Discrimination, Mental Health, and Suicidal Ideation among Sexual Minority Adults in Latin America: Considering the Roles of Social Support and Religiosity

: Despite the high rates of discrimination and mental health issues experienced by lesbian, gay, bisexual, and transgender (LGBT) adults at an international level, very little research has focused on this population in Latin America. As a result, the purpose of this study was to explore relationships among discrimination experiences, mental health (i.e., anxiety and depressive symptoms), suicidal ideation, religiosity, and social support in LGBT adults in Latin America. A sample of 99 participants from the region completed an online survey assessing these constructs. Results suggested that 51.5% of the sample reported lifetime suicidal ideation. Experiences with harassment/rejection-type discrimination predicted suicidal ideation, while work/school-type discrimination predicted mental health problems (anxiety and depression). Depressive symptoms predicted suicidal ideation, while anxiety symptoms alone did not. Further, depressive symptoms mediated relationships between work/school discrimination and suicidal ideation, as well as between harassment/rejection discrimination and suicidal ideation. Moderated mediation analyses suggested that high, but not low or moderate levels of social support and religiosity buffered LGBT adults in Latin America against discrimination, speciﬁcally by weakening links among discrimination, depressive symptoms, and suicidal ideation. Clinical intervention research with LGBT adults in Latin America should focus on increasing social support networks for this population and on helping LGBT adults reconcile conﬂicts between religious identities and sexual or gender minority identities, as healthy religiosity may serve as a protective factor against mental health problems generally, and suicidal ideation speciﬁcally.


Introduction
Research from the U.S. suggests that gender-and/or sexuality-based discrimination (hereafter "discrimination") toward lesbian, gay, bisexual, and transgender (LGBT) individuals is all too prevalent, with over 50% of LGBT individuals reporting microaggressions, slurs, sexual or non-sexual harassment, and/or violence [1]. Discrimination occurs in a range of social settings, including public places such as stores and restaurants [2] and places of employment [3]. Moreover, LGBT individuals are disproportionately affected by mental health problems; for example, relative to their heterosexual counterparts, sexual minority (SM) individuals are more likely to experience anxiety and depression [4,5], while transgender individuals report psychological distress at rates that vastly exceed those of the general population [3]. Of particular concern are the high rates of suicidal ideation and suicide attempts among the LGBT community. In fact, almost half of LGBT individuals report lifetime serious contemplation of suicide [6]. Relative to their heterosexual peers, gay men are approximately three times more likely to report lifetime suicidal ideation and almost four times more likely to report lifetime suicide attempts, while bisexual women are five times more likely to report lifetime suicidal ideation when compared with heterosexual women [7]. The link between LGBT identity-based discrimination and poor mental health outcomes has been well established within literature from the U.S.; for example, SM women who report lifetime discrimination are at an increased risk of experiencing anxiety and depression relative to women reporting no discrimination [8], and transgender individuals who report higher rates of discrimination report increased symptoms of posttraumatic stress disorder [9], anxiety, and depression [10].
Despite the attention that the detrimental effects of discrimination have received in U.S.-based literature, societal non-acceptance of the LGBT community is not geographically constrained, but rather represents a global issue influenced by political and cultural factors [11]. Specifically, several Latin American countries have highly conservative leaders wishing to, along with faith-based groups, uphold traditional family and gender systems [12]. Such efforts consequently facilitate barriers to the constitutional protection of LGBT individuals and create cultural strain which in turn provides fertile ground for legal and social exclusion as well as discrimination [12]. Despite the LGBT non-inclusive sociopolitical climate in some regions of Latin America, peer-reviewed research from this region, while increasing, remains limited.
However, a set of cross-sectional research conducted among Mexicans highlights similar concerns as U.S.-based findings; for example, transgender individuals reported high rates of discrimination from family members, friends, and neighbors (all >30%), as well as violence (>24%), sexual harassment (>34%), and threats/insults (50%) [13]. Among the same sample, past-year suicidal ideation as well as past-year suicide were high, with over 58% reporting suicidal ideation and over 55% reporting suicide attempts [13]. Moreover, relative to those reporting low levels of violence, LGBT individuals who reported average or high levels of violence were more likely to report greater levels of depressive symptoms [14]. Among Mexican SM individuals, the most frequently reported forms of violence included verbal (32%) and sexual harassment (18%), while the most frequently reported forms of discrimination included not being hired (13%) and mistreatment by police (11%) [15]; additionally, 39% of participants reported suicidal ideation, while 15% reported having attempted suicide in the past [15]. Past-year discrimination in work settings was linked to an increased risk of mental health problems and suicide attempts. Similarly, physical, and sexual violence was linked to suicidal ideation and attempt. Sexual harassment was linked to suicide attempts, while sexual violence was shown to be a risk factor for suicidal ideation and suicide attempts [15]. Among a sample of Mexican adolescents, those who reported having had sexual relationships with individuals of the same gender were more likely to report depressive symptoms, suicidal ideation, and suicide attempts relative to adolescents reporting sexual relationships with the opposite gender [16].

Protective Factors
The high rates of discrimination and suicidal ideation and attempts necessitate investigation of protective factors in order to mitigate detrimental health outcomes among LGBT individuals in Latin America. U.S.-based literature has highlighted the role of social support. For example, social support has been linked to decreased depression in SM individuals [17][18][19]. Among transgender/gender non-conforming individuals, depressive symptoms fully explained the link between perceived harassment-and rejection-based discrimination and suicidality, while social support received from a significant other buffered the effect of harassment and rejection on suicidal ideation [20]. Moreover, past research suggests that transgender individuals who were out and perceived family members to be supportive reported lower levels of psychological distress relative to those who perceived their family members as unsupportive or neutral [3]. Lastly, Mexican SM women who report greater levels of community connectedness reported lower levels of depressive symptoms [21].
While the buffering effect of social support on mental health outcomes among LGBT individuals is supported by past research, evidence for religiosity is less clear. Yet, religiosity remains an integral factor of many Latin American cultures [22], and has been deemed to serve as a protective factor for suicidality among non-LGBT individuals from Brazil [23]. However, religion often includes conservative teachings that encourage heterosexual marriages while discouraging same-gender marriages [24]. Unsurprisingly, in LGBT samples from developed countries, religiosity has been linked to internalized homophobia [25] and sexual identity-based feelings of shame [26]. Additionally, non-heterosexuals who belong to non-LGBT-affirming religious communities face psychological distress [27][28][29]. Additionally, findings from mixed-methods research conducted among a sample among behaviorally bisexual Latino men living in the U.S. identified a link between religiosity and internalized homonegativity as well as increased loneliness, discrimination, and violence [30].

The Current Study
The current study expands the literature on mental health and its correlates for LGBT individuals in Latin America in two specific ways. As reviewed above, there is a broad literature from developed countries, and particularly the U.S., documenting relationships among discrimination experiences, mental health (i.e., anxiety and depressive symptoms), and risky behaviors (i.e., suicidal ideation) in LGBT adults. There is also literature, though scant, exploring connections between some of these constructs for LGBT individuals in Latin America, and with Latino LGBT individuals residing in the U.S. Within this literature, discrimination experiences have been found to be associated with higher levels of mental health problems [15,[31][32][33]. Thus, the primary aim of this study was to examine the relationships among discrimination, mental health, and suicidality in LGBT adults from Latin America. Given the literature linking religiosity and social support to mental health for LGBT adults in developed countries [17,18,[34][35][36][37], a secondary aim of the proposed study was to examine whether these relationships vary as a function of participants' social support or religiosity.

Participants
Participants (N = 99) were self-identified lesbian, gay, bisexual, and transgender (LGBT) adults, who were age 18 or older, currently residing in Latin America, and were able to read and respond to a survey in Spanish. Participants were recruited as part of an online survey study of sexual and gender minority adults in Latin America. In general, participants from the current study tended to be younger (M age = 24.38, SD = 5.97), upper (42.4%) and lower (42.2%)-middle class, residing in Mexico (92.9%), and have a bachelor's degree (68.7%). Additionally, most participants identified as a cisgender man (51%) or cisgender woman (40%), and gay/lesbian (75.8%). See Table 1 for more detailed demographic information pertaining to the study sample.

Measures
Consistent with Chapman and Carter's (1979) translating procedures for the crosscultural use of measures, all study measures were translated into Spanish by a bilingual and bicultural researcher and then back-translated into English by another bilingual and bicultural researcher. If any discrepancies emerged between the back-translated version and original English version, they were addressed mutually. The only measure that was validated in Spanish is the Multidimensional Scale of Perceived Social Support. However, this study was initiated in 2014, before Cobb and Xie's (2015) translation was available [38] and Edwards (2004) validated the measure among a sample of Mexican adolescents raising concerns about developmental appropriateness for use with adults [39].
Demographics. A researcher-designed form collected participant demographics, including participants' age, gender identity (man, woman, transgender man, transgender woman, intersex, other), sexual orientation (heterosexual (screen-out criterion), "heterosexual AND transgender, intersex, or other gender identity," gay/lesbian, bisexual, other), education level, family's social class, romantic relationship status, and country of residence.
LGBT Discrimination. Experiences with LGBT discrimination were measured using the Heterosexist Harassment, Rejection, and Discrimination Scale (HHRDS) [40]. The scale consists of 14 items, is comprised of three subscales (Harassment/Rejection, Work/School, and Other), and measures the frequency of discriminatory experiences occurring within the past year. Participants respond to items on a six-point Likert scale ranging from 1 (the event has never happened to the individual) to 6 (the event has occurred almost all of the time (more than 70% of the time)). In the current study, internal consistency for the HHRDS was acceptable overall (α = 0.89) and across the three subscales: Harassment/rejection (α = 0.85), Work/school (α = 0.77) and Other (α = 0.66).
Suicidality. To assess suicidal ideation, the Suicidal Ideation subscale of the Suicide Behaviors Questionnaire (SBQ-14) [41] was administered to participants. This subscale contains five items and assesses the frequency of past and current suicidal ideation. Its scoring algorithm weights current suicidality more highly than past suicidality in order to tap the frequency of a respondent's current ideation more strongly. Thus, the bulk of the score's index of suicidal ideation reflects ideation occurring sometime within the year during which participants complete the survey. A total score is calculated whereby higher scores indicate greater suicidal ideation. In the current sample, the SBQ had an acceptably high α = 0.87.
Mental Health. The Hopkins Symptom Checklist-25 (HSCL-25) [42] was used to assess the presence and severity of anxiety and depressive symptoms. The HSCL-25 is a 25-item self-report questionnaire, comprised of a 15-item Depression subscale and a 10-item Anxiety subscale. Individuals are asked to indicate how often each symptom has been bothersome or distressing over the past week. Responses range from 1 ("not at all") to 4 ("extremely"), with higher scores indicating more severe symptomatology. In the current study, item 6 of the HSCL-25 was inadvertently omitted from the survey when putting the scale into the online survey software. Item 6 is part of the Anxiety Subscale and reads: "Trembling." To address this issue, linear interpolation was used, whereby the last value before the missing value and the first value following the missing value were averaged to provide an estimate of the missing value for each participant's item score. Additionally, given that the Depression subscale includes an item of suicidal ideation, which would significantly overlap with the suicidal ideation outcome variable (r = 0.595, p < 0.001), this item was removed from the calculation of the subscales and total score, and this item was not used in calculating internal consistency of the total score or Depression subscale. Internal consistency was found to be good in the current sample across the total scale (α = 0.92), as well as the Anxiety (α = 0.87) and Depression (α = 0.90) subscales. Social Support. The Multidimensional Scale of Perceived Social Support (MSPSS) [43] is a 12-item measure used to assess degree of social support across three dimensions, including social support from family, from friends, and from a significant other. Individuals were asked to indicate, on a five-point scale ranging from 1 ("Strongly Agree") to 5 ("Strongly Disagree"), the extent to which they agree with each of several statements. Higher scores on each subscale reflect greater perceived social support. In the current sample, internal consistency was found to be strong across the subscales and total scale: Total scale α = 0.92, Friends α = 0.95, Family α = 0.91, and Significant Other α = 0.96.
Religiosity. The Religious Commitment Inventory-10 (RCI-10) [44] was used to assess level of religiosity in participants. The RCI-10 is comprised of 10 items, and maps to two subscales, including Intrapersonal and Interpersonal subscales. Individuals were asked to indicate the extent of their agreement with each of several statements on a five-point scale ranging from 1 ("not at all") to 4 ("totally"). The current study demonstrated acceptable internal consistency across the total score (α = 0.84) as well as Interpersonal (α = 0.88) and Intrapersonal (α = 0.70) subscales.

Procedure
This study, fully completed online via SurveyMonkey, took place from January 2015 to September 2016 and was cross-sectional in nature. Given that sexual and gender minorities generally, and those in some Latin American countries specifically, represent an extremely difficult-to-reach population, a chain-type online snowball sampling method was employed in order to recruit participants. Recruitment efforts included building partnerships with other professionals invested in improving the overall well-being of sexual minorities in Latin America, working directly with stakeholders in Latin America (e.g., directors of non-profit organizations aimed at stenting HIV transmission), reaching out to LGBT organizations with an online presence (e.g., Asociación Hombres y Mujeres Nuevos de Panamá; Acción Gay), and advertising over social media resources across Latin American countries. Once a potential participant was identified, they were provided with a survey link to the online survey study and provided informed consent. All subjects gave their informed consent for inclusion before they participated in the study. The study was conducted in accordance with the Declaration of Helsinki, and the study was approved by the host institution's ethics committee.

Data Analysis Plan
Preliminary Analyses. Prior to conducting the primary statistical analyses reflecting the study's primary aims, descriptive statistics (i.e., means, standard deviations, frequencies and percentages) of participants' mental health, suicidality, level of social support, and level of religiosity were computed. Based on the clinical cutoff scores, the percentage of participants that reported clinically significant scores on the HSCL-25 total scale as well as Anxiety and Depression subscales were reported. A correlation matrix was created to examine bivariate correlations among discrimination experiences, depressive symptoms, anxiety symptoms, suicidal ideation, social support, and religiosity, and to identify relevant demographic variables (age, education, and social class).
Primary Analyses. In order to identify the patterns of connections among discrimination, mental health, and suicidal ideation, in LGBT adults from Latin America, a series of simultaneous multiple regressions were performed. The first and second regressions included the three subscales of the HHRDS (Harassment/Rejection, Work/School, and Other) as independent variables and suicidal ideation as the criterion variable. The third regression included the three subscales of the HHRDS regressed onto the total score of the HSCL-25 (Anxiety and Depression subscales combined) as the criterion variable. The fourth regressed the subscales of the HSCL-25 onto suicidal ideation.
Two meditational models were developed using the PROCESS macro, Version 2.13 [45] using 5000 bootstrap samples to combine patterns of relationships that emerged among the primary variables under scrutiny in the prior series of regressions, whereby the variables with the strongest unique associations from the regressions were chosen for the mediational models. Because only the Work/School subscale of the HHRDS uniquely was associated with mental health, and only the Harassment/Rejection subscale of the HHRDS uniquely was associated with suicidal ideation (see Results section), separate mediational models were run testing Work/School discrimination to suicidal ideation through depressive symptoms and Harassment/Rejection discrimination to suicidal ideation through depressive symptoms.
Subsequently, these two meditational models were each expanded to moderated mediations (producing six moderated mediation models) with the PROCESS macro. The two mediations (for Work/School discrimination and Harassment/Rejection discrimination as independent variables and suicidal ideation as the criterion variable) were examined differentially as a function of participants' level of social support, as well as their level of Intrapersonal Religiosity and Interpersonal Religiosity, respectively.

Bivariate Correlations
Bivariate correlations were conducted for all primary study variables ( Table 2). Both Harassment/Rejection and Work/School discrimination were positively associated with suicidal ideation, depression, and anxiety. Other discrimination was only positively associated with suicidal ideation.

Descriptive Statistics
Descriptive statistics (i.e., means, standard deviations) of participants' mental health, suicidal ideation, level of social support, level of religiosity, and level of discrimination appear in Table 2. Based on the subscale clinical cutoff item average of 1.75 for the HSCL-25 [46], 5% of the sample met or surpassed the threshold for clinically significant depressive symptoms, and 3% for anxiety symptoms. Of the sample, 51.5% of individuals reported any lifetime suicidal ideation.

Regressions
In the first multiple regression with all three HHRDS discrimination subscales as independent variables and suicidal ideation as the outcome, the overall model was significant and explained 37.5% of the variance in suicidal ideation (F(3, 95) = 5.18, p = 0.002). Only the Harassment/Rejection subscale uniquely predicted suicidal ideation (β = 0.367, p = 0.015).
In the second multiple regression with the three HHRDS subscales as independent variables and the total score of the HSCL-25 as the outcome, the overall model was significant, accounting for 19.0% of the variance in mental health (F(3, 95) = 7.40, p < 0.001). Only the Work/School subscale uniquely predicted mental health problems (β = 0.349, p = 0.003).
In the third multiple regression, the Anxiety and Depression subscales of the HSCL-25 were entered as independent variables and suicidal ideation as the outcome. The overall model was significant, with anxiety and depressive symptoms accounting for 30.6% of the variance in suicidal ideation (F(2, 96) = 21.13, p < 0.001). Only depressive symptoms (β = 0.625, p < 0.001) uniquely were associated with suicidal ideation.

Mediations
In the meditational models, the most highly explanatory index of discrimination was specified to lead to the most highly explanatory index of symptoms of anxiety/depression, which was then specified to lead to suicidal ideation. In the multiple regressions, the Harassment/Rejection subscale of discrimination was the strongest explanatory factor of suicidal ideation and the Work/School subscale of discrimination was the strongest explanatory factor of mental health problems (combined score). Depressive symptoms were the strongest explanatory factors of suicidal ideation. Thus, two mediation models were run to examine relationships among Work/School or Harassment/Rejection discrimination, depressive symptoms, and suicidal ideation.
In the first mediational model, Work/School discrimination was specified to have a direct effect on suicidal ideation, as well as an indirect effect through depressive symptoms. The direct paths from Work/School discrimination to depressive symptoms (b = 0.33, p < 0.001) and from depressive symptoms to suicidal ideation (b = 9.21, p < 0.001) were both statistically significant. Further, the indirect effect of Work/School discrimination on suicidal ideation through depressive symptoms was statistically significant (b = 3.04, 95% CI (0.94, 7.33)), indicating a full mediation because the direct path from Work/School discrimination to suicidal ideation (c' path) was not statistically significant (b = 1.36, p = 0.344).
In the second simple mediation model, the same model was run but with Harassment/Rejection discrimination instead of Work/School. The direct paths from Harassment/Rejection discrimination to depressive symptoms (b = 0.16, p = 0.003) and from depressive symptoms to suicidal ideation (b = 9.04, p < 0.001) were both statistically significant. Further, the indirect effect of Harassment/Rejection discrimination on suicidal ideation through depressive symptoms was statistically significant (b = 1.46, 95% CI (0.29, 3.90)), indicating a full mediation because the direct path from Harassment/Rejection discrimination to suicidal ideation (c' path) was not statistically significant (b = 1.46, p = 0.098).

Moderated Mediation: Work/School as Predictor and Social Support as Moderator
The two meditational models were each expanded to moderated mediations (producing six moderated mediation models) with the PROCESS Macro using social support, Interpersonal Religiosity, or Intrapersonal Religiosity as the moderator. The first model with Work/School discrimination and social support was significant, F(5, 93) = 11.43, p < 0.001, R 2 = 0.38. A conditional indirect effect of Work/School onto suicidal ideation through depressive symptoms was observed: depressive symptoms were significant mediators of Work/School discrimination in explaining suicidal ideation when social support was low to somewhat high (10-75th percentile), but not when social support was high (90th percentile; Table 3). This pattern is reflective of a moderated mediation, and the mediational effect decreased linearly as social support increased.

Moderated Mediation: Work/School as Predictor and Intrapersonal Religiosity as Moderator
The second model explaining suicidal ideation was significant, F(5, 93) = 8.34, p < 0.001, R 2 = 0.31. A conditional indirect effect of Work/School onto suicidal ideation through depressive symptoms was observed: depressive symptoms were significant mediators of Work/School discrimination in explaining suicidal ideation when Intrapersonal Religiosity was low to somewhat high (10-75th percentile), but not when Intrapersonal Religiosity was high (90th percentile). This pattern is reflective of a moderated mediation, and again the mediational effect decreased linearly as Intrapersonal Religiosity increased.

Moderated Mediation: Work/School as Predictor and Interpersonal Religiosity as Moderator
The third model explaining suicidal ideation was significant, F(5, 93) = 8.88, p < 0.001, R 2 = 0.32. A conditional indirect effect of Work/School onto suicidal ideation through depressive symptoms was observed: depressive symptoms were significant mediators of Work/School discrimination in explaining suicidal ideation when Interpersonal Religiosity was low to somewhat high (10-75th percentile), but not when Interpersonal Religiosity was high (90th percentile). This pattern is reflective of a moderated mediation.

Moderated mediation: Harassment/Rejection as Predictor and Social Support as Moderator
The fourth model explaining suicidal ideation was significant, F(5, 93) = 10.74, p < 0.001, R 2 = 0.37. A conditional indirect effect of Harassment/Rejection onto suicidal ideation through depressive symptoms was observed: depressive symptoms were significant mediators of Harassment/Rejection discrimination in explaining suicidal ideation when social support was low to moderate (10-50th percentile), but not when social support was somewhat high to high (75-90th percentile). This pattern is reflective of a moderated mediation.

Moderated Mediation: Harassment/Rejection as Predictor and Intrapersonal Religiosity as Moderator
The fifth model explaining suicidal ideation was significant, F(5, 93) = 8.94, p < 0.001, R 2 = 0.32. A conditional indirect effect of Harassment/Rejection onto suicidal ideation through depression was observed: depressive symptoms were significant mediators of Harassment/Rejection discrimination in explaining suicidal ideation when Intrapersonal Religiosity was low to somewhat high (10-75th percentile), but not when Intrapersonal Religiosity was high (90th percentile). This pattern is reflective of a moderated mediation, with a mediational effect that decreased linearly as Intrapersonal Religiosity increased.

Moderated Mediation: Harassment/Rejection as Predictor and Interpersonal Religiosity as Moderator
The sixth model explaining suicidal ideation was significant, F(5, 93) = 9.50, p < 0.001, R 2 = 0.34. However, there were no conditional indirect effects as a function of Interpersonal Religiosity of Harassment/Rejection onto suicidal ideation through depressive symptoms. Depressive symptoms were significant mediators of Harassment/Rejection discrimination in explaining suicidal ideation at all levels of Interpersonal Religiosity (10-90th percentile). This pattern indicates that a moderated mediation was not found.

Discussion
The purpose of this study was to explore relationships among discrimination experiences, mental health (i.e., anxiety symptoms, depressive symptoms), suicidality, religiosity, and social support in LGBT adults residing in Latin America. In the current study, 5% of the sample reported clinically significant levels of depressive symptoms and 3% reported clinically significant levels of anxiety symptoms. These percentages are considerably lower than those found in samples of LGBT adults in the U.S. [47] and in studies utilizing other measures (e.g., the HSCL-25) in LGBT adults from other developed countries [48]. A possible explanation for these findings is that within Latin American cultures, psychological distress may be expressed through psychosomatic rather than psychological complaints. Indeed, this notion is well-supported by research with Latino individuals residing in the U.S. [49,50].
Participants from this study reported significantly lower levels of religiosity than even nonreligious U.S. adults [44]. Although Latin America as a region is heavily influenced by religion [51], it is possible that because the traditional tenets of predominant religions in Latin America (Catholicism, Protestantism) condemn same-gender attraction, identity, and behavior [52,53], LGBT adults in Latin America may avoid socializing with from religious communities to avoid discrimination and rejection. Although there is a paucity of research examining religiosity among LGBT adults residing in Latin America, research conducted with general LGBT adult populations and with LGBT Latino populations in the U.S. seems to support this notion [54,55].
LGBT adults in the present sample reported similar rates of suicidal ideation compared to other samples of sexual minority adults in the U.S. [37,56], rates which far exceed their cisgender and heterosexual counterparts [57]. Research with Latino adults from the U.S. has found responsibility to one's family and moral objection to suicide to be protective against suicidal ideation [58,59]. It is possible that among LGBT adults in Latin America, familismo (though not directly measured in the current study) was high, and mitigated levels of suicidal ideation.
Participants in the current study reported comparable levels of social support across the Significant Other, Friends, and Family subscales relative to a U.S. sample of cisgender SMW [56], and reported slightly higher levels of total social support compared with a sample of transgender adults [34]. These comparisons make sense in light of research documenting that women in general adult populations tend to report higher levels of social support than men across subscales of the MSPSS [43], as well as the research documenting lower levels of social support among transgender compared with cisgender individuals [60]. The current study's participants reported considerably lower levels of total social support relative to a sample of cisgender LGB adults [61].
Results were mixed in comparing levels of discrimination reported by participants in the current study to samples of LGBT adults in the U.S. Specifically, rates of Other and Work/School discrimination in the current study were comparable to rates reported in Szymanski's [40] sample of predominantly White and well-educated lesbian adults residing in the U.S., whereas rates of Harassment/Rejection discrimination reported in the current sample were higher. Compared with a sample of LGBT people of color in the U.S., rates of all three types of discrimination were lower [48]. One plausible explanation for the lower levels of discrimination reported in the current sample relative to Sutter and Perrin's [48] sample may relate to the impact of multiple minority identities on discrimination. Specifically, research from the U.S. has documented that compared with White LGB individuals, sexual minority people of color experience higher levels of heterosexist discrimination [62].
The overall pattern for discrimination is fairly consistent with other studies, with the exception of the Other subscale [46,61,62]. There are a couple of factors that may help to explain why the Other subscale of the HHRDS was not associated with anxiety and depression in the current study. First of all, because sexual and gender minority statuses may reflect "hidden" or invisible minority identities [63], it is possible that individuals may not be aware of others' sexual or gender minority status, and therefore may be less prone to enact this type of discrimination. Additionally, the type of discrimination tapped into by the Other subscale of the HHRDS represents discrimination that occurs through isolated encounters. Thus, the transient nature of this form of discrimination may not exert the same deleterious effects on LGBT individuals' mental health.
Discrimination, specifically the Harassment/Rejection subscale, was significantly associated with suicidal ideation in the first regression model, consistent with findings from previous studies using samples of LGBT adults from the U.S. [20,48]. In the second multiple regression, discrimination was associated with mental health (a composite of Anxiety and Depression subscales). This overall finding is in line with prior research conducted with LGBT populations from the U.S. [40,64]. Within this regression model, only the Work/School subscale uniquely was associated with mental health, which is surprising given that participants in the current sample reported the highest mean levels of Harassment/Rejection discrimination (relative to Work/School and Other discrimination). It is possible that relationships between participants and their family members and friends are not uniformly negative. For example, participants may have initially faced discrimination and rejection by family members and friends upon disclosing their sexual or gender minority status, but perhaps some friends and family were able to reconcile participants' stigmatized sexual or gender identities and maintain connected and supportive relationships with participants. Research on parents' reactions to sexual minority youths' disclosures of their sexual orientation has found that parents may progress through multiple stages and over time, moving towards acknowledgement and acceptance of their youths' sexual identities [65,66]. This same process may have occurred in families of LGBT adults in Latin America.
In the third multiple regression model, mental health (with the Anxiety and Depression subscales as separate independent variables) was associated with suicidal ideation, and while both anxiety and depressive symptoms uniquely were associated with suicidal ideation, depressive symptoms were the strongest unique associations with suicidal ideation in the model. There is some research linking anxiety symptoms to suicidal ideation [67]. However, there is a much larger body of research documenting positive relationships between depressive symptoms and suicidal ideation across a variety of populations, including Latin American immigrants [68], LGBT adults [69], and general samples of adults from the U.S. [70]. For this reason, it is not surprising that depressive symptoms were more powerful explanatory factors of suicidal ideation in the current study than anxiety symptoms.
Depressive symptoms fully mediated the relationship between Work/School discrimination and suicidal ideation. Depressive symptoms also fully mediated the relationship between Harassment/Rejection discrimination and suicidal ideation. These findings are consistent with prior research with LGBT adult samples in the U.S. documenting that mental health problems mediate the relationship between discrimination and suicidal ideation [20,48]. Sutter and Perrin [48] found that mental health (a latent variable comprised of anxiety symptoms, depressive symptoms, and satisfaction with life) mediated the relationship between LGBT discrimination (comprised of the three subscales of the HHRDS) and suicidal ideation in LGBT people of color. Additionally, in a sample of transgender adults in the U.S., Trujillo and colleagues [20] found that depressive symptoms mediated the relationship between Harassment/Rejection discrimination and suicidal ideation. Although LGBT participants in the current Latin American sample reported lower levels of depressive symptoms and suicidality relative to U.S. LGBT samples, harassment and rejection (Harassment/Rejection subscale), as well as unfair treatment by employers, supervisors or by educators (Work/School subscale) still negatively impacted participants' mental health, explaining suicidal ideation through depressive symptoms.
The present study documented a moderating effect of social support on the relationships among both Work/School and Harassment/Rejection forms of discrimination, depressive symptoms, and suicidal ideation. From a theoretical standpoint, these findings are in line with Meyer's [69] minority stress conceptualization, which posits that distal minority stressors such as discrimination are associated with mental health problems in LGBT individuals, and furthermore that protective factors such as social support may mitigate the impact of distal stressors on mental health. These findings are also consistent with research from developed countries documenting the buffering effect of social support on the relationship between discrimination and mental health for transgender individuals [20,71], and with literature documenting inverse relationships among social support and depressive symptoms and suicidal ideation among LGBT individuals [18,19,71]. While social support has been found to serve as a buffer in the relationship between discrimination and mental health problems, it seems that this buffering effect may only be apparent at high levels of social support.
The current findings suggest that high Intrapersonal Religiosity only may serve as a protective factor against discrimination by buffering the relationship between discrimination and negative mental health outcomes in Latin American LGBT individuals. This finding seems to contradict the trend in extant research in this area suggesting that intrapersonal facets of religiosity may confer risk for mental health problems among LGBT adults [37,72]. There are several considerations regarding this finding. First, it is possible that different types of discrimination negatively impact LGBT individuals' mental health to varying degrees. While unfair treatment in work and school environments has been found to negatively impact the mental health of LGBT adults [2], harassment and rejection by family members and friends may have especially pernicious effects on mental health problems for LGBT adults in Latin America given the close nature of relationships that LGBT adults likely have with these individuals. Second, religiosity should likely be considered in the context of its teachings and practice in which other individuals play a role in how religiosity serves or inhibits the well-being of LGBT individuals. For example, among a U.S.-based sample of LGBT individuals, attendance at a rejecting church was significantly correlated with greater levels of depression as well as increases in internalized homonegativity while religiosity was significantly correlated with lower levels of depression among participants reporting attendance of an accepting church [73]. Perceived religiosity of others and how their religiosity relates to one's LGBT identity may influence if, and to what degree, one's own religiosity can buffer or exacerbate negative mental health outcomes. For example, LGBT identity-related familial stigma has been found to mediate the relationship between parental religiosity and depressive symptoms and substance use among LGBT individuals, with greater familial stigma being linked to greater depressive symptoms and substance use [74]. Additionally, the relationship between religiosity of one's parents and familial stigma of one's sexual orientation was found to be greater in men than in women [74].
Given unique cultural influences present in Latin America surrounding religiosity, family, and gender roles, it is critical to consider the current findings in the context of cultural values. Many Latin American countries have made progress towards legal inclusivity of LGBT individuals in recent years, including marriage equality and antidiscrimination laws [12]. However, cultural factors embedded in Latin American cultures may contribute to perceptions of non-acceptance and discrimination of LGBT individuals. Latin American countries have high prevalence rates of Christianity [75]. Conservative religious views categorize homosexuality as morally faulty [76], and Christians report greater homonegative attitudes relative to non-Christians [77]. Particularly Protestantism has gained followers [78,79] who may exceed Catholics in their opposition to same-sex marriage [51]. Male gender socialization in Latin America is marked by cultural concepts including machismo, which assumes stereotypically masculine traits including aggressiveness and patriarchic dominance over family members [80]. Past qualitative research suggests that when negotiating one's gender and sexual orientation, some gay men preserve stereotypically male behavior while distancing themselves from gay men whose behavior is perceived as nonconforming to male gender traits and associated with stigmatization [81]. Familismo is the cultural concept of the centrality of family and the support derived from it; some research findings suggest familismo to be associated with positive health outcomes [82]. However, rejection of one's sexual orientation by family members has been found to negatively affect mental health of LGBT individuals; for example, U.S.-based research found a relationship between family rejection and depression and suicidality, with Latino gay and bisexual men reporting more negative responses from family than their White counterparts [83].

Clinical Implications
Given that depressive symptoms mediated relationships between Harassment/Rejection and Work/School discrimination and suicidal ideation for LGBT adults from Latin America, mental health professionals conducting research with or working with LGBT populations from this region might consider investigating and employing evidence-based techniques that have been found to attenuate symptoms depressive symptoms, in turn reducing suicidality for this population. This might be especially important for LGBT adults experiencing discrimination in work or school environments, as discrimination in these domains was most strongly tied to mental health issues for this population. Cognitive techniques such as challenging automatic negative thoughts and behavioral strategies such as behavioral activation are core components of cognitive-behavioral therapy and have been found to decrease depressive symptoms [84]. Likewise, because harassment or rejection by family members and friends was found to predict suicidal ideation in participants, family systems approaches might be used to simultaneously promote adaptability of families (e.g., to an individual's LGBT identity) and connectedness among family members, two hallmark components of family systems work [85].
Social support moderated relationships among discrimination, depressive symptoms, and suicidal ideation for participants in the current study. This finding underscores the importance of high levels of social support in protecting LGBT adults from Latin America against mental health problems. Accordingly, it seems important for mental health workers and other stakeholders in the LGBT community to expand existing LGBT social support networks (i.e., by increasing visibility and accessibility for LGBT adults), and also to create new networks in order to provide safe and supportive spaces for LGBT adults to connect and receive support. In Latin America, internet-based approaches might be used to establish online forums and groups, and to organize in-person clubs or meetings. Likewise, increasing social support from family members might involve helping families to reconcile conflicts between love for and connection with LGBT family members and anti-LGBT attitudes within family structures, moving towards acceptance and integration of the family, a value that is consistent with familismo.
In the current study, Intrapersonal Religiosity buffered mediational relationships among Work/School and Harassment/Rejection discrimination with depressive symptoms and suicidal ideation. As Haldeman [86] noted, religion and spiritual practice may "create a rich internal spiritual framework that soothes the anxieties stemming from sexual dissonance with social expectation and heals the wounds of a homonegative world" (p. 694). Likewise, research conducted with religious LGBT individuals in the U.S. has found that developing a spiritual or religious identity distinct from institutionalized religions may promote mental health for religiously affiliated LGBT adults [87]. Consistent with this research and in light of the current study's findings, clinicians working with LGBT adults in Latin America may focus on helping LGBT individuals to negotiate conflicts between their religious beliefs and LGBT identity and experiences in order to promote healthy and cohesive personal religious or spiritual identities.

Limitations and Future Directions
These findings should be considered within the context of several limitations. First, the current sample included very few transgender individuals, which may be in part due to the snowball recruitment methodology that limited which communities could be accessed. For this reason, study results may not be fully generalizable to transgender populations. Additionally, the vast majority of participants were from Mexico. The snowball-type sampling used by Mexican collaborators may have resulted in a particularly large subsample of participants from Mexico. Thus, results may not be fully generalizable to disparate Latin American countries or regions. Additionally, the bulk of participants were middle-class, well-educated, and very young (M age = 24.38, SD = 5.97). Thus, the current sample represents a small sub-group of young, primarily sexual (not gender)-minority individuals, residing in Mexico, and likely with much higher levels of literacy compared with the majority of individuals in this region.
Of note, a requirement for participating in the current study was having access to the internet, as data were obtained solely via an online survey. In order to capture a more heterogenous sample with respect to education level, class, gender identity, and location, similar research in the future could be carried out in-person and via paper-and-pencil administration in community venues traditionally frequented by LGBT individuals. Inperson recruitment may have fostered interpersonal relationships and personal alliances between researchers and potential participants, lending greater trust between potential participants and the research team, and in turn increasing the likelihood that more diverse LGBT adults from Latin America would participate in the study.
Another limitation to the current study is that the design was cross-sectional rather than longitudinal in nature, making it impossible to assert causality and establish temporality among variables. For example, although the present study's measure of suicidal ideation had a scoring algorithm that weighted current suicidality more highly than past suicidality to tap into the frequency of a respondent's current ideation more strongly, it was impossible to establish that discrimination experiences occurred prior to suicidal ideation.
Another potential limitation to the current study is that the RCI-10 does not tap other important aspects of religion, such as religious affiliation, religious coping, religious conflict, and spirituality (a concept that may overlap considerably with religiosity), religious identification, or discrepancy between parent and child religious beliefs. Religious affiliation in particular may be important to measure with LGBT individuals, as different religions and denominations within those religions have been found to be more or less tolerant of LGBT individuals [88,89]. Additionally, spirituality has been found to promote emotional well-being in LGBT adults [90], while religious conflict has been linked to suicidality for this population [72]. Additionally, the present study did not examine religious discrimination. This limitation is especially pronounced in light of high levels of discrimination that LGBT individuals face in religious communities, which negatively impact their mental health [91,92]. Future research should examine relationships between various types of discrimination (including religious discrimination), and their relative impact on mental health issues for LGBT adults in Latin America.

Conclusions
The current study exists among many other studies in its examination of relationships among discrimination, mental health, and suicidal ideation in LGBT adults from Latin America, and is the first study to the authors' knowledge to test a mediational model linking these constructs. The present study also appears to be the first to examine the potential moderating effect of social support and religiosity on the relationships among discrimination, mental health, and risky behaviors for this population. At high levels, social support and Intrapersonal Religiosity buffered LGBT adults in Latin America against mental health issues (depressive symptoms, suicidal ideation) in the face of different types of discrimination. Interpersonal Religiosity only protected LGBT adults in Latin America against mental health problems in the face of discrimination at work or school. Findings from this study suggest that clinical intervention research with this population should focus on increasing social support and, where appropriate, assisting LGBT adults from Latin America in negotiating conflicts between their religious and sexual identities in order to harness religiosity as a potential protective factor. Future work incorporating these implications may have the potential to improve mental health for a population that has generally been overlooked to date in the scientific literature.  Institutional Review Board Statement: The study was conducted in accordance with the Declaration of Helsinki, and was approved by the Virginia Commonwealth Institutional Review Board IRB #HM14577.
Informed Consent Statement: All subjects gave their informed consent for inclusion before they participated in the study.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy concerns.

Conflicts of Interest:
The authors declare no conflict of interest.