted access to supportive networks.^5–10 These social determinants shape risk factors across the life course, including stress exposure, health behaviors, and opportunities for cognitive stimulation and engagement.^11–14 Yet, despite an expanding literature on cognitive aging, sexual and gender minority (SGM) adults remain largely overlooked in research on cognitive health disparities.^15–21

SGM individuals face disproportionate exposure to chronic stressors resulting from stigma, discrimination, and marginalization over their lifetimes.^22–27 According to the Minority Stress Model, SGM people endure distinct and additive stressors beyond those experienced by the general population, including experiences of rejection, expectations of stigma, and internalized negativity.²² Such stressors can have cumulative negative impacts on health across the life course. For instance, recent studies have linked minority stress to depressive symptoms, physical health problems, and even premature mortality.^16,28–31

Emerging research has begun to examine how minority stress shapes cognitive aging, though this literature is still relatively sparse. Several studies have suggested that chronic exposure to discrimination and stigma may accelerate cognitive decline by increasing allostatic load, disrupting neural pathways, disrupting sleep, and contributing to sustained dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis.^32–37 For example, one study found sleep problems mediated the relationship between major problems at work and reports of cognitive symptoms consistent with mild cognitive impairment among older SGM adults.¹⁰ Another study showed that internalized ageism, a specific form of social stigma, predicted worse memory performance over time, demonstrating the relevance of internalized and anticipated stigma processes for cognitive trajectories.^38–40 These studies indicate that minority stressors, whether rooted in sexual/gender stigma or other marginalized identities, can negatively influence cognitive functioning through pathways of chronic stress, social isolation, and reduced opportunities for cognitive engagement. However, empirical research directly linking minority stress processes to cognitive aging in diverse SGM older adults – particularly those living in structurally disadvantaged rural environments – remains limited.²⁰ Addressing this gap is essential for advancing a comprehensive understanding of health inequities in cognitive aging.

Geographic context further shapes SGM health inequities. Rural environments often present distinct structural challenges, including health professional shortages, lower socioeconomic resources, underdeveloped public health infrastructure, and limited availability of culturally competent care.^41–43 In the U.S. South, these challenges are compounded by hostile policy environments and conservative social norms, contributing to heightened stigma, fewer supportive services, and more profound social isolation for SGM older adults.²⁰ Rural SGM residents may therefore face a “double disadvantage” of geographic and identity-based marginalization, exposing them to chronic minority stress while restricting their access to protective resources.^44,45 While a handful of studies have described the mental health burdens of rural SGM people, few have examined their cognitive health, despite evidence that social isolation and chronic stress are critical risk factors for cognitive decline.^2,44,45

Community involvement is widely theorized to influence the effects of chronic stress by promoting social connectedness, fostering collective identity, and providing tangible and emotional support.^46,47 For SGM older adults, community engagement – whether through LGBTQ+ peer networks, chosen family, advocacy groups, or faith communities – may be a vital resource that counters minority stress and supports cognitive resilience.^16,48,49 Yet findings on its effectiveness remain mixed. Some work has shown that community networks improve mental health and quality of life among SGM populations,⁵⁰ but other studies suggest these networks may be insufficient in the face of structural exclusion, particularly in rural regions.⁴⁵ While social support is protective for SGM mental health, little empirical work has tested whether community engagement influences cognitive outcomes.^18,48 Moreover, no published studies to date have explicitly examined whether community involvement mediates minority stress effects on cognitive health among rural SGM older adults. This is a critical gap, given both the importance of social capital for successful aging and the challenges of sustaining such capital in rural, under-resourced settings.

Furthermore, an intersectional perspective is essential for understanding how rural SGM adults experience cognitive health risk. Intersectionality theory recognizes that people are situated within multiple, interlocking systems of oppression, including structural racism, sexism, classism, and cisheteronormativity.^51,52 For SGM people of color, or those with low socioeconomic resources, the rural U.S. South may represent a context of particularly deep disadvantage due to longstanding histories of racial violence, economic disinvestment, and restrictive social policies.^41,53 These intersecting structural forces likely shape patterns of stress exposure, community participation, and cognitive health across the life course. Despite strong theoretical rationale, empirical research rarely tests these intersectional dimensions of cognitive aging among rural SGM populations.

Taken together, these theoretical frameworks suggest a pressing need to examine the pathways linking rural life course exposure, minority stress, community involvement, and cognitive health among SGM older adults. The present study leverages a life course perspective, drawing on the Minority Stress Model, social capital theory, and intersectionality frameworks, to address four key research questions: (1) Does rural life course exposure influence minority stress among SGM older adults? (2) Does minority stress relate to community involvement? (3) Do these factors shape cognitive difficulties? and (4) Does community involvement mediate these relationships? In addition to providing evidence for interventions in the rural U.S. South, this study offers lessons relevant to other global contexts where marginalized communities face similar intersecting disadvantages.

Methods

Study Design and Sample

This study used data from the LGBTQ+ Social Networks, Aging, and Policy Study, a longitudinal, multi-wave survey assessing social, community, and health outcomes among a diverse sample of SGM adults in the U.S. South. The study aimed to explore how cumulative exposures affect health outcome in later life.

Participants were recruited through a combination of purposive and snowball sampling strategies, leveraging LGBTQ+ community organizations, pride events, health centers, and online platforms. Inclusion criteria included identifying as LGBTQ+, being age 50 or older at baseline, and residing in Alabama, Georgia, North Carolina, or Tennessee. The study did not employ any purposive oversampling for rurality. Data were collected through structured online surveys administered at three waves over approximately six years, beginning in 2020.

The analytic sample for the structural equation modeling (SEM) analyses included 1,259 participants with complete responses on measures of lifetime rural exposure, minority stress exposure, community involvement, and cognitive change. For growth curve modeling, a subsample of 962 participants with cognitive problem data at all three waves was used to ensure valid estimation of change over time. This analytic strategy balanced sample size and data completeness to optimize power while preserving temporal ordering for longitudinal change models.

Measures

Lifetime Rural Exposure.

Lifetime rural exposure was measured retrospectively as a cumulative index ranging from 0 to 5, reflecting the number of developmental life-course stages in which participants resided in a rural or small town area. Specifically, participants were asked to “describe the areas where you lived when you were: under18 years old, 18 to 29 years old, 30 to 49 years old, 50 or older, and area lived in the last year.” Response options included “city (downtown); city (neighborhood); suburban areas; small town, and rural.” Responses marked as “small town” and “rural” were identified as “rural” in our study. Missing responses were indicated as 0s. A cumulative count approach was chosen to align with life-course models of exposure accumulation, capturing the breadth of rural living across multiple developmental windows.

Minority Stress Exposure.

Minority stress exposure was measured retrospectively as a cumulative index ranging from 0 to 65, summarizing self-reported reported minority stress exposures across life stages. The measure is not time-varying, with the intention of capturing accumulated structural and psychosocial disadvantage. Specifically, respondents were asked, “The following is list of experiences that LGBT people sometimes have. Please read each one carefully, and then respond to the follow question: When in your life have you experienced this, if ever?” The experiences included: having very few people you can talk to about being LGBT; watching what you say and do around heterosexual people; hearing about LGBT people you know being treated unfairly; hearing about LGBT people you don’t know being treated unfairly; hearing about hate crimes (e.g., vandalism, physical or sexual assault) that happened to LGBT people you don’t know); hearing someone make jokes about LGBT people; hearing politicians say negative things about LGBT people; being verbally harassed by strangers because you are LGBT; being verbally harassed by people you know because you are LGBT; people laughing at you or making jokes at your expense because you are LGBT; feeling like you don’t fit in with other LGBT people; people staring at you when you are out in public because you are LGBT; feeling invisible in the LGBT community because of your gender expression; avoiding talking about your current or past relationships when you are at work; hiding part of your life from other people; feeling like you don’t fit into the LGBT community because of your gender expression; difficulty finding clothes that you are comfortable wearing because of your gender expression; being misunderstood by people because of your gender expression; being raped or sexually assaulted because you are LGBT; having object thrown at you because you are LGBT; having property (like a car, houses, locker) damaged because you are LGBT. Response options included: This never happened to me; this happened when I was 18 or younger; this happened when I was about 18 to 29 years old; this happened to me when I was about 30 to 49 years old; this happened when I was 50 or older; this happened in the last year. Missing responses were indicated as 0s. A cumulative index was justified based on minority stress theory, which emphasizes the additive impact of repeated stress experiences over time on health and well-being. Although the items included here are not drawn from a single validated scale, they reflect widely studied domains of minority stress and were selected based on theoretical relevance and prior empirical studies.^{9,54, 55}

Community Involvement.

Community involvement was measured retrospectively as a cumulative index (0–99) capturing participation in community activities across multiple life-course periods (i.e., under 18; 18 to 29 years old; 30 to 49 years old; 50 or older; in the last year). Specifically, respondents were asked, at each life stage, how often (i.e., never; rarely, once or twice over the year; sometimes, once every other month; often, once a month or more; not applicable) they participated in the following: a professional association or business group; organized sports team; neighborhood or block association; a service organization (like the Rotary, Red Cross, or a food bank); a religious organization like a church or synagogue; a local performing arts or cultural organization (like a choir or amateur theater); an LGBTQ organization; a political organization; a student or school organization; another kind of organization. Higher scores indicated greater or more sustained community participation. Missing responses were indicated as 0s. This measure was standardized (z-score) prior to modeling to facilitate interpretability of coefficients and comparability with other predictors on different scales.

In this study, we operationalized exposures (i.e., minority stress, rural residency, and community involvement) as cumulative measures spanning multiple developmental periods. This approach draws on life course epidemiology frameworks that examine how “accumulation of risk” models may shape later-life health outcomes.^56,57 While we acknowledge that more contemporary socio-ecological models recognize that minority stressors may ebb and flow across historical, developmental, and contextual domains,^23,53 our cumulative approach reflects a pragmatic strategy to model chronic exposure burden. We do not assume that these effects are strictly linear or irreversible but rather treat the accumulation of stressors as a meaningful proxy for the long-term embedding of disadvantage.

Cognitive Problems.

Cognitive problems were assessed through prospective self-reported counts of cognitive difficulties at each survey wave. Specifically, at each wave, responded were asked if they have difficulty doing each of the following: remembering things that have happened recently; recalling questions a few days later; remembering my address and phone number; remembering what day and month it is; remembering where things are usually kept; remembering where to find things that have been put in a difference place than usual; knowing how to work familiar machines around the house; learning to use a new gadget or machine around the house; learning new things in general; following a story in a book or on TV; making decisions on everyday matters; handling money for shopping; handling financial matters, that is, the pension or dealing with the bank; handling other everyday arithmetic problems, such as knowing how much food to buy, knowing how long between visits from family or friends; understand what’s going on and reasoning through things. Higher counts reflected greater perceived cognitive concerns.

To evaluate cognitive change, two modeling strategies were justified. First, a log-transformed difference score from wave 1 to wave 3, used in SEM, to reduce skewness in change scores and support path modeling assumptions of continuous outcomes. Second, a growth curve model treating the counts directly as repeated measures, using generalized structural equation modeling (GSEM), to more precisely estimate developmental change in cognitive problem trajectories across waves while accounting for the count distribution of the outcome.

Sociodemographic Covariates.

Sociodemographic controls included gender minority status (coded 0 = non-minority, 1 = minority, i.e., transgender man, transgender woman, or gender nonconforming), race (0 = white, 1 = racial minority), educational attainment (0 = less than bachelor’s degree, 1 = bachelor’s degree or higher), and chronic health conditions (0 = no reported chronic conditions, 1 = at least one chronic condition [i.e., Diabetes or high blood sugar, asthma or another breathing issue, arthritis or rheumatism, cancer]). These variables were selected based on prior literature demonstrating their associations with minority stress, community participation, and cognitive health disparities.^9,10,15

Statistical Analysis

A two-stage analytic plan was implemented. First, a structural equation model (SEM) was specified to test pathways among lifetime rural exposure, minority stress exposure, community involvement, and cognitive change. SEM was selected because it allows simultaneous estimation of direct, indirect, and total effects within a single framework, which is well-suited for evaluating theorized mediation pathways. Cognitive change was modeled as a log-transformed difference score to address skewness and meet the SEM assumptions of continuous, normally distributed outcomes. SEM was estimated with maximum likelihood with missing values (MLMV) to minimize bias from incomplete data and maximize statistical power.

Second, a GSEM was estimated to examine cognitive problems as repeated count outcomes across three waves. GSEM was selected to account for the non-normal, discrete distribution of cognitive problem counts. A Poisson growth curve framework was implemented with a linear slope factor (loadings set to 0, 1, and 2), enabling the estimation of within-person change over time. Predictors of the slope included rural exposure, minority stress, community involvement, gender minority status, race, education, and chronic conditions. Poisson family was chosen for the primary analysis, with the model monitored for evidence of overdispersion to justify possible negative binomial extension.

Model fit for the SEM was assessed using standard indices including chi-square, root mean square error of approximation (RMSEA), and comparative fit index (CFI), justifying these as common SEM criteria for evaluating global model adequacy.^58,59 GSEM model adequacy was evaluated based on convergence criteria, model log-likelihood, parameter estimates, and interpretability.

The analytic sample varied across models based on the nature of the outcome and the data requirements for estimation. The SEM utilized a log-transformed difference score of the cognitive problems between Wave 1 and Wave 3. This allowed us to include participants with missing data at Wave 2, resulting in a larger sample. In contrast, the GSEM employed a Poisson growth curve to model count data across all three waves, which required complete observations at each wave. As a result, the GSEM analysis was limited to participants with full cognitive data across waves, yielding a smaller sample size.

All analyses were conducted in Stata 17.0 (StataCorp). The full conceptual framework is illustrated in Figure 1.

Figure 1. Conceptual SEM and GSEM Pathways Among Life Course Rural Exposure, Minority Stress, Community Involvement, and Cognitive Outcomes.

Conceptual and empirical pathways linking lifetime rural exposure, minority stress exposure, community involvement, and cognitive outcomes among SGM older adults living in the U.S. South. Solid black arrows indicate significant estimated pathways, while dashed gray arrows indicate pathways tested but not statistically significant in the final SEM and GSEM models.

Results

Descriptive Statistics

The analytic sample included 1,259 respondents with data on lifetime rural exposure, minority stress exposure, community involvement, and cognitive change (see Table 1). Repeated measures of cognitive problem counts were available for 962–1,208 respondents across three waves.

Table 1.

Sample Characteristics

Characteristic	Mean or % (SD)
Age (years)	Wave 1: 59.28 (6.37); Wave 2: 60.72 (6.38); Wave 3: 62.09 (6.44)
Gender Minority Status
Gender Minority	10.33%
Not Gender Minority	89.67%
Race
Racial Minority	14.06%
White	85.94%
Education
Bachelor’s Degree or Higher	71.22%
Less than Bachelor’s	28.78%
Chronic Health Conditions
No Diagnosed Chronic Conditions	42.18%
Has Diagnosed Chronic Conditions	57.82%
Lifetime Rural Exposure
Never Lived in Rural	42.49%
One Life Course Stage	25.73%
Two Life Course Stages	15.73%
Three Life Course Stages	8.74%
Four Life Course Stages	4.37%
Five Life Course Stages	2.94%
Minority Stress Exposure	19.77 (16.07)
Community Involvement	28.37 (22.08)
Cognitive Problems
Wave 1	.96 (1.79)
Wave 2	.98 (1.78)
Wave 3	1.10 (2.25)

Descriptive characteristics of the analytic sample (N = 1,259) of SGM older adults in the U.S. South from the LGBTQ+ Social Networks, Aging, and Policy Study. Means and standard deviations are reported for continuous variables; proportions are shown for categorical variables.

Cognitive problems were reported as count variables at each wave, with a mean baseline of approximately 0.39 cognitive problems (exp(–0.93) on the log scale from the growth model) and substantial individual variation.

Structural Equation Model Results

We first estimated a structural equation model examining pathways from lifetime rural exposure to minority stress exposure, community involvement, and cognitive change, with cognitive change modeled as a log-transformed outcome. The SEM included direct, indirect, and total effects, while controlling for gender minority status, race, educational attainment, chronic conditions, and age.

Predicting Minority Stress Exposure (Table 2).

Table 2.

SEM Paths to Minority Stress Exposure

Predictor	Estimate	SE	Z	P	95% CI
Rural Exposure	2.21	.35	6.79	<0.001***	1.57, 2.84
Gender Minority	4.12	1.42	2.90	0.004**	1.34, 6.90
Racial Minority	−7.08	1.26	−5.63	<0.001***	−9.54, −4.62
Bachelor’s Degree	5.92	0.97	6.13	<0.001***	4.03, 7.81
Chronic Conditions	1.07	.88	1.21	0.23	−0.66, −0.14
Age	−0.29	0.08	−3.65	<0.001***	−0.45, −0.14

Structural equation model estimates for predictors of lifetime minority stress exposure among sexual and gender minority (SGM) older adults in the rural U.S. South (N = 1,259). Coefficients represent direct effects of lifetime rural exposure and sociodemographic covariates on cumulative minority stress experiences reported across the life course. Standard errors (SE), z-statistics, p-values, and 95% confidence intervals (CI) are reported.

Higher lifetime rural exposure was significantly associated with increased minority stress exposure (B = 2.20, SE = 0.32, z = 6.79, p < .001, 95% CI: 1.57 to 2.84). Gender minority respondents also experienced higher lifetime minority stress exposure (B = 4.12, SE = 1.42, z = 2.90, p= 0.004, 95% CI: 1.34 to 6.90). Racially minoritized respondents reported significantly lower lifetime minority stress (B = −7.08, SE = 1.26, z = −5.63, p < 0.001, 95% CI: −9.54 to −4.62). Respondents with a bachelor’s degree or higher showed higher lifetime minority stress as well (B = 5.92, SE = 0.97, z = 6.13, p < 0.001, 95% CI: 4.03 to 7.81). Chronic health conditions were not significantly associated with minority stress (B = 1.07, SE = 0.88, z = 1.22, p = 0.23), while older age was inversely related to minority stress exposure (B = −.29, SE = .08, z = −3.65, p < .001).

Predicting Community Involvement (Table 3).

Table 3.

SEM Paths to Community Involvement

Predictor	Estimate	SE	Z	P	95% CI
Minority Stress	0.008	0.002	4.21	<0.001***	0.004, 0.011
Rural Exposure	0.044	0.022	2.08	0.038*	0.003, 0.085
Gender Minority	−0.009	0.099	−0.10	0.918	−0.188, 0.169
Racial Minority	0.167	0.081	2.06	0.040*	0.008, 0.326
Bachelor’s Degree	0.406	0.063	6.47	<0.001***	0.283, 0.529
Chronic Conditions	−0.016	0.057	−0.27	0.783	−0.095, 0.126
Age	−0.002	0.005	−0.37	0.712	−0.01, 0.012

Structural equation model estimates for predictors of lifetime community involvement among SGM older adults in the rural U.S. South (N = 1,259). Coefficients represent direct effects of minority stress exposure, lifetime rural exposure, and sociodemographic covariates on cumulative community participation across the life course. Standard errors (SE), z-statistics, p-values, and 95% confidence intervals (CI) are reported.

Higher minority stress exposure was significantly associated with greater community involvement (B = 0.0076, SE = 0.0018, z = 4.21, p < 0.001, 95% CI: 0.0041 to 0.0112). Lifetime rural exposure also positively predicted community involvement (B = 0.044, SE = 0.022, z = 2.08, p = .038). Racially minoritized respondents reported significantly higher community involvement (B = 0.167, SE = 0.081, z = 2.06, p = 0.041, 95% CI: −0.327 to −0.007). Education remained a strong predictor of community involvement (B = 0.406, SE = 0.063, z = 6.47, p < 0.001, 95% CI: 0.282 to 0.529). Other covariates, including chronic conditions and age, were not significant.

Indirect effects revealed that both gender minority status (B = 0.031, SE = 0.013, z = 2.39, p = 0.017) and educational attainment (B = 0.045, SE = 0.013, z = 3.49, p = 0.001) predicted greater community involvement indirectly through minority stress exposure.

Predicting Cognitive Change (log-transformed) (Table 4).

Table 4.

SEM Paths to Cognitive Change (log-transformed)

Predictor	Estimate	SE	Z	P	95% CI
Minority Stress	0.001	0.002	0.50	0.614	−0.002, 0.004
Community Involvement	−0.012	0.022	−0.53	0.593	−0.054, 0.031
Rural Exposure	0.004	0.015	0.26	0.796	−0.026, 0.034
Gender Minority	−0.103	0.068	−1.53	0.126	−0.237, 0.03
Racial Minority	−0.098	0.063	−1.57	0.116	−0.221, 0.027
Bachelor’s Degree	0.14	0.057	2.47	0.014*	0.029, 0.251
Chronic Conditions	0.054	0.041	1.33	0.184	−0.026, 0.134
Age	0.002	0.003	0.58	0.564	−0.004, 0.008

Structural equation model estimates for predictors of cognitive change, modeled as the log-transformed difference in cognitive problem counts between Wave 1 and Wave 3, among SGM older adults in the U.S. South (N = 1,259). Coefficients represent direct effects of minority stress exposure, community involvement, lifetime rural exposure, and sociodemographic covariates. Standard errors (SE), z-statistics, p-values, and 95% confidence intervals (CI) are reported.

No statistically significant predictors emerged for the log-transformed cognitive change variable. The direct effects of minority stress exposure (B = 0.00077, SE = 0.0015, z = 0.50, p = 0.614), community involvement (B = – 0.0116, SE = 0.0217, z = −0.53, p = 0.593), rural exposure (B = 0.0039, SE = 0.0149, z = 0.26, p = 0.796), and other covariates were all non-significant.

Indirect and Total Effects in the SEM.

Significant indirect effects were identified for pathways leading to community involvement. Minority stress acted as a mediator linking gender minority status and education to increased community involvement. No significant indirect or total effects were detected for cognitive change.

Generalized SEM Growth Model Results

Given that cognitive problems were reported as count variables across three waves, we next estimated a Poisson latent growth curve using GSEM, specifying an intercept and a linear slope across three time points, with predictors of the slope factor (Table 5). This approach better captures the longitudinal development of cognitive problems as a count process.

Table 5.

GSEM Poisson Growth Model: Predictors of Cognitive Problem Slope

Predictor	Estimate	SE	Z	P	95% CI
Minority Stress	0.005	0.002	2.20	0.028*	0.001, 0.009
Rural Exposure	0.005	0.021	0.25	0.802	−0.038, 0.046
Community Involvement	−0.005	0.020	−0.24	0.810	−0.045, 0.036
Gender Minority	−0.024	0.093	−0.26	0.794	−0.205, 0.158
Racial Minority	−0.083	0.043	−1.94	0.052	−0.167, 0.001
Bachelor’s Degree	−0.093	0.045	−2.06	0.039*	−0.183, 0.003
Chronic Conditions	0.274	0.064	4.25	<0.001***	0.148, 0.4
Age	−0.012	0.019	−0.65	0.514	−0.049, 0.025

Generalized structural equation model (GSEM) Poisson growth curve estimates for predictors of the slope of cognitive problem counts across three survey waves among SGM older adults in the U.S. South (N = 962). Coefficients represent effects of minority stress exposure, community involvement, lifetime rural exposure, and sociodemographic covariates on the rate of change in cognitive problems over time. Standard errors (SE), z-statistics, p-values, and 95% confidence intervals (CI) are reported.

Growth Factors.

At baseline, respondents had an estimated log count of cognitive problems of – 0.93 (SE = 0.075, p < 0.001), corresponding to approximately 0.39 cognitive problems on average (exp(–0.93)). Across time, the slope parameter was positive and significant (Wave 2 slope = 0.95, SE = 0.05; Wave 3 slope = 0.75, SE = 0.06; both p < 0.001), indicating that cognitive problems increased over time among respondents.

Predictors of Cognitive Problem Growth (Slope).

Higher minority stress exposure predicted significantly a faster growth rate of cognitive problems over time (B = 0.0045, SE = 0.0021, z = 2.20, p = 0.028. Chronic health conditions were also associated with worsening cognitive problems (B = 0.2735, SE = 0.064, z = 4.25, p < 0.001). Educational attainment was protective, associated with slower growth (B = −0.0931, SE = 0.045, z = −2.06, p = .039). Other predictors including rural exposure, community involvement, gender minority status, race, and age were not significant.

Variances and Covariances.

The variance of the slope factor was 0.25 (SE = 0.03), suggesting moderate individual heterogeneity in rates of cognitive change, while the intercept variance (Var = 1.94, SE = 0.19) reflected substantial variation in baseline cognitive problem counts.

Discussion

This study explored how lifetime rural exposure, minority stress, and community involvement influence cognitive health among sexual and gender minority (SGM) older adults living in the U.S. South. Guided by minority stress theory and intersectionality frameworks,^22,23,51,52 we applied structural equation modeling (SEM) and generalized structural equation modeling (GSEM) to assess whether community involvement could mediate the harmful effects of minority stress on cognitive difficulties. These findings contribute new evidence to the literature on cognitive health inequities in rural SGM older adults, highlighting theoretically and practically important patterns.

First, the SEM results demonstrated that greater lifetime rural exposure was strongly associated with elevated minority stress exposure, a finding that underscores the enduring health costs of growing up or living in structurally hostile social environments. SGM adults in rural areas – especially in the U.S. South – are often embedded within politically conservative environments characterized by limited legal protections, restricted access to affirming healthcare, and heightened experiences of stigma, discrimination, and rejection.^45,53 The accumulation of such exposures over the life course appears to produce a sustained burden of minority stress among rural SGM older adults, in line with minority stress theory.²²

Interestingly, education and gender minority status were also positively associated with minority stress exposure, suggesting that higher social visibility and educational mobility may heighten exposure to stigmatizing environments over time. In contrast, racially minoritized respondents reported lower cumulative minority stress, a pattern that may reflect underreporting, stress normalization, or different coping schemas developed across multiple marginalized identities. Older age was inversely associated with minority stress, consistent with research suggesting that some older adults experience reduced exposure to institutional settings where discrimination is common or develop more effective coping mechanisms with age.

Second, minority stress exposure was positively associated with community involvement. Although on its surface this association may appear counterintuitive, prior scholarship on resilience and SGM coping strategies suggests that adversity can drive individuals to mobilize social resources, seek solidarity, and build community ties.^48,50 These results support the idea that, faced with structural hostility, SGM older adults may turn to community networks as adaptive strategies to protect their well-being, even if those networks cannot fully address structural inequalities.

Education and rural exposure also predicted greater community involvement, suggesting that both access to resources and experiences of marginalization can motivate civic and social participation. However, the positive association between rural exposure and community engagement should not be overinterpreted as evidence of resilience alone; in many rural areas, community participation may be one of few available outlets for affirmation, even within constrained or stigmatizing environments.

Importantly, indirect effects indicated that gender minority status and education predicted higher community involvement indirectly through minority stress, highlighting community engagement as a response to – not protection from – stress exposure.

Third, community involvement was not significantly associated with cognitive change in the SEM, nor did it mediate the relationship between minority stress and cognitive outcomes. This challenges optimistic assumptions that social participation alone can offset the neurocognitive consequences of minority stress. Several explanations are possible. Measures of community involvement may not capture the quality, trust, or reciprocity of relationships, which might be more crucial for cognitive health.⁴⁶ Moreover, community participation in under-resourced rural contexts may be insufficient to address the structural disadvantages that drive stress and health inequities, such as limited healthcare access or persistent poverty. Participation alone may deliver some emotional benefit but cannot fully overcome the social and economic determinants of cognitive health.

The GSEM growth curve model, however, revealed that higher minority stress exposure significantly accelerated the rate of cognitive problem growth over time, even after accounting for community involvement and other covariates. This finding indicates that minority stress could operate as a cumulative process, shaping the trajectory of cognitive decline rather than a statis cross-sectional difference. Consistent with stress-process theory, chronic stress exposures may dysregulate neurobiological systems, elevate allostatic load, and progressively undermine cognitive functioning.⁶⁰

In contrast, educational attainment was protective, slowing cognitive decline, while chronic health conditions independently accelerated it. Race, gender minority status, and rural exposure were not significant in the longitudinal model, suggesting that their effects are mediated or superseded by cumulative stress and health status over time.

Collectively, these findings underscore the importance of modeling cognitive outcomes longitudinally, as static difference scores may obscure the dynamic, cumulative harm associated with minority stress across the life course.

The joint SEM and GSEM findings reveal a complex interplay between social disadvantage, adaptive engagement, and health decline. Minority stress emerges as both a driver of community involvement and a predictor of accelerated cognitive decline. Community participation, while adaptive, does not appear sufficient to mitigate the neurocognitive consequences of lifelong discrimination. These patterns reinforce the view that community-based resilience, though vital, cannot substitute structural and policy-level interventions to dismantle the root causes of stress exposure.

Although sample size limited formal multi-group analyses, patterns in exploratory analyses suggested that the effects of minority stress on cognitive trajectories may vary across subgroups, particularly by race and education. While race was not a significant predictor in the final models, intersectionality perspectives suggest that overlapping structural marginalizations, including racism, classism, and cisnormativity, may compound stress exposure and constrain access to protective resources.^51,52 Future research should intentionally power subgroup analyses to clarify these intersectional dynamics.

Several other limitations warrant attention. First, cognitive difficulties were assessed using self-reported items rather than objective clinical evaluations, which may introduce reporting biases. Second, while the study leverages a longitudinal panel, the timing of measurements does not fully address concerns about residual confounding or reverse causality. Additionally, although our analytic approach estimated indirect effects using SEM and GSEM sequentially, future work may benefit from unified modeling frameworks that estimate mediation within a single longitudinal model. Third, although the sample includes diverse SGM respondents, the absence of a non-SGM comparison group prevents us from assessing whether these pathways are unique to SGM older adults or generalize to other rural populations. Fourth, measures of community involvement did not assess the quality or trust dimensions of social ties, which may be more influential for health. Finally, the study focuses on the U.S. South, a region with unique structural barriers, limiting generalizability to other global or urban settings.

The results of this study underscore the urgent need for structural, policy, and clinical interventions to address cognitive health inequities among SGM older adults in rural communities. Policy reforms should prioritize expanding access to culturally competent, SGM-affirming healthcare providers in rural regions, alongside broader investments in public health infrastructure to address longstanding socioeconomic disadvantage.^20,61 Clinicians and mental health professionals working with SGM populations should be trained to recognize and address minority stress as a risk factor not only for mental health concerns but also for cognitive decline. Community-based organizations can be supported to develop programs that foster not only social participation but also high-quality, trusting, and resource-rich networks that may better protect cognitive health. Finally, broader anti-discrimination policies and legal protections remain essential to reduce chronic exposure to stigma, marginalization, and stress across the life course, with the potential to yield long-term cognitive health benefits for SGM communities.

Conclusion

Overall, this study extends minority stress scholarship by demonstrating that the consequences of structural stigma reach beyond mental health to shape cognitive trajectories. Minority stress exposure was found to accelerate the rate of cognitive problem growth over time, highlighting its cumulative and embodied effects. At the same time, community involvement, while associated with greater engagement and solidarity, did not buffer against cognitive decline, underscoring that resilience strategies alone are insufficient to counteract lifelong structural discrimination. The findings challenge models that overstate the protective potential of social participation and instead call for bolder, upstream interventions that address the social, economic, and policy-level conditions that generate minority stress in the first place. Promoting cognitive health equity for SGM older adults will require interventions that dismantle systemic inequities rather than rely solely on individual or community adaptation.

Highlights.

Lifetime rural exposure and education predicted minority stress

Minority stress increased community involvement but not cognitive protection

Higher minority stress exposure accelerated cognitive decline over time

Structural reforms are needed to reduce stress and promote equity